Evaluation of Pathologic Criteria for Acute Renal Allograft by hkksew3563rd


									 Evaluation    of Pathologic     Criteria                                                                                                                                        for Acute Renal
 Allograft   Rejection:    Reproducibility,                                                                                                                                          Sensitivity, and
 Clinical   Correlation

                                          ROBERT   B. COLVIN,*     ARTHUR     H. COHEN,t     CYNTHIA     SAIONTZ,
                                          STEPHEN   BONSIB,     MELISSA     BUICK,    BARBARA      BURKE,1    SHELLY   CARTER,t
                                          TITO CAVALLO,#      MARK    HAAS,**    ANNE    LINDBLAD,      J. CARLOS    MANIVEL,1
                                          CYNTHIA   C. NAST,     DANIEL    SALOMON,tt       CASEY    WEAVER,      and MARK     WEISS#
                                           *Massacht,setts                            General               Hospital,              Boston,                  Massachusetts;                       tUniversitv                  of C’alzfornia,                       Los        Angeles,                  California;
                                          Tlie              EMMES                  Corporation,                    Potomac,                  Maryland;                   Universitv                     of Iowa,             Iowa             City,         Iowa;           1Washington                        Hospital
                                          Center,               Washington,                      DC;            Universitv                  of Minnesota,                        Minneapolis,                       Minnesota;                                                   of Cincinnati,
                                          Cincinnati,                      Ohio;                                            of Chicago,                       Chicago,             Illinois;              Scripps                  Clinic             and       Research                Foundation,
                                          La       Jolla,               (alifornia;                and       Universitv                       ofAlabama,                       Birmingham,                       Alabama.

Abstract.                    This          study             was          designed              to evaluate                  the      pathologic                         diagnosis                  of rejection                   was        9 1 % for              a single               core         and        99%           for
criteria               used         for         acute          renal         allograft               rejection           that       were              devel-             two cores.                   To validate                  the diagnostic                      criteria,            the thresholds                        for
oped              by         a     panel               of      renal            pathologists                     participating                        in     the         number              of tubules                  with        tubulitis               and the percent                        infiltrate                were
Cooperative                         Clinical                 Trials           in Transplantation,                             a National                      In-        varied,             and         the       pathologic                    diagnosis                 was         compared                     with          the
stitutes               of        Health-supported,                                 multicenter                   research            group.                 The          clinical            course.             The       greatest               agreement                  occurred                with           a thresh-
panel             defined                 three             categories                of acute               rejection.              (1)         Type              I:    old of                 1 tubule            with          tubulitis            and       5%              cortex            with         interstitial
mononuclear                             infiltrate                 in      5%            of     cortex,              a total          of         at        least         infiltrate             (9 1 %).            Clinically                severe            rejection              episodes                 were            cor-
three            tubules            with           tubulitis                in 10 consecutive                         high-power                       fields            related            with          the type            of rejection                    (type           I, odds              ratio        [OR]            6.2;
from             the        most          severely                     affected          areas,          and         at least         two             of the             type         II,       OR         37.9).          Type             II     rejection                was         more               likely          to      be
three             following                      features:                  edema.             activated                lymphocytes.                            or       clinically                 severe          than          type        I (OR             6. 1 ). Analysis                     of other                 mdi-
tubular                injury.            (2) Type                  II: arterial,             or arteniolar,                  endothelialitis                            vidual             pathologic                  features              revealed               a correlation                       with          clinical
with            or      without                 the          preceding                 features.             (3)      Type           III:        arterial                severity               for      endothelialitis                         (OR         1 3.2),          interstitial                 hemorrhage
fibninoid                   necrosis               or transmural                        inflammation                      with        or without                         (OR          13.2),           and       the      presence                of glomerulitis                        (OR          3.7)          (all      P       <

thrombosis,                       parenchymal                            necrosis,            or hemorrhage.                        Using              these             0.05).           The          extent          of tubulitis                   or of the             interstitial                  infiltrate              did
criteria,               and         without                  any          knowledge                   of the          clinical             course               or       not correlate                    with         severity             (P     >     0.05).            It is concluded                       that        these
original                diagnosis,                     a rotating               panel         of three             pathologists                   agreed                 criteria           are simple,                   reproducible,                      and clinically                     relevant.                  These
with         the         original                study           pathologist’s                     diagnosis                of the           presence                    data       should              lead       to further               refinement                  of the          diagnostic                   systems
or absence                       of rejection                      in 259             of the 286                 biopsies           (91%)                  used          for      renal         allograft              rejection.                (J Am           Soc         Nephrol                8: 1930-1941,
for       this         analysis             (kappa                 =      0.80).        The        sensitivity               to establish                    the         1997)

Pathologists                        have            struggled                   to define              and         standardize                   criteria                olds       more              precisely,             such         as the            minimum                 criteria               of rejection,
for       the        diagnosis                   of renal                 allograft            rejection.              The         recent              effort            and        specifies                   five        grades               of         acute          rejection                 or        borderline
by Solez                    and colleagues                             led to the Banff                     working               classification                         changes.                Banff             has       proved                useful            in       provoking                     debate              and
(“Banff”).                       which                 has      stimulated                    much              discussion                  in         many              evaluation                    studies,            but       is rather               complex                in practice                      and         has
transplant                    centers              (   I ). Banff               attempts               to define             certain             thresh-                 limitations                   (2).
                                                                                                                                                                                Motivated                     by the         need           to develop                     practical               criteria             for       the
                                                                                                                                                                         large         multicenter                       trials          supported                   by       National                    Institutes                 of
                                                                                                                                                                         Health-National                            Institute             of Allergy                    and        Infectious                    Diseases,
                                                                                                                                                                         such          as       the           Cooperative                     Clinical               Trials            in      Transplantation
Received               January            31.      1997.        Accepted              June     25.      1997.
Drs.       R. B. Colvin,                   A. H. Cohen,                     S. Bonsib.          M. Buick,             T. Cavallo.                M. Haas.                (CCTT),                 and          benefiting                 from          the       nascent               Banff              formulation,
J. C. Manivel,                      C. C. Nast.                 C. Weaver,               and M. Weiss                   are members                    of the            the      center              pathologists                  reached              a consensus                      on       the       definitions
Cocperativc                  Clinical            Trials        in Transplantation                     Pathology          Review             Panel.                       for       a simplified                        scoring            system                (the        “CCTT                  System”).                    We
Correspondence                       to Dr.            Robert            B. Colvin,           Department              of Pathology,                   Massa-
                                                                                                                                                                         believe             that        a useful             classification                        system             must          fulfill           at least
chusetts             General            Hospital.              WRN            225       MGH.           32    Fruit      Street.       Boston.                MA
                                                                                                                                                                         four       major             goals.        (1) It must                  be practical                 to implement                          and easy
                                                                                                                                                                         to describe                   to unfamiliar                     personnel.                  (2)      Within               single           raters           or
1046-6673/0S0l                      2- 1930$03.0()/()
Journal           of the          American                Society          of Nephrology                                                                                 institutions                  and       between               many            raters          or institutions,                      the        system
Copyright               © 1997             by the            American              Society      of Nephrology                                                            must          be         reproducible.                       (3)         Sensitivity                    and          specificity                     rates
                                                                                                                                                                                                                                        Pathologic                Criteria           of Rejection                                    I 931

should            be high               to ensure                  that       positive              diagnoses                 and         negative               Diagnostic                         Criteria
diagnoses                 are         accurately                   identified.               (4)      The         system              should           be               Acute         rejection                 was        classified               into        one        of three               mutually                exclusive
clinically               informative,                   i.e.       , predictive                of course               and/or             response               types        according                    to the highest                      type        of lesion                 present.
to therapy.                  The            study           presented              here            describes                this      new          clas-                Type          I.           At      least         5%        of the cortex                        must          have             interstitial               mono-
                                                                                                                                                                 nuclear             infiltration                  with          at least            two        of the              following                  three         features
sification               schema                and       evaluates                 its performance                            in a series               of
                                                                                                                                                                 present:             edema,                   tubular            degeneration/injury,                                    or      activated                 lympho-
293       biopsies                with          respect              to reproducibility,                             sensitivity/speci-
                                                                                                                                                                 cytes.         In     addition.                 tubulitis               must          be present.                    with            at least            a total          of
ficity,        and          clinical               correlation.
                                                                                                                                                                 three        tubules              affected              in 10 serial                  high-power                        fields         (hpf:         X40)          from
                                                                                                                                                                 the areas              with            the      most           infiltrate.
                                                                                                                                                                        Type          II.           Arterial               or     arteriolar               mononuclear                           cell        endothelial                in-
Materials                         and              Methods                                                                                                       flammation                   (endothelialitis                         or endartenitis)                          is present                (with       or without
Patients                                                                                                                                                         features             of type                 I). A threshold                       of at least                    one         cell        under        the endo-
      The CCTT                   is a multicenter                      group           supported             by National                  Institutes             thelium             is required.                  A lymphocyte                         adherent                   to the luminal                         surface          of
of Health                whose              goal       is to develop                   and         evaluate           novel           therapeutic                an endothelial                         cell      is not sufficient                        for this                diagnosis.
protocols             in renal              transplantation.                     The      patients            in these             studies         were                 Type          III.              Arterial              or arteriolar                   fibrinoid                  necrosis              or transmural
all treated                 with         baseline              cyclosporin                   A (CsA),                 azathioprine,                  and         inflammation                           is present                and         may          or       may             not         be      accompanied                     by
corticosteroids,                      and      rejection             was        treated         with         pulse          steroids.           OKT3,            thrombosis,                   parenchymal                         necrosis/recent                         infarction,                  or hemorrhage.
antithymocyte                     globulin             (ATG).             tacrolimus,               mycophenolate                       mofetil.           or        These  categories                             are based on histologic                                         pattern. which     correlates
rapamycin.                  The         details         of the            protocols            (donor-specific                       transfusion,                with pathogenesis:                               types I and II are known                                           to be T cell-dependent.
donor marrow transplantation.         and rapamycin      treatment)     will be de-                                                                              whereas              type          III is generally                       regarded                 as antibody-mediated                                     (22).
scribed elsewhere.     Patients were recruited      by the individual       centers,                                                                                 Note            that neither  interstitial                              mononuclear      cell infiltration     alone                                              nor
and informed     consent    was obtained.    The patient      demographics        are:                                                                           tubulitis            alone is considered                                sufficient    for the pathologic       diagnosis                                                of
mean age 39 yr (range,          17 to 65 yr), 94 (68. 1%) men, 44 (31.9%)                                                                                        rejection             and          that        the presence                    of infarction                       or hemorrhage                           does       not
women,    89 (64.5%)       white,   42 (30.4%)    black,     2 (1.4%)      Asian,      5                                                                         constitute             type             III rejection. The                          area immediately                                under            the capsule
(3.6%)             other.             The       kidney              grafts         were         33      (23.9%)                cadaveric,               93       was not              scored,               nor was tubulitis                          or interstitial                          infiltrate             in areas of
(67.4%)             living-related.                    12 (8.7%)              living-unrelated.                      Eight          centers         con-         tubular      atrophy.     Other                              specific    diagnoses                          were recorded     (acute  pye-
tributed           a minimum                   of 21 (7.2%)                   biopsy         samples            to a maximum                       of 56         lonephritis,       thrombotic                                microangiopathy,                             etc.). Several  other features

( 19. 1%)           biopsy             samples.             Biopsies             were        taken       to evaluate                 episodes              of    that        were            not         part         of        the       definition                  of          rejection                were           scored            to

graft dysfunction  (n = 261)                                       or as protocol                   biopsies           (n           32), taken 9                 investigate                 correlation                    with         outcome                (e.g.        .    glomerulitis)                     and      to eval-

to 22 mo after transplantation.                                                                                                                                  uate        for potential                      inclusion               in future               classifications                         (Figure             1).
                                                                                                                                                                        The       following                    definitions               were          used         to identify                  pathologic                 features:

                                                                                                                                                                 .      Interstitial      Edema:    Tubular                                     separation     (tubular    basement                                                mem-
Process                  of Review                                                                                                                                      branes       showing     no points                                    of direct    contact)     by interstitial                                             fluid
      The CCTT                   Pathology                  Committee,                 consisting             of at least one pathol-                                   accumulation                           resulting               in clear              or       pale           areas.             Trichrome                   stain,
ogist        from           each            center,          was          formed          to        ensure           comparability                    and               when           used,             shows             slender             connective                        tissue          fibers            separated              by
consistency                  of the            diagnoses               across           centers.            The        committee                  estab-                irregular             nonstaining   areas.
lished       standardized                      pathologic              criteria         during          a series            of meetings               and        .      Tubular              Injury: Tubules lined by epithelial                                                     cells that may show                               one
has       since       reviewed                 all biopsies               taken         on or before                  October             25,      1995.                or more               alterations,                      such       as thinning.                          detachment.                   vacuolization,
The committee                         reached           a consensus                 definition               on adequacy,                     methods                     individual                cell         necrosis.               nuclear              enlargement.                            cytoplasmic                   baso-
of analysis,                    and      classification                   of acute             rejection.             The          criteria         were                philia.            or mitotic                 activity.            In rejection,                     the injury                 is seen            primarily
applied            to the             biopsy          by the           center          pathologist                who          completed               the                in association                       with        the        infiltrate.
pathology   score sheet (Figure      1) and made the “center diagnosis.”                                                                                          .     Tubulitis:                 Tubular                 inflammation                       in which                    one         or more               mononu-
These slides were then reviewed       by a rotating panel of three members                                                                                              clear         cells         have           penetrated                   into       tubules                 and          are        located          between
from the 1 1-member    Pathology     Committee     in five sessions held from                                                                                           tubular             epithelial                cells       or between                    tubular              cells           and      their        basement
December    1994 to January     1996, and a “panel diagnosis”       was estab-                                                                                          membranes.
lished.           This      diagnosis                 was         established             on slides            blinded             to the center                  .     Activated     Lymphocytes:       Lymphocytes                                                        that          are larger                than normal
pathology   diagnosis,      center                                 of origin, and any clinical information.                                                 If          (small    or resting)   lymphocytes.      which                                                    have            basophilic                 cytoplasm.
the committee’s        diagnosis                                   of rejection  was subsequently      found                                               to           a large             vesicular              nucleus               with         an open                    chromatin                 pattern.          and       one
differ        from     the center   pathology     evaluation,                                              the         committee      docu-                             or more nucleoli.
mented           its findings    on the pathology       score                                           sheet          (Figure   1 ) so the                       .     Endothelialitis:  Intimal inflammation                                                          resulting   from infiltration                                       of
nature of the discrepancy       could be determined.                                                                                                                    one or more mononuclear         cells into                                                       the subendothelial       space                                     of
    The minimum        criteria for adequacy   to ensure correct evaluation                                                                                               arteries           or,        less      commonly,                     arterioles                 usually               associated                 with        en-
included:    two cores of renal cortex and at least seven glomeruli        and                                                                                          dothelial              cell           cytoplasmic                  and         nuclear             enlargement.
two arteries    (larger than arterioles).    Specimens   were stained     with                                                                                    .       Fibrinoid                Necrosis:                  Necrosis               of arterial                    or arteriolar                     medial           and
hematoxylin                      and        eosin           and      either         periodic            acid-Schiff                  or periodic                          intimal            layers             and         replacement                      by       a homogeneous                                 eosinophilic
acid-methenamine      silver stains; Masson’s    tnichrome   was not required                                                                                             fibrin-like               material.                   Nuclear              debris             is also                usually             present.            and
but often used. A specimen       that did not meet these criteria was judged                                                                                              sometimes                 thrombi                are present.
adequate    for the presence    of rejection  only if it showed   features    of                                                                                  .     Glomerulitis:                         Segmental                  or global                glomerular                     hypercellularity                      due
rejection                (see         below).           Typing               the       rejection             required               two         arteries                  to mononuclear                           cell infiltration                       and in severe                              form         termed           acute
(unless           endothelialitis                     was      observed            in the single               artery          present           and/or                   allograft           glomerulopathy.                             with         endothelial                    cell        enlargement                     result-
arterioles).                                                                                                                                                              ing in occlusion                         of capillary                     loops.
 1932                        Journal          of    the        American                 Society             of Nephrology

                                                                                                                      COOPERATIVE                       CLINICAL        TRIALS           IN TRANSPLANTATION
                                                                                                                                                   PATHOLOGY                             FORM
                                     SITE                               PROTOCOL                                           CASE           NO.                 INITIALS

                                    L_J                             I     lI              I]                               Lu_i

                                    L_LJ                  LLJ                   L....LJ              BIOPSY            DATE
                                          M                     D                   V

                               51                                                                                                                                                        OBSERVATIONS:
                                                   Li Was                this a perioperative                         biopsy?             (1-No,        2-Yes)
                                                                                                                                                                                         Li                    Tubulitis   in 10 serial high-powered                                      fields
                                                   SPECIMEN                      ADEQUACY:                                                                                                                            (0: None, 1: 1 or 2 tubules,                                    2: 3 tubules)
                                                                                                                                                                                         LLJ                   Number             of high-powered                        fields       with
                                                   U                              Number               of cores
                                                                                                                                                                                                                         tubulitis          in series             of 10 fields
                                                   LLLJ                  Number                 of glomeruli                                                                             LJ_LJ                 Cortex          with      interstitial             infiltrate       (%)

                                                                                                                                                                                         LI                    Interstitial           edema             (0-None,               1-Present)
                                                   1_Li                  Number                 of small          or large         arteries
                                                                                                                                                                                         U                     Tubule           injury (0-None,                      1-Present)
                                                   L_LJ                  Number                 of sections
                                                                                                                                                                                         Li                    Reactive            lymphoblasts                      (0-None,            1-Present)

                                                   Li      Specimen                     adequate               (1-No, 2-Yes)                                                             Li                    Glomerulitis                 (0-None,               1-Focal,           2-Severe)
                                                                                                                                                                                         Arteries                (0-None,             1-Present)
                                                   Li      H and               E sections              prepared                (1-No,          2-Yes)
                                                                                                                                                                                                    Li         Endarteritis
                                                   Li      PAS or silver sections                                 prepared               (1-No,         2-Yes)
                                                                                                                                                                                                    Li         Fibrinoid           necrosis

                                                   Li      Trichrome                      sections            prepared            (1 -No,         2-Yes)                                            Li         Mucoid            intima        change

                                                   DIAGNOSIS:                        (1-No,           2-Yes)                                                                                        U          Transmural                   arteritis

                                                                                                                                                                                                    Li         Intimal fibrosis
                                                   Li      Morphologic                          abnormalities
                                                                                                                                                                                                    Li         Medial          degeneration,                      arterioles
                                                   Li      Acute               tubular           necrosis
                                                                                                                                                                                                    U          Hyaline           arteriopathy
                                                           Findings                consistent                 with CsA toxicity                                                                                                                                                                     54
                                                   Li                                                                                                                                    Thrombosis                        (0-None, 1-Present)

                                                   U       Acute               rejection            (1 -No,        2-Yes)                                                                           Li         Glomeruli                                Li        Arteries
                                                                         Ifyes,           specify           type:                                                                                   Li         Arterioles                               U         Venous
                                                                         U        Type of rejection                        (1 -Type            I, 2-Type         II, 3-Type              Other            (0-None,               1-Present)

                                                   U       Hyperacut                           rejection                                                                                 Li                    Venulitis

                                                   Li      Non-diagnostic                           acute         injury                                                                 Li                     Plasma           Cells         (0-None,              1-Focal,            2-Diffuse)

                                                                                                                                                                                         Li                     Interstitial          hemorrhage
                                                   U       Chronic                rejection                (Complete              Late
                                                                                                                                                                                         Li                     Renal          infarction
                                                                                   Biopsy/Chronic                      Rejection               Form)
                                                                                                                                                                                         Li                     Interstitial          fibrosis/tubular                   atrophy
                                                   Li      Non-diagnostic                           chronic           injury
                                                                                                                                                                                         Li                    Other        features
                                                   U       Other               diagnosis             (e.g., AIN, BNS, CMV)                                                               If present,               specify
                                                   If yes,          specify
                                          (Am           mcnt

                                          Signature                     of CCTT                Pathologist:.

                                                                                                                                Figure            1. Pathology                form        score           sheet.

.        Venulitis:           Mononuclear                       cell           infiltration                 of the intima                  and      media          of         of >0.75              is interpreted                     as excellent                    agreement,                 and       a kappa             of    <0.40
         small      veins         or venules.                                                                                                                                 represents                 poor        agreement:                   intermediate                    values            represent            fair        to good
                                                                                                                                                                              agreement                  (3).
         The       kappa          statistic              was             used           to test             the       degree              of agreement
between             the center            and the panel                           diagnosis                 in the presence                      or absence                   Sensitivity
of rejection                and     in the type                     of rejection.                    This         statistic          is a measure                  of            An ad hoc                       investigation                    was          conducted                 during          three      panel             review
inter-rater            reliability                 between                independent                        raters        (3).      The          kappa          sta-         sessions           to       assess           the        absence                or     presence              of    rejection           in      individual
tistic         estimates            the       agreement                      between                 raters         beyond               the      agreement                   cores       from            I 30       multiple-core                      biopsies,               and       to estimate              the      sensitivity
that could                 be observed      based on chance      alone.   If there is complete                                                                                of these         criteria.              The diagnosis                          of presence               or absence              of rejection               was
agreement                  between   raters. the kappa statistic     will be 1 . If there is some                                                                             determined                  by       the     pathologic                   diagnosis                based         on     the     assessment                 of at
agreement                  beyond         chance,                   it is between                     0 and            I . Generally,                   a kappa               least      two cores.                 Single-core                   biopsies              were          excluded              from    this analysis.
                                                                                                                                                                                                                                       Pathologic                Criteria        of Rejection                                         1933

Single-core                sensitivity                   was calculated   as the probability   that                                               each     core      considered               negative               for        clinical          rejection.                 These          biopsies                 were             per-
is positive              for rejection                     given that rejection   was present.                                                                       formed           to comply                  with       protocol              requirements                       and      were            not obtained
                                                                                                                                                                     due     to clinical                  suspicion of rejection.       None was                                        treated            for rejection.
Correlation                         with             Clinical                    Course                                                                                    Odds ratios                   (OR) for clinically      severe rejection                                         were           calculated      fr
                                                                                                                                                                     several         pathologic                 characteristics.                    The          OR         describes             the     odds          of being
    To determine                       whether                  the CCTT                pathologic                  criteria        correlated             with
a clinical             course               compatible                   with         rejection              (or       no rejection),                 a defi-        clinically            severe           for each              unit       increase              of the particular                      characteristic

nition        of       acute           rejection                 based           solely            on       clinical             observations               was      compared                with         the      odds          of the           reference                 group.          Logistic              regression
                                                                                                                                                                     analysis          (4)         was      also         used         to examine                   multivariately                   the         relationship
necessary.               In consultation                           with         CCTT               transplant              physicians.              the fol-
                                                                                                                                                                     between              clinical             severity            and       pathologic                     findings.             The          usual            linear
lowing         simple               objective                criteria            for a clinical                    course          likely         to be due
to rejection                  were          developed:                                                                                                               regression              analysis              setting            assumes             the         response              variable              to      he con-
                                                                                                                                                                     tinuous.          Logistic                 regression               models               are      designed                  to relate              a binary
1 . Peak             serum             creatinine                  (Cr)          during             the      first         I 4 d after             onset        of   response              variable,             such           as clinical              severity.              to explanatory                         variables
        suspected               rejection                 > 120%                of baseline                 Cr                                                       (e.g..        type       of rejection.                      edema,             activated                 lymphocytes.                       etc.).          Each
2. Episode                    is treated                 with         antirejection                  therapy            (e.g..       pulse         steroids,         regression               coefficient                  obtained               from           logistic            regression                  analyses                   is
   OKT3,                ATG)                                                                                                                                         interpreted              as the log of the OR for that                                       particular               explanatory                   variable
3. One             or both             of the following:                                                                                                             after        controlling               for other              variables              in the             model.

.   With            antirejection                    therapy.              Cr returned                     to within              20% of baseline
    within             14 d                                                                                                                                          Results
S   Cr on day                    14 is lower                    than       the highest                    level       during         the first            3 d of     Samples
    rejection                 and      is less             than         5 mg/dl                                                                                            Two         hundred                  ninety-three                      renal            allograft               biopsies                taken               for
                                                                                                                                                                     diagnostic                    or       protocol-scheduled                                     purposes                 on           I 38          patients.
    Cases            that did not respond                                 completely                    to steroid             therapy            were also
                                                                                                                                                                     sampled                from            4 to 700                  d posttransplantation                                      (median.                  42 d),
considered                to be clinical                        rejection.             Cases              with       a clinical             diagnosis           of
obstruction,                   urine             leak,       or renal                artery          stenosis              were       excluded.               Be-    were evaluated                            in this report.       Seven cases were removed                                                                   from
cause        of the known                         variability                  of cyclosporine                       levels        and      the inherent             the reproducibility                             analysis     because   of no kidney    tissue                                                              (ii

problem             with specifying                             a known              nontoxic              level for cyclosporine.                            our     I ), insufficient                        material       for the committee   to review      (ii                                                        =          2).
clinical            definition                   does           not      include              cyclosporine                       levels.       However,              insufficient                   histologic                   stains           (ii     =         1 ). or missing                           data        on           the
cases        are       treated              as     nonrejection                      if the         only          treatment              is the      discon-         original              pathology                     score              sheet          (ii        =        3).      The             review              panel
tinuation             or reduction                       of cyclosporine.                                                                                            scored            34          specimens                     as      adequate,                     which               did          not        meet                the
    This           clinical           definition                  of     rejection             is based               on      objective            observa-          adequacy                definition                  according                  to the            reported              center              pathologic
tions. It requires that Cr rises,                                          that antirejection treatment  is initiated.                                               data.         These            cases          are      included                in the            reproducibility                          analysis                 as
and that the Cr subsequently                                               lowers, which is the accepted      clinical
                                                                                                                                                                     nonrejection     for                        the center      assessment                                     and as scored                             for          the
picture        of acute                rejection.
                                                                                                                                                                     panel assessment                             (seven    rejections      and                                 27 nonrejections).
    To evaluate                       the        thresholds                of the            pathologic                criteria            for rejection,
the pathologic    diagnostic                                    criteria (tubulitis.    percentage    infiltrate)  were
varied to see what effect                                       that had on the sensitivity        and specificity    of                                             Reproducibilirs’
the pathologic     diagnosis                                    of rejection    (using the clinical diagnosis         of                                                The center  pathology                                         and     the panel                     assessments                       agreed              with
rejection   as the standard).                                     We recognize       that normally     the pathologic                                                respect           to the presence                            or absence                      of acute              rejection                  in 259               of
diagnosis               is      the         standard                  rather          than          the      clinical             course,           but       this   the 286 biopsies     (9 1% ) (Table     I ). The review     panel  agreed
approach               was          necessary                   to evaluate                  the      most           appropriate               thresholds            with a center diagnosis    of rejection     regardless  of type in 167 of
for the pathologic                               findings.                                                                                                            179 (93%)                    of the cases                    and        with         a diagnosis                      of no rejection                              in
                                                                                                                                                                     92 of           107       (86%).              Inter-rater                 reliability                    between               the          center               and
Correlation                           with           Clinically                       Severe                Rejection                                                panel          diagnosis                   yielded               a kappa                 statistic              of 0.80              (95%              confi-
        Biopsies              were          analyzed               to determine                      whether               the type          of rejection            dence           interval              [CI],         0.73         to 0.87).               Thus,            the      agreement                      between
or any specific pathologic feature correlated with clinical severity, as                                                                                             the centers                     and the panel     was in the excellent      range.      The
defined  in the CCTT protocols    (an episode that is steroid-resistant,
                                                                                                                                                                     agreement                     rates for acute rejection  in the five review      sessions
treated with ATG, OKT3, or FK5O#{212}. began within 10 d of trans-
plantation).                  Steroid              resistance                  was     defined              as: (1) failure                  to improve
serum         Cr to within                         35%          of baseline                  within           5 d of initiation                    of pulse
steroid        therapy   (50()     mg/d with no additional                                                         antirejection  treatment
                                                                                                                                                                     Table           1. Table                  of center                 versus           review                panel’s                 assessment                      of
during         that time);     (2) a value  of serum    Cr                                                       48 h after steroid     pulse                                                rejection’
initiation            that       exceeded                   that        of the Cr on the day                                of initiation;                or (3)
subsequent                    treatment                  with         OKT3,            ATG,             or FK506.                                                       Review              Panel                                                                                                                 Total
    Two            hundred              five biopsies                          were       evaluated                  for this portion                     of the                                                    Rejection                           No Rejection
analysis,            including                    17 1 of 189 biopsies                             taken          at the time               of suspected
rejection             and 32 scheduled                                 protocol              biopsies              taken         from        9 to 22 mo               Rejection                                  167(93%)                                 15(14%)                                         182 (64%)
posttransplantation.                               Fifteen              cases          were           removed                 because             of made-            No          Rejection                         12 (7%)                               92 (86%)                                        l()4         (36%)
quate        Cr data.                  and          three          cases             were          removed                 because            of missing
                                                                                                                                                                      Total                                      179 (100%)                             107 (100%)                                       286 ( l()0%)
pathologic      data                  (percent   cortex                         or the number       of tubules   with tubuli-
tis). Available                       Cr information                             on protocol-scheduled         biopsies   was                                                 a   Kappa        =        0.80       with         a 95%          confidence                     interval            of 0.73              to 0.87.
limited         and inadequate      to assess for                                                   clinical            rejection    as defined                      The percent observed                                 agreement                 is 90%,                 and the percent                       expected
above.         However,   the protocol-scheduled                                                     biopsies            are included and are                        agreement is 53%.
 1934                            Journal           of the American                      Society           of Nephrology

 Table              2. Table                   of center                  versus           review                panel’s               assessment                         of rejection’

                    Review               Panel                                                                                                                                                                                                                                                                             Total
                                                                                      No Rejection                                                                   Type             I                                              Type        II

                  No       Rejection                                                       92 (86%)                                                             12 (9%)                                                             0 (0%)                                                                         104         (36%)
                  Type           I                                                         12(11%)                                                            115(80%)                                                              6(17%)                                                                         133 (47%)
                  TypeIl                                                                    3(3%)                                                               16(11%)                                                        29(83%)                                                                              48(17%)
                  Total                                                                107(100%)                                                              143(100%)                                                        35(100%)                                                                            285 (100%)
       a    The          one      type           III case           (who           went       on to experience                             graft              failure)                is not        included             in the table.                  Kappa              =     0.72 with                 a 95%           confidence
interval                of 0.65            to 0.79.             The        percent           observed                 agreement                       is 83%,             and             the percent                expected               agreement                   is 39%.

ranged                   from            86 to 95%.                        The        overall                  agreement                   rate               for       the                specimens                     showing                  type             II     rejection,                  14       (82%)             showed                  the
presence                   or absence                     of endothelialitis                               was         91%             (260              of 285                            diagnostic                   changes               in only                   one      core.
with             a kappa                 value          of 0.65             (95%            CI, 0.52                 to 0.78).                There                  were                         The        clinical                course           was          compared                    with         the pathologic                          diag-
modest                   but not statistically                              significant                   differences                      in the kappa                                    nosis         (Table                4).      The           center             pathology                   diagnosis                  had          a sensi-
statistic                when             biopsies                  were           grouped               for         time          after              transplan-                           tivity        of 85%                 (    106 of           1 24)         and         a specificity                    of 72%               (58       of 81)
tation            of the biopsy                      (range,              0.68       to 0.89),                 for center              (range,                 0.46          to            with         respect                to the         clinical                  definition              of acute                  rejection.              Sim-
0.91).              and         for therapeutic                           protocol               (range,              0.67          to 0.81)                    as as-                     ilarly,           the      panel           diagnosis                  had          a sensitivity                 of 86%              (109          of 127)
sessed              by overlapping                             CI     without               adjustment                     for      correlation.                                           and a specificity       of 72% (56 of 77).                                                            The           clinical            and center
       For          type         I and            type          II rejection,                    the agreement                             between                      the                pathologic    diagnosis      agreed in 84%                                                            (83           of 99)             of biopsies
review                  panel            and       center             diagnosis                  was            80%          (1 15 of 143)                             and                 diagnosed                   as type              I rejection                  and       in 92%              (23         of 25)          of biopsies
83%              (29      of     35),          respectively                      (Table            2). The             total           agreement                         for               diagnosed                    as type               II rejection.                        Of         the       32         protocol                 biopsies
type             I/type          IL/nonrejection                            was         83%             (236          of 285).                    The           kappa                      taken          from            9 to 22 mo                          after            transplantation                         in patients                   with
statistic                 was           0.72       (95%               CI,        0.65        to 0.79)                  for       the        presence                      or               normal              function,                    93.8%                showed                no       rejection,                  3. 1 % showed
absence                   of       endothelialitis,                           also          indicating                     very            good                 agree-                     type         I rejection,                   and        3.1%              showed               type         II rejection.                     To      inves-
ment.              The          most         common                   discrepancy                       was          an upgrading                              of type                     tigate            the        sensitivity                    of        the          agreement                    rates          to      the         clinical
I to type                 II rejection                    by the panel                      (n      =          16) due             to appreciation                                         definition,                  the          definition                    was          modified                   and         agreement                    rates
of         the      initially               unrecognized                            endothelialitis.                         Other                discrepan-                               were          recalculated.                        When                 the         increase               from             baseline               Cr      was
cies         included    upgrading                                     from no rejection     to type                                              I    (ii      =       12)                changed                  to 5, 10, or 30%,                               rather          than         the 20%,                   no substantial
and          downgrading       from                                   type I to no rejection      (n                                          =          12).         The                  change              in agreement                           was           observed                  (data          not          shown).              There-
upgrading                       changes                 were           due          to the          review                 panel            recognizing                                    fore, the                 original      definition                             (20%            increase)                    was         used            in all
increased                     tubulitis,                and         the      downgrading                             changes               were                 due       to               subsequent                   calculations.
the         review               panel            lowering                  the       percent                  of cortical                 infiltrate.
                                                                                                                                                                                           Evaluation                         of the Pathologic        Criteria                                             of Rejection
Sensitivity                       of       Diagnosis                                                                                                                                          A primary                         purpose  of correlating         the                                         clinical   course                         and
       To assess                        the diagnostic                       sensitivity                   of an individual                                    biopsy                      pathologic                   diagnosis                 of rejection                     was         to evaluate                     the appropri-
core,             we          analyzed                  the         frequency                 of        discrepancy                        in          multiple-                           ateness             of       the          chosen             pathologic                     thresholds.                     Altering               the     de-
core             biopsies,                i.e. , those               in which               at least             one         core          was               positive                      gree         of         tubulitis            and           percent                 cortex           with          interstitial                infiltrate
for         rejection                    and       one           of        the       companion                        cores            was               negative                          components                          in the            CCTT                   definition               of        pathologic                    rejection
(Table                  3). Of the                   130            multiple-core                       biopsies                 analyzed,                           three                 substantially                       affected                 the        agreement                    rates            for       rejection                 (and
were              inadequate                      (no         rejection               and        too           few         glomeruli).                         Of       the                nonrejection)                        (Table            5).         As        the      threshold                 for      percent              cortex             is
 127 adequate                            biopsies,              only         one        of the           cores             met       the criteria                       for                increased                  to 25%,               a marked                    and       progressive                      loss        of agreement
rejection                  in      13 cases                   (10%).              There           were               131      positive                       cores        of               between                  a clinical               course              of rejection                   and         pathologic                   rejection
 144         total         cores            from          patients                 with          rejection.                  so that                  the calcu-                           is evident,                 which            falls          from             82 to 63%.                   At the             same          time,          only
lated             sensitivity                  of a single                   core          was      91%               (131         of 144).                     Of the                     a modest                  increase               in the            agreement                  with         a clinical                course              of no
 I 14            cases          (90%)              showing                   concordance                             among              all           cores,             58                rejection                 occurs            (from            76 to 84%).                         Decreasing                     the number                     of
showed                   no      rejection                in all cores,                     39 showed                       type        I rejection                       in               tubules             with           tubulitis               from          3          tubules           to              1 tubule             causes             the
all cores,                 and           three          showed               type          II rejection                    in all cores.                        In the                     agreement     rate for rejection   to rise from 82 to 9 1 %,                                                                                     whereas
remaining                        14       specimens,                       type         II rejection                       was          diagnosed                         in               the agreement      rate for no rejection   decreases slightly                                                                                    from 76
only              one         core,            whereas                the          other          core           showed                 either                  type              I        to 72%.             The         maximum                      agreement                      with         clinical            rejection               (91%)
rejection                  (n      =        12) or was                     inadequate                    (ii     =      2). Thus,                      of the             17               occurs             with         the        criteria              of            1 tubule              with           tubulitis                and         5%
                                                                                                                                                                                           cortex             with             interstitial                   infiltrate.                Removing                       the        criteria              for
                                                                                                                                                                                           edema,              tubular                injury,            and            activated               lymphocytes                         resulted                in
.1   One case was                       removed          due        to inadequacy                  for assessment                    of endothelialitis.
                                                                                                                                                                                           the       reclassification                           of       two            cases          (that          had          been           deemed                 no
                                                                                                                                                                                                                                    Pathologic                   Criteria          of Rejection                                           1935

Table             3.       Summary               of characteristics                           of      multiple-core                        specimens

             No. of Cores                                                                                                                                                                                                                                    Total No. of Multiple-Core
             in Specimen                                                                                                                                                                                                                                          Specimens Reviewed
                                                              In All Cores                                          In One              Core        Only                           In No Cores

                       2                                       48(45%)                                                      13(12%)                                                  46(43%)                                                                                      107(100%)
                       3                                           8(40%)                                                       0(0%)                                                12(60%)                                                                                       20(100%)

                       Total                                   56                                                           13                                                       58                                                                                           127
    a    Percentages                  are based              on row           totals.

Table             4.       Presence            or absence                   of clinical                rejection                by                               P      >   0.05).              Two           variables                 that        did correlate                             with         severity                  (P       <

                           pathological                rejection                                                                                                 0.05)        were             glomerulitis                  (OR               =      3.7)            and interstitial                           hemorrhage
                                                                                                                                                                 (OR        =        13.2).
                                            Pathologica            I Rejection
    Clinical                                                                                                                                                            Pathologic                     characteristics                         with              at least                four            cases                with           the
                                                 +                                    -                                                                          characteristic                        present            as determined                                   by      center              pathologic                          data
                                                                                                                                                                 were         considered                      jointly          for logistic                           analyses                 (the        type               of rejec-
         +                            106(85%)                              23(28%)                                       129           (63%)
                                                                                                                                                                 tion       and          endothelialitis                        were                not          used             simultaneously                                  due         to
         -                              18 (15%)                            58 (72%)                                       76 (37%)
                                                                                                                                                                 their        high             correlation).                   The             final             logistic                model                  includes                     the
         Total                        124 (100%)                            81 (100%)                                     205           (100%)                   following:                     type          I rejection                      (OR               =         2.0,         P       =         0. 1 5),                type           II
                                                                                                                                                                 rejection               (OR                  1 1 .8, P         =        0.03),              edema                (OR           =        4.3,         P       =     0.01),
                                                                                                                                                                 and        tubular                injury           (OR             =      4.9.              P        =        0.02).           The              presence                     of
                                                                                                                                                                 edema             and          tubular            injury,                as well                 as increasing                            type               of rejec-
rejection)                 as rejection               (< I %).              One           of these        cases           had a clinical
                                                                                                                                                                 tion,        increased                   the       odds             of        rejection                       being           clinically                         severe.
diagnosis                   of rejection,               and        the       other           did      not.
                                                                                                                                                                 Logistic                regression                  analysis                      examining                       clinical                 severity                      was
                                                                                                                                                                 also       performed                     on       122     biopsies                       in which                  the         review                 panel               had
Correlations                          with Clinical Severity
                                                                                                                                                                 diagnosed                     rejection.               When              the         same                pathologic                      characteristics
    For this evaluation,                                 171       biopsies                  taken        at the           time            of sus-
                                                                                                                                                                 were           considered,                       the      following                             model                 was            chosen:                      edema
pected    rejection     had adequate    clinical     data to determine          clinical
severity     (see Materials      and Methods).        Table 6 lists the descrip-                                                                                 (OR          =      5.9,          P      =       0.02)         and                endothelialitis                             (OR              =         5.7.           P       =

tive statistics      and univariate   OR for diagnosis            of rejection         and                                                                       0.10).           Similar               to the above                       model,                     the presence                         of edema                        and
the acute pathologic         features  significantly       associated        with din-                                                                           type       II rejection                    (endothelialitis)                             increases                     the odds                    of rejection
ical severity       (P < .05). Individual          pathologic       features       found                                                                         being            clinically                severe.            Forward                       stepwise                   selection                     procedures
not to be associated                               with clinical severity   (considered   as van-                                                                with         ‘.     =         0.20         for     inclusion                       were               used            for       model                    selection;
ables  independent                               of the presence     or type of rejection)      in-                                                              backward                  stepwise                procedures                        were              used         to validate                        the model
cluded:                acute         tubular          necrosis               (ATN),                mucoid            intima               change,                selection.
thrombosis                       (glomeruli     and                      arterioles),                  venulitis,       interstitial                                    Compared                    with          type      I rejection,                          the odds                    of type                 II rejection
plasma     cells                  (focal or diffuse),                       interstitial               fibrosis/tubular           atro-                          being            clinically                  severe            were                6. 1 (95%                      CI,          0.8         to         47.8).                An
phy,         percent             infiltrate,           and the number                          of hpf with                  tubulitis               (all         additional                    logistic           regression                       analysis                 was          performed                            to exam-

Table             5.       Agreement                 rates     between                     clinical           and     pathological                         definitions

                                                                                                                                                               Percent            Cortex           with         Interstitial               Infiltrate
                                                                                                             >.5a    ii    (%)b                                   lO         ii (%)b                                       l5              n (%)‘                                                   25           n (%)‘

              3            tubules          with       tubuliti&’
                       rejection                                                                             106(82%)                                                103(80%)                                                   92(71%)                                                               81(63%)
                       no rejection                                                                             58 (76%)                                              61 (80%)                                                  63 (83%)                                                             64 (84%)
                       overall                                                                               164(80%)                                                164(80%)                                               155(76%)                                                                145 (71%)
                       1 tubule           with        tubulitis
                       rejection                                                                             117(91%)                                                109(85%)                                                   97(75%)                                                               84 (65%)
                       no rejection                                                                           55 (72%)                                                59 (78%)                                                  63 (83%)                                                             64 (84%)
                       overall                                                                               172(84%)                                                168(82%)                                               160(78%)                                                                148(72%)
    a    Current               definition            of type I rejection.
    I)   Percentages                  based          on the number                        of clinical           rejections                (n    =    129).      the number                 of clinical               no rejection                      (ii        =       76),      and         total           (ii       =       205).
1936                         Journal            of the    American               Society        of Nephrology

Table           6.        Table           of clinical               severity               by pathology                        charactenisticsa

                                                                                                                                                                      Clinically                Severe
                                                                                                                                                                                                                                                                                         Odds      Ratio
                                                                                                                                                  No                                             Yes                                                                                    >(95%        Cfl’
                                                                                                                                       ii (Row                %)                      ii (Row                (7)

                 no                                                                                                                        30(61%)                                      19(39%)                                             49                                    -

                 type         I                                                                                                            19 (20%)                                     75 (80%)                                            94                                   6.2 (2.9,13.4)
                 type         II                                                                                                             I (4%)                                     24 (96%)                                            25                                37.9       (4.7,303.7)
                 missing                                                                                                                     1                                              2                                                    3
            tubular injuryc
                 none                                                                                                                      20(83%)                                       4(17%)                                             24                                   -

                 present                                                                                                                   30 (21%)                                   1 16 (79%)                                           146                                19.3       (6.2,60.8)
                 missing                                                                                                                     I                                           0                                                       1
                 none                                                                                                                      49(36%)                                      89(64%)                                        138                                       -
                 present                                                                                                                     1 (4%)                                     24 (96%)                                            25                                13.2       (1.7,100.7)
                 missing                                                                                                                     1                                              7                                                    8
            interstitial                hemorrhagec
                 none                                                                                                                      50(33%)                                    103(67%)                                         153
                 present                                                                                                                     0 (0%)                                     13 (100%)                                           13                                13.2       (07722609)d

                 missing                                                                                                                     I                                           4                                                    5
            interstitial                edemac
                 none                                                                                                                      24(75%)                                          8(25%)                                          32                                   -

                 present                                                                                                                   26 (19%)                                   1 12 (81%)                                           138                                12.9       (5.2,32.0)
                 missing                                                                                                                     1                                           0                                                       1
            tubulitis              in      10 serial              high-powered                      fieldsc
                 none                                                                                                                      17(71%)                                       7(29%)                                             24                                    -

                 1 or 2 tubules                                                                                                            1 1 (39%)                                    17 (61%)                                            28                                   3.8 (1.2,12.0)
                 3           tubules                                                                                                       22 (19%)                                     95 (81%)                                       1 17                                   10.5       (3.9,28.4)
                 missing                                                                                                                     1                                              1
            activated               lymphocytesc
                 none                                                                                                                      31 (53%)                                     27 (47%)                                            58                                   -

                 present                                                                                                                   19 (17%)                                     93 (83%)                                       1 12                                      5.6(2.8,11.5)
                 missing                                                                                                                     1                                           0                                                       1
                 none                                                                                                                      46(34%)                                      91 (66%)                                       137                                       -
                 present                                                                                                                     4 (12%)                                    29 (88%)                                        33                                       3.7 (1.2,1           1.1)
                 Illissing                                                                                                                   1                                           0                                                       1

      a    Hyperacute                   rejection,            fibrinoid           necrosis,          transmural                  arteritis,             thrombosis              (arteries              and         venous),           and            renal     infarction             are omitted            due   to
having           less       than         four       samples          with         the      characteristic                 present.               The      three       samples           that          could          not      be     assessed                for   pathological              rejection         have
been         included              in the table,               bringing            the sample                   size      to 17 1 . The                 data       in this      table           are      based        on center                  pathology               findings.          CI. confidence
interval:            -.    indicates              reference          point.
      h    Univariate             odds ratio.
      C    Denotes            significance    at the a                              0.05        level.
      (I   Adjusted              odds           ratio.

inc        whether            any         specific             pathologic                  characteristics                      were          related                schemata                   for       acute            rejection                  (Table           7).     Finkelstein                et al.       (5)
to endothelialitis.                              Only          glomerulitis                   (OR           =      4.4.         P      <         0.001)              proposed                    one          of      the          first         semiquantitative                           systems,           which
was         significantly                       associated                with          endothelialitis.                                                             grades           the         interstitial,                vascular,                     and      glomerular                lesions        with          a
                                                                                                                                                                     resulting                  acute          rejection                   index.             Their          analysis           demonstrated
Discussion                                                                                                                                                           that       the     pathologic                     features                  correlated               with        the    clinical          course
      The                 CCTT              classification                        system                 of            acute               rejection                 and      provided                   a solid             basis          for further                  refinement.              The        features
(“CCTT”)                     draws                on     the        insights               gained               from           several                 past          included               many              that         define           acute            rejection           in the         CCTT          system
                                                                                                                                                                                                                                                  Pathologic              Criteria         of Rejection                                1937

Table           7. Comparison                                   of CCTT                with           other           classification                     systems            for acute                     renal          allograft                rejectiona

                 CCTT                                                     Banff         ( 1)                                                          Mata       s (7)                                                             Banfi           (6)                                          Finkelstein               (5)

            Type               I                                       Borderline’                                                   Categories                  1 through               4                                   Mild                                                                 Mild
                                                                       Grade           IIAb
            Type               II                                      Grade           IIBC                                          Categories                 5 through                7
                                                                       Grade           III
              Type             III                                     Grade           III                                           Category              8                                                                 Moderate                                                             Moderateh
                                                                                                                                                                                                                             Severer                                                              Sever&
      a   CCTF,       Cooperative       Clinical      Trials     in Transplantation.                                                            Numbers               in parentheses       indicate    reference     number.
      h   Differ    in the degree        of tubulitis       and infiltration.        Does                                                      not require              tubular    injury.   activated     lymphocytes,      or edema,                                                            versus         CCTT             type        I.
      C   Differ    in the degree       of endothelialitis.
      d   Endothelialitis         not recognized         separately.
      C   Subset     of grade III with ‘severe”               endothelialitis.
      ‘   Subset    of grade III with fibrinoid     necrosis.
      g    Differ  in the extent of fibrinoid     necrosis.
      h    Also includes    glomerular   lesions,    not considered                                                                  in the other               systems.
      I   Includes    those with parenchymal       necrosis.

(interstitial                    infiltrate,                   edema,          tubulitis,                   and          vascular              necrosis)                   and III).                These             categories                   have           some         degree              of correspondence
and        some             that           do            not      (glomerular                       inflammation,                         endothelial                      with          CCTT                 and        differ         primarily                      in the         separation                 of the           major
swelling,              and               mural             inflammation                       of       arteries).               Banfi            and      col-             divisions                    (Table           7).       Banff            divides              grade            II into          A and            B on         the
leagues              (6)           had         four            categories              of          acute           rejection,              from          mild              basis         of         the presence    or absence     of “mild or moderate    intimal
with        interstitial                       infiltrates                without                  glomerular                   or       vascular               le-        antenitis.”                 We  regarded    the presence     or absence  of endotheli-
sions           to     irreversible                             with        morphologic                             alterations                  of      large             alitis        as         potentially                    more             fundamental.                       and           it therefore                  is the
vessels              and            extensive                    necrosis              and           infarction.                  Both           of     these              basis             of         separating                   type           I from             type II. Banff  puts “severe
early schemata                            had more                      categories                   at the severest       end of the                                      intimal                arteritis”                into       grade              III;       however,   we do not grade    the
spectrum    and                          emphasized                       thrombotic                    features,    reflecting   the                                      endothelialitis,                           in part           because                   the       number                of affected                   vessels
patterns           of pathology                              in that           era.         Neither                recognized                    endothe-                  is usually                     small,          and         the         distinction                 is too            subjective.                 Some             of
lialitis         as a distinct                           lesion.                                                                                                           the originators                            of the Banff                         schema              have             regrouped                 their        cat-
      Matas           et al. developed                                  a classification                         scheme                similar          to the             egonies                 in a way                 similar               to CCTF                  in a subsequent                         publication
CCTT,                although                      with         finer       gradations                      (7).         They           scored           eight             for       clinical                 correlations                    (grade              I with             hA          and       grade           IIB         with
features              of           acute            rejection               (which                  were           not        mutually                 exclu-              III)       (8).          Banff             includes               recent               focal            infarction              and           interstitial
sive):          categories                         I through                4, “minimal                          to severe               acute          tubu-              hemorrhage                          without               transmural                        artenitis           as     grade           III,      whereas
lointerstitial                      rejection,”                   corresponding                           to our           type         I; categories                      CCTT               requires                 an      artery             with           fibrinoid                necrosis              or transmural
5 through                   7, “minimal                           to     moderate                    acute           vascular              rejection,”                     inflammation.                              Another      difference      is that the CCTT    system
corresponding     to our type II; category     8, “severe                                                                         acute vascular                           does not have                              a borderline       category.    Instead, we have added
rejection”    with fibninoid   necrosis,   corresponding                                                                              to our type                          criteria               that         reflect          ongoing                   parenchymal                        injury           (edema,                tubu-
III. The outcome     at 1 yr correlated    with category.                                                                           Categories    1                        lar      injury)               or immunologic                                 activity            (activated                 lymphocytes)                         to
and        2 (n        =         41 of 204                      biopsies)              had          a 78%              I-yr       graft          survival.                 help          separate                   active           rejection                   from         inactive               infiltrates.                 CCTT
Category                   3        (n         =         28)       corresponding                            to      CCTF                type          I with               excludes                     the subcapsular                            area          for      scoring               (which           often            shows
infiltrates                 of        50           to      75%           had       a        1 -yr          graft           survival              of     71%;               mild          inflammation                           and fibrosis)                        and does                not score              tubulitis                in
category               4 (n                =        5)     had          a 1-yr          graft              survival               of      100%.            The             areas             of         tubular             atrophy                 (where                it is often                  seen        nonspecifi-
moderate                   acute               vascular                rejection             (ii      =      23       biopsies),                 in which                  cally).
“five or more                            vessels showed    a prominent      endothelialitis,”                                                            had                      The         CCTT                 scoring             system                meets            several             performance                        cnite-
a 1-yr survival                           of 61 % (not statistically   different      from                                                        survival                 ria      to be               accepted               as a widely                       applicable                 diagnostic                   classifica-
with        milder                  degrees                of endothelialitis).                                  Category                8 ( 14 biop-                      tion;         these                 include               reproducibility,                              ease         of      use,        sensitivity,
sies)       had        the          poorest                prognosis,               with            29%            graft       survival               at 1 yr.             specificity.                        validation                    of          criteria,             and           clinical              correlation,
Matas      et              al. did not use                                glomerular                       changes  to distinguish                                         which              are reviewed                           below.
categories                 of acute rejection,                               consistent                    with the CCTT     system.
   The Banff     working                                          schema     was the                               first effort   to get a                                 Reproducibility
multicenter  consensus                                          on allograft   grading                               and has achieved   its                                       The         interobserver                          reproducibility                         of the CCTF                        classification
goal to stimulate     progress   in standardization                                                                  ( 1 ). Banff has five                                 in       these               286        biopsies                 was           quite           satisfactory;                    in     9 1%            of     the
categories   for acute rejection     (borderline                                                                 and grades I, hA, IIB,                                    cases,                 the         panel            and          the          center             pathologist                    agreed               on       the
1938                             Journal           of the               American               Society               of Nephrology

presence                   or absence                             of rejection.                      As judged                      by          the     0.80          kappa             report            the       number                  of arteries                     sampled,                 since           the       false           negative
statistic,                 the          agreement                              was           excellent.                      The            agreement                        was        rate       is high               for         endothelialitis                          in a single                    core           (82%).
almost              as good                  for the type                         of rejection                        (kappa                =      0.72)             and      the
presence                  or absence                          of endothelialitis                                 (kappa                =     0.65).           It should                 Specificity
be pointed                       out      that           the         design                 of our              study            did        not        intrinsically                       The specificity                                  of the                 pathologic                  criteria              is difficult                     to de-
favor              agreement.                           in that                 the          scoring                  was           done              over           a long             termine                 because                  the          biopsy,             although                   not        infallible,                   is widely
period              of time              (3 yr)                   by a large                   group             of pathologists                             (n       =      1 1).      regarded                   as         the          “gold              standard”                  for         the          diagnosis                    of      acute
Thus,              the CCTT                        classification                              system                 appears                to be a statisti-                          rejection.                 However.                       even             when           rejection                was       defined                  solely           by
cally          robust              classification                               system.                  The          kappa                values            for          inter-        clinical                criteria              and             the          biopsy             was           interpreted                    without                   any
observer                   reproducibility                                 of         acute              rejection                  classification                          sys-        clinical                information,                           the          CCTT              classification                           performed                       re-
tems           for        the      liver               (9)         and         heart           ( 10)           are      much                lower            (kappa                =    markably                   well,             with             a sensitivity                     of      86%           and          a specificity                        of
0.40 to 0.55 and 0.51,       respectively),                                                                           which                some            have             said        72%.          A comparison                                    of the           clinical                and         pathologic                    diagnosis,
argues   for a revision (10).                                                                                                                                                           using         the         Banff              criteria,               yielded              an agreement                          rate       of 77%                  with
    The reproducibility  of the Banff                                                                     classification                          was        tested            by       a low             degree                of          specificity,                     i.e. , the              usual            discrepancy                           was
five         pathologists                          who              reviewed                       77      biopsies.                   with           “weighting”                       rejection                 on       biopsy                 with             a clinical                course             of nonrejection                              (2).
to       minimize                      the         contribution                               of        distinctions                        considered                       less       The        specificity                       depends                   on the pathologic                                  features              present:                In
important.                       such             as      normal                  versus                  borderline                       ( 1 1 ). The                kappa            one        series,              the         specificity                    of endothelialitis                             was          100%             (i.e.,         all
score           for        interobserver                                grading                of the             rejection                     was        0.40           over-         patients                with           endothelialitis                            had         a clinical                  diagnosis                    of      rejec-
all,         increasing                      to         0.56             for          the       presence                      or       absence                  of        acute         tion),        and           the specificity                             of tubulitis                    was         only           50%           (8).         Obvi-
rejection                  (normal                     and           borderline                         versus               grades               I through                   III)      ously,            in actual                      clinical              practice               both           the          pathologic                    features
( 1 1 ).      This             compares                       unfavorably                           with          the         CCTT                kappa              values.            and        clinical                features                   should              be considered                            to maximize                             diag-
It is obviously                            hazardous                           to equate                       kappa             values               between                two        nostic            accuracy.
studies.                  However,                           it     is     notable                      that         the         agreement                        rate         for             It is clear                 that            when             judged              by the clinical                         course,                a signif-
endothelialitis                           is equivalent                               in the             two          studies               (0.65          and         0.66).           icant         percentage                            of         patients               who             lack          clinical               evidence                     of
Because                   this         feature                is defined                     similarly                  by the              two        groups,               this       rejection                 (28%               in the             CCTT                series           and          38%         in a Banff                      study;
result             suggests                that          the            other          differences                         in agreement                         rates         are       reference                  2) had                  a biopsy                  that         met         the         pathologic                    criteria               for
intrinsic                 to the classification                                         schemes                  rather              than             the design                   of   rejection.                  In biopsies                         of normally                          functioning                       kidneys                     1 wk
the       studies                 or the               skills             of the              pathologists.                                                                             after        transplantation,                                  the         presence               of        tubulitis               correlated                      with
                                                                                                                                                                                        subsequent                      rejection                     episodes                (13).          It must               be considered                            pos-
Ea.ce           of         Use                                                                                                                                                          sible,        if not            likely,             that         these          discrepancies                         are       not        due         to a lack
        The          pathologists                             in this             study              found              that           the CCTT                       classi-           of      specificity                    of the                 biopsy           criteria,               but         that      subclinical                       rejec-
fication                  of      the         biopsies                         takes           no          more               time           than            the          usual         tion         is present.                         Approximately                               30%             of      stable             patients                   have
diagnostic                       examination                             of a transplant                              biopsy.                Little          has           to be        biopsy             features                   of         rejection                (tubulitis)                 and           have           an         increased

quantified:                       The             estimated                       percentage                          of cortex                       involvement                       loss         of         renal           function                      at       1 yr.            even              without               overt               clinical
with          the         infiltrate                   has         to exceed                   5%,             and         the      tubulitis                occasion-                  rejection                episodes                   (14).            Thus,          tubulitis                seems           to be a significant
ally         has      to be counted                                 to be certain                        that         at least             three         tubules               are      finding             in biopsies,                         regardless                  of whether                     a rejection                   episode                   is
affected.                  Nothing                      else             is graded:                     There              are       no         “mild,”               “mod-             clinically                 evident                  at that             time.
erate,”              or “severe”                         degrees                  of any                 lesions.                As a measure                             of the
efficiency.                      the       review                   panels              typically                    took          approximately                             4 to       Validation                         of Criteria
8      mm           per         biopsy.                                                                                                                                                        In transplant                         pathology,                       “a major                  problem                 in the            histologic
                                                                                                                                                                                        diagnosis                  is the             relative                weighting                   of the             respective                   criteria              of
Sensitivity                                                                                                                                                                             acute rejection,    such as interstitial  infiltrate and tubulitis”                                                                                                  (6).
   These                   studies                suggest                      that         the CCTT                         criteria              have            a sensi-             Therefore,     we attempted   to validate     the CCTT   thresholds                                                                                                   for
tivity         of 9 1 % for                        detection                     of rejection                         in a single                     core           biopsy.            diagnosis                  of         rejection.                     Maximum                     agreement                      with            the         clinical
This           predicts                  that           two              cores              have          a sensitivity                          of 99%,                   quite        course             was          found               with             a threshold                  of 5%                     infiltrate;                 progres-
satisfactory                      for clinical                           management.                            Others               have             published                 on      sively            less       agreement                         was         found           with         higherthresholds                                    (10,       15,
the       concordance                             rate             by      grade               in paired                   renal            transplant                     biop-        or       25%).             This             was            due         to decreased                          sensitivity                   for         rejection
sies,         using              the       Banff                   schema                   (12).          Although                    no         direct           data         on      with          less             change                    in      specificity.                     Banff              has           a much                     higher
the pathologic                               sensitivity     for the diagnosis                                                          of rejection                           are      threshold                  of infiltrate                       for grade                  I rejection                     (25%          to distinguish
provided.     the                          study       found    that a clinical                                                        decision     to                      treat       from          borderline).                             Our           results              indicate                 that       the          infiltrate                  in-
rejection                      would               be             missed                in          10.5%               of         single              cores,              quite        volves              <25%                    of       the         cortex              in       3 1%           of       the          cases          of          type           I
compatible                        with             the             CCTT                data.            In the             CCTT                  biopsies,                  88%         rejection.                 Thus,              CCTT                   would            classify               many            of the Banff                           bor-
showed                 concordance                                  in all            cores              for      type             of rejection.                       In the           derlines                as a type                   I rejection.                    In fact,                it has been                 noted               that       the
Sorof study ( 12). concordance                                                                 of        Banff             grades               between                    cores        Banff             borderline                       category                  behaves                 like         rejection                in a substan-
was found in 29% of paired                                                              samples.                     Thus,             the CCTT                      system             tial      percentage                        of the              cases           (37%)                (2).
is at least                     as sensitive                             and          far      better             in concordance                                  between                      The        originally                       defined                 CCTT              threshold                for tubulitis                         is prob-
cores.               Clearly,                 it is important                                      to      indicate                    on        the       pathology                    ably         appropriate                         at 3                tubules            per          10 hpf.              When           the          threshold
                                                                                                                                                                                                                                Pathologic              Criteria         of      Rejection                                 1939

for        tubulitis              was           raised            beyond                3 tubules                  per          10 hpf,               the     biopsies,               with            acute          cellular               rejection                  in two               series              (18,23),
sensitivity                 declined                  steeply,            similar             to     the     result             with          the     in-     somewhat                    higher            than          the      present              study,            in which                     27%           of the
terstitial              infiltrate.             Lowering                   the      threshold                for      tubulitis               to      >1      biopsies             with             acute         rejection              had endothelialitis.                               Endothelialitis
tubule per 10 hpf improved                                              sensitivity, as judged                            by correlation                      typically              affects            the        small          arteries              but        can        also         extend               into         the
with clinical course   (from                                             82 to 91%)     partly                           balanced      by a                   arterioles                 (22).         Previous                 studies            have            suggested                     that         endothe-
decrease      in specificity      (76 to 72%).   Overall,                                                           the clinical   cor-                       lialitis          may            have         prognostic                    significance                      but        have             not        always
relation    improved        from 80 to 84%. However,                                                               a small improve-                           been         conclusive.                      Typically,                   the early                 studies             did           not       separate
ment may not be clinically             useful, considering                                                             the hazards    of                      endothelialitis                          from          fibrinoid                   necrosis                 and            medial                 damage
overdiagnosing         rejection.                                                                                                                             (5,6,24).             OKT3                treatment                  reversed                   acute           cellular                rejection                 in
       We       included                 the       lesions              of activated                   lymphocytes,                       edema,              patients              without                 endothelialitis                        more               often            than             those             with
and tubular     injury to minimize                                           the potential                    for false positives                             endarteritis                     in two studies:                           75%    versus                        62%           (20)             and          91%
from incidental       infiltrates that                                       may be seen                      in renal allografts                             versus            86%            (23). One-year                           graft survival                        was          not         affected              by
(15-17).                 In this            biopsy               series,          eliminating                      these          criteria            re-     endothelialitis                        in the largest                     series       (61%              versus            64%            in biopsies
sulted    in two additional                                    specimens                being              scored   as pathologic                             without          endothelialitis)                            (7), but                  was in another     study    (58%
rejection    (only one had                                    rejection             clinically).               This suggests      that                        versus         75% without                           endothelialitis)                    (23). These differences      did
these         criteria            are       not        necessary                 for biopsies                 taken             either         at the         not        reach statistical                          significance.                    In children,   the presence      of
time         of suspected                   rejection               or during                 a rejection             episode,                 which          arteritis            was           associated                 with           a high              rate      of graft                    loss       (6 of 6)
represent the vast majority    of the samples    in this study.                                                                                 How-          despite             the use of steroids                             and OKT3                      (25). A subsequent                                     study
ever, we believe   that the criteria  should  remain,    because                                                                                 they         addresses                  the        question              of whether                    the        degree            of endothelialitis
may          be valuable                  in the protocol                         scheduled/surveillance                                      biopsy          (number                 of         vessels,             number                 of         cells)            affects                outcome                    (V.
setting.           Requiring                      activated                lymphocytes                       and         edema                should          Nickeleit,                 J. Poletti,               E. Vamvakas,                          and          R. Colvin,                     unpublished
also help distinguish     CsA toxicity   from                                                        rejection.   Sibley                           et al.     data).
noted that edema      and “large lymphocytes                                                            with vesicular                         nuclei            The data did not permit    a breakdown                                                               of steroid                      responsive-
and        small          nucleoli,”                  which             we would                term         “activated                  lympho-              ness per Se. However.      a subsequent                                                              study.    using                      the CCTT
cytes,”          are         more           characteristic                       of acute             rejection             than          of CsA              criteria,            has           demonstrated                       that         type          I rejection                    is more                  often
toxicity           (I 8). Further   evaluation                                    of the diagnostic                             contribution                  completely                    steroid-responsive                              than          type           II or type                   III rejection
of these            features   is warranted.                                                                                                                  (V. Nickeleit,                         J. Poletti,                E. Vamvakas,                          and         R. Colvin,                       unpub-
    How            well these criteria       will                                distinguish                 ATN           deserves                  fur-     lished            data).           Response                  to therapy                    (steroids                or antibodies)                            has
ther    study.   A                      mononuclear        infiltrate                                 and tubulitis                        can be             been         correlated                   with          the         Banff             grade:               A complete                           response
present     in ATN,                      but are generally        less                             prominent     than                     in acute            occurred               in 100%                  of borderline,                       93%           of grade                I. 79%               of grade
cellular           rejection.                  In the            studies          of ATN                   in native              kidneys              by     II, and 47%                      of grade              III (26)             cases.          The          Banff           grades                II and           III
Marcussen  and colleagues,                                          the lymphocyte                          interstitial               infiltrate             are heterogeneous                                with         regard               to endothelialitis,                                 preventing                      a
had a volume    percentage                                            of 8.2%,     and                       tubulitis                 was less               direct        comparison                      with the CCTT       results.    However,                                                          the vas-
extensive    than in acute rejection  (19).                                                        Similarly,              little tubulitis                   cular        score was                    higher   in the steroid     nonresponders.                                                            which   is
was detected     in ATN in renal allografts                                                          in another               study (mean                     compatible                    with            the CCTT                    results.
of 0.04            on a scale                    of 0 to 3) (1 7). However,                                         interpretation                       of         When             individual                    histologic                    features               among                all        cases             with
biopsies                from          grafts           with       primary               nonfunction                      could           be     diffi-        rejection             or nonrejection                             were         examined                    for correlation                             with            a
cult         if the            infiltrate               and       tubulitis              are         not      extensive.                      In that         clinically                severe           rejection.                all of the                 features              we        used            to define
setting,            more     stringent     criteria                                 may be required                                (e.g. , also               rejection             appeared                   significant,                  i.e. , tubulitis,                     interstitial                    edema,
requiring             activated        lymphocytes                                     and attempting                                to judge                 tubular             injury,            activated              lymphocytes,                           and        interstitial                   infiltrate.
whether             the tubular                       injury        is proportionate                         to the             infiltrate).                  This         is      to          be      expected,                   because                 a       rejection                  needed                 to       be
                                                                                                                                                              diagnosed                    and        treated             to receive                 OKT3/ATG                            (one           of the en-
Clinical                 Correlation                                                                                                                          tenia for severe  rejection).
   In the study presented    here, we sought      correlations     between                                                                                        The two features       that                                     were            not          part         of       the         definition                   of
the type of rejection   and the clinical severity     of the rejection.    A                                                                                  rejection             that            correlated              with           a clinically                  severe              rejection                 were
clinically                severe            rejection               had          been         predefined                   by      the        CCTT            glomerulitis                       (OR          =      3.7)         and        interstitial                     hemorrhage                        (OR              =

protocols                 as one               that        was          steroid-resistant,                         was          treated             with      13.2),            consistent                    with          previous                 studies.                 Glomerulitis                           is      re-
OKT3   or ATG,                           or occurred                   in the first 10 d after                                  transplanta-                  garded            either as a T cell (27,28)  or antibody-mediated                                                                                (29,30)
tion. We found                             that the                 type II rejection    was                                    more       often              lesion.           Glomerulitis  has been associated     with graft                                                                      loss      in most
clinically                severe            than           type          I (OR            =        6. 1). We               did         not          have      (27-3         1),      although                  not        all     (32),           series.             Interstitial                    hemorrhage
enough             cases of type III rejection       to confirm     the already      well                                                                     has        long       been             associated                  with        graft          loss         (6,7,33,34).                        Thus,           the
known              adverse  prognostic    significance        of fibninoid     necrosis                                                                       presence                or         absence             of         these        features                  should              be          specifically
(6,7,20,21).                                                                                                                                                  noted         in biopsies.                      The         interobserver                        reproducibility                              of scoring
       We       have             separated                  biopsies               with            endothelialitis                       in small             milder glomerulitis                                 in other              series          is low            (kappa                 =      0.26)             ( 1 1),
arteries           or      arterioles                 (endarteritis),                   because              it is a common                          and      and some argue                             it should                not       be       included                  (34).          However,                      our
diagnostically        significant                                  manifestation                        of cellular                     rejection             data,        and           that        of others,                 support              the        inclusion                  of hemorrhage
(22). Endothelialitis           has                               been    reported                     in 48%       of                 the renal              and glomenulitis                           as markers                     of more                severe             rejection.                  Whether
1940                        Journal           of the American                     Society            of Nephrology

they         should               be         incorporated                        into          the      classification                    scheme                 2.   Dooper          MM.            Hoitsma             AJ. Koene               RA, Bogman                      Mi: Evaluation
more          formally                 remains                  to be determined.                          Clearly,           it would                be              of the Banff criteria      for the histological     diagnosis  of rejection      in
useful     to identify  prognostic                                              features     to subdivide                            the       large                  renal allograft  biopsies.      Transplant      Proc 27: 1005-1006,         1995

percentage      of cases with type                                              I rejection.                                                                     3.   Fleiss        JL:      Statistical               Methods           for     Rates           and       Proportions,                 New
                                                                                                                                                                      York,        John          Wiley         & Sons,              1981
    Two standard                             features  did not correlate      with severity:                                                      the
                                                                                                                                                                 4.   Hosmer            DW         Jr. Lemeshow                   S: Applied                Logistic           Regression,              New
extent  of tubulitis                           and the extent of the infiltrate    (among                                                      those
                                                                                                                                                                      York,        John          Wiley         & Sons,              1989
cases         with          rejection).                  Matas            et al. also                concluded               that        a severe
                                                                                                                                                                 5.   Finkelstein                FO.       Siegel        NJ,     Bastl         C, Forrest               JNJ.      Kashganian               M:
tubulointerstitial                        infiltrate  alone (>75%                                      of the cortex) does not                                        Kidney    transplant      biopsies                         in the diagnosis   and management                                          of
predict      poorer                    graft survival      compared                                    with those with lesser                                         acute rejection      reactions.                          Kidney   hit 10: 171-178,   1976
degrees     of               infiltrate     (7).                     Other             studies    have reached                          similar                  6.   Banfi        G, Imbasciati                       E, Tarantino                  A, Ponticelli                 C: Prognostic
conclusions                   (6,20,32,35,36)                           (V.            Nickeleit,    J. Poletti,                      E. Vam-                         value        of renal          biopsy          in acute          rejection             of kidney           transplantation.
vakas,         and          R. Colvin,                    unpublished                       data).        Therefore,                we      see       no              Nephron              28: 222-226,                  1981
value         in grading                    the degree                   of tubulointerstitial                         infiltrate             or the             7.   Matas         AJ,          Sibley        R, Mauer              M,        Sutherland                 DE,       Simmons               RL,
tubulitis            beyond                 the     threshold                  needed              to classify           as rejection.                    It          Najarian             JS:      The        value         of needle               renal         allograft          biopsy.            I. A
is possible                 that disruption                          of the tubular                   basement             membrane                    in             retrospective                 study        of biopsies               performed               during          putative            rejec-
                                                                                                                                                                      tion       episodes.             Ann      Surg         197:      226-237.                 1983
association                  with           tubulitis                may         have          prognostic                implications.
                                                                                                                                                                 8.   Solez        K, Hansen                  HE,       Kornerup               Hi.         Madsen           S. Sorensen                 AW,
    In summary.                         the CCTF      system                                 described    here has certain
                                                                                                                                                                      Pedersen EB, Marcussen     N, Benediktsson                                                  H, Racusen                 LC, Olsen
objective   and                        major  advantages                                    over existing     classification
                                                                                                                                                                      5: Clinical validation and reproducibility                                                of the Banff                schema for
schemes               for      acute              renal          allograft               rejection,             notably             simplicity
                                                                                                                                                                      renal       allograft            pathology.              Transplant              Proc         27:      1009-101            1 , 1995
and reproducibility,                                while            retaining               sensitivity,            specificity,                   and         9.    Demetris Al, Seaberg  EC,                                      Batts KP, Ferrell LD, Ludwig       J,
clinically             relevant                   prognostic                   implications.                  Ideally,           one        classi-                   Markin RS, Belle SH, Detre                                     K: Reliability and predictive value
fication             system                 would               be       universally                   accepted            based              on      its             of the National                   Institute      of Diabetes      and Digestive                                   and Kidney
merits          and           used            by          all        transplantation                      centers.            The           CCTT                      Diseases   Liver                 Transplantation         Database     and grading                                   system  for
system          should               promote                    further          discussion                and      studies            that        will               cellular        rejection               of liver          allografts.                Hepatologv               21 : 408-416,
lead        to this          goal.                                                                                                                                     1995
                                                                                                                                                               10.    Neilsen H, Sorensen                            F. Neilsen            B, Bagger     i. Thayssen     P. Baan-
Addendum:                     In March                  1997, the Banff                       categories           for acute             rejection
                                                                                                                                                                      drup U: Reproducibility                            of the            acute rejection     diagnosis    in hu-
were        changed            to substantially                          correspond                  to the types          of rejection                in
                                                                                                                                                                      man cardiac                 allografts:            The Stanford classification                                  and the inter-
the CCT1’ system                            (borderline   is now termed                                 suspicious            for rejection:
                                                                                                                                                                      national        grading              systems.        J Heart Lung Transplant                                    12: 239-243,
K. Solez, personal                           communication).
                                                                                                                                                               1 1.   Marcussen       N. Olsen TS, Benediktsson           H, Racusen    L, Solez K:
Acknowledgments                                                                                                                                                       Reproducibility       of the Banff   classification        of renal  allograft
      We acknowledge                              Dr. Lisa Beaudet                       for her assistance,                  Susan           Quinn                   pathology:                 Inter-        and       intraobserver                     variation.           Transplantation
for     preparing             all      of    the slides                for review,              the clinical             coordinators                 for             60: 1083-1089, 1995
their       assistance,              and Kate                   Tomlin    for               preparation            of the manuscript.                          12.    Sorof        JM,        Vartanian                RK,      Olson          JL.     Tomlanovich                     Si,     Vincenti
This        work was                supported                   by National                  Institutes          of Health  (National                                 FG, Amend      WI: Histopathological concordance                                                            of paired             renal
Institute           of Allergy                    and      Infectious                  Diseases)           Grants          UOl-AI-3                1440               allograft biopsy cores: Effect on the diagnosis                                                          and management
(Francis            L. Delmonico,                         M.D.,           Principal             Investigator,              Massachusetts                              of acute            rejection.            Transplantation                      60:      1215-12           19,     1995
General          Hospital),                  U0l-AI-31457                        (Arnold              G. Diethelm,               M.D.,             Prin-       13.    Kooijmans-Coutinho                          MF, Bruijn                    IA,         Hermans I, Schindler                            R,
cipal        Investigator.                   University                   of     Alabama,                Birmingham).                    UOl-AI-                      Frei U, Schrama                        E, van Es LA,                     Daha         MR. van der Woude                              FJ:
31492         (J. Thomas                    Rosenthal,                 M.D/Alan                    Wilkinson,            M.D.,           Principal                    Evaluation              by histology,                  immunohistology                        and PCR            of protocol-
Investigator,                University                   of California,                     Los      Angeles),            U0 1-AI-3               1445               lized       renal       biopsies               I week         posttransplant                   in relation              to subse-
(Richard          Thistlethwaite,                     Jr., M.D.,                 Ph.D.,   Principal                Investigator,                Uni-                  quent        rejection              episodes.          Nephrol            Dial         Transplant               10: 847-854,
versity         of Chicago),                        UOl-AI-31449                      (J. Wesley                   Alexander,                  M.D.,                  1995
Principal             Investigator,                       University                   of      Cincinnati),                UOl -AI-3               1490        14.    Rush        DN,        Jeffery           JR.      Gough          I: Sequential                    protocol            biopsies        in
(Lawrence               Hunsicker,                      M.D.,             Principal             Investigator,                University                of             renal       transplant               patients.         Clinico-pathological                           correlations               using
Iowa),        UOl-AI-3l559                          (Arthur              J. Matas,             M.D.,       Principal           Investigator,                          the Banff schema. Transplantation       59: 5 1 1-5 14, 1995
University             of      Minnesota),                       and       U01-AI-31442                     (Jimmy            Light.          M.D.,            15.    Burdick  IF, Beschorner    WE, Smith Wi, McGraw            Di,                                                             Bender
Principal            Investigator,                   Washington                    Hospital             Center).                                                      WL. Williams     GM, Solez      K: Characteristics    of early                                                              routine
                                                                                                                                                                      renal allograft    biopsies. Transplantation                                            38: 679-684,                      1984
References                                                                                                                                                     16.    McWhinnie       DL, Thompson     IF, Taylor                                           HM, Chapman                      JR. Bolton
  I.       Solez       K, Axelsen                       RA, Benediktsson                           H, Burdick             iF, Cohen                AH,                EM, Carter NP, Wood RF, Morris P1: Morphometric                                                                       analysis        of
          Colvin            RB,      Croker             BP, Droz                D, Dunnill              MS.       Halloran           PF, Hayry                        cellular infiltration assessed by monoclonal antibody                                                                 labeling        in
           P. Jennette               JC, Keown                       PA, Marcussen                      N, Mihatsch                 Mi, Moro-                         sequential              human             renal        allograft           biopsies.                Transplantation                       2:
          zumi K, Myers                           BD. Nast CC,                    Olsen 5, Racusen   LS, Ramos   EL,                                                  352-358,              1986
          Rosen S. Sachs                           DH, Salomon                     DR. Sanfilippo  F, Verani  R, von                                           I 7.   Solez        K, Racusen                 LC, Marcussen                    N, Slatnik               I, Keown             P. Burdick
          Willebrand                   E. Yamaguchi                            Y: International                    standardization                     of             IF, Olsen             5: Morphology                      of ischemic                 acute        renal      failure,       normal
          nomenclature         and                      criteria for               the histologic        diagnosis                     of renal                       function,            and cyclosporine                      toxicity            in cyclosporine-treated                            renal
          allograft    rejection:                        The Banff                  working     classification                       of kidney                        allograft           recipients.   Kidney                    Jot 43:            1058-1067.         1993
          transplant                pathology.                  Kidney           In!     44:       41 1-422,           1993                                    18.    Sibley        RK.           Rynasiewicz                  I, Ferguson                  RM.         Fryd       D, Sutherland
                                                                                                                                                                                                  Pathologic            Criteria       of Rejection                                1941

      DE, Simmons        RL, Najarian       IS: Morphology      of cyclosporine                                                                   cells     in acute          allograft           glomerulopathy.                      Am        J Pathol           129:         1 19-
      nephrotoxicity    and acute rejection     in patients immunosuppressed                                                                      132, 1987
      with cyclosporine     and prednisone.       Surgery   94: 225-234,    1983                                                            28.   Hiki      Y, Leong              AY,        Mathew           TH,        Seymour             AE,        Pascoe        V. Woo-
19.   Marcussen      N, Lai R, Olsen TS, Solez K: Morphometric                  and                                                               droffe        Al:      Typing          of intraglomerular                        mononuclear                cells            associ-
      immunohistochemical                         investigation                    of renal          biopsies         from        pa-             ated     with       transplant             glomerular              rejection.          Clin Nephrol                 26: 244-
      tients      with     transplant           ATN,         native          ATN,         or acute          graft    rejection.                   249, 1986
      Transplant Proc 28: 470-476,  1996                                                                                                    29.   Maryniak             R, First          RM,       Weiss            MA:        Transplant               glomerulopathy:
20.   Visscher  D, Carey J, Oh H, Turza  N, Kupin                                                      W,        Venkat         KK,               Evolution            of morphologically                           distinct         changes.           Transplantation
      Zarbo        R: Histologic                and immunophenotypic                                 evaluation             of pre-               41:      262-264,            1985
      treatment           renal      biopsies             in OKT3-treated                       allograft           rejections.             30.   Trpkov          K,        Campbell              P.     Pazderka               F,     Cockfield              5,      Solez              K,
      Transplantation                51: 1023-1028,                         1991                                                                  Halloran            PF: Pathologic                features           of acute          renal      allograft         rejection
21.   Burke        GW,      Cirocco             R, Markou               M, Viciana               A, Ruiz            P, Allouch                    associated             with         donor-specific                    antibody:          Analysis   using the
      M, Esquenazi                 V. Roth              D, Nery              J, Miller           J: Acute            graft       loss
                                                                                                                                                  Banff        grading            schema.          Transplantation                     61: 1586-1592.     1996
      secondary           to necrotizing                vasculitis:               Evidence         for cytokine-medi-
                                                                                                                                            31.   Axelsen             RA,      Seymour             AE,        Mathew               TH,       Canny          A,      Pascoe               V:
      ated       Shwartzman             reaction               in clinical                kidney           transplantation.
                                                                                                                                                  Glomerular                transplant            rejection:             A distinctive                  pattern           of early
      Transplantation                59: 1 100-1               104, 1995
                                                                                                                                                  graft       damage.          Clin       Nephrol             23:       1-1 1 , 1985
22.   Colvin        RB:      The        renal         allografts            biopsy.          Kidney         mt      50:      1069-
                                                                                                                                            32.   Olsen        5, Spencer               E, Cockfield                 S. Marcussen                  N, Solez           K: Endo-
      1082,        1996
                                                                                                                                                  capillary           glomerulitis              in the renal              allograft.             Transplantation                     59:
23.   Schroeder           TI,     Weiss         MA,          Smith           RD,    Stephens      GW, First                    MR:
                                                                                                                                                  1421-1425,                 1995
      The efficacy              of OKT3           in vascular                rejection.     Transplantation                     5 1:
                                                                                                                                            33.   Hsu       AC,       Arbus        GS,       Noriega           E, Huber              I: Renal           allograft           biopsy:
      312-315,            1991
                                                                                                                                                  A satisfactory                  adjunct          for    predicting                 renal       function           after         graft
24.   Magil        A, Rubin          I, Ladewig               L, Johnson                  M, Goldstein               MB,       Bear
      RA: Renal biopsy                      in acute allograft  rejection:   Significance      of                                                 rejection.           C/in Nephro/                5: 260-265,                  1976
      moderate vascular                    lesions in long-term    graft survival.     Nephron
                                                                                                                                            34.   Kasiske          BL,        Kalil      RS,       Lee        HS,      Rao         KV:       Histopathologic                      find-

      26: 180-183,                1980                                                                                                            ings        associated              with        a chronic,               progressive                decline             in      renal
25.   Corey        HE,      Greenstein                 SM,         Tellis         V,     Schechner              R, Greifer             I,         allograft           function.          Kidney          mt     40:       5 14-524,              1991
      Bennett B: Renal allograft                          rejection in children and young adults:                                           35.   Herbertson                BM,       Evans         DB,        Calne           RY,       Banerjee            AK:          Percuta-
      The Banif classification.                          Pediatr Nephrol 9: 309-312,     1995                                                     neous         needle         biopsies           of renal             allografts:           The        relationship                 be-
26.   Gaber        LW,       Moore         LW,           Alloway             RR,        Flax     S. Gaber            AO:        Cor-              tween        morphological                   changes          present              in the biopsies                and         subse-
      relation           between           Banff           classification,                   acute         renal          rejection               quent        allograft          function.            Histopathologv                    I : 161-178,                1977
      scores,       and reversal  of rejection.                        Transplant               Proc 27:            1019, 1995              36.   Kiaer        H, Hansen               HE,      Olsen          S: The          predictive            value          of percuta-
27.   Tuazon         TV, Schneeberger        EE,                       Bhan AK,                 McCluskey              RT, Co-                    neous        biopsies           from        human           renal       allografts             with      early       impaired
      simi       AB,      Schooley              RT,       Rubin             RH,        Colvin        RB:      Mononuclear                         function.           C/in Nephro/                 13: 58-63,                  1980

To top