Original Articles
                                                         In 1979 Birch and Fairley using phase
URINARY RED CELL                                     contrast microscopy demonstrated that ab-
                                                     normal morphology of red cells in urine can
                                                     differentiate glomerular from nonglomeru-
SITE OF HEMATURIA                                    lar hematuria in adults(l). According to
                                                     them, red cells coming through the glomer-
Kumud Mehta                                          ulus were dysmorphic whereas, those from
Deepak Tirthani                                      the lower urinary tract were eumorphic.
Uma Ali                                              Subsequently, usefulness and sensitivity of
                                                     this method in children was confirmed by
                                                     several workers(2-4). Later on examination
                                                     of urinary sediment stained by Wright's
ABSTRACT                                             stain under light microscope was reported to
    We studied the urinary RBC morphology in         be as good as phase contrast microscopy
87 consecutive cases of significant hematuria by     with the added advantages of easy avail-
3 commonly used methods: (a) light microscopy        ability, being less expensive as well as pro-
of the unstained urinary sediment; (b) phase         viding permanent record(5-7). There are no
contrast microscopy of the unstained urinary
                                                     studies comparing these three methods
sediment; and (c) Wright's staining of the
urinary sediment, in order to compare the sensi-
                                                     simultaneously in children. Hence we de-
tivity of these methods in detecting dysmorphic      signed this study to assess the sensitivity of
RBCs and thus predicting the site of hematuria.      (i) phase contrast microscopy (PCM), (ii)
The clinical data and the relevant investigations    Wright's staining of urinary sediment (WS),
were made available after the morphology of          and (iii) simple light microscopy (LM) of
RBCs in the urine was identified. Out of the 87      unstained sediment as a preliminary test to
patients, 45 had a glomerular and 42 had a non-      localize the site of hematuria and to guide
glomerular cause o hematuria.                        further line of investigations.
     Phase contrast microscopy showed a sensi-
tivity of 91.1%, Wright's stain of 82.2% and         Material and Methods
light microscopy of 66.7% in detecting a glome-
rular source of hematuria. Nonglomerular he-             Urinary red cell morphology of 87 con-
maturia could be detected in 92.9% cases by
each of the 3 methods.
                                                     From the Nephrology Division and Research
    It is concluded that phase contrast micro-          Department, Bai Jerbai Wadia Hospital for
scopy is most sensitive for the detection of dys-       Children and Research Centre, Parel,
morphic RBCs in the urine, Wright's stain near-         Bombay 400 012.
ly as sensitive whilst light microscopy of the un-   Reprint requests: Dr. Kumud P. Mehta, Hono-
stained sediment is least sensitive. Urinary RBC        rary Professor of Pediatrics, Chief, Neph-
morphology is a useful adjunct in the diagnosis         rology Division, Bai Jerbai Wadia Hospital
of hematuria and saves the patients from un-            for Children and Research Centre, Parel,
necessary investigations.                               Bombay 400 012.
Key words: Hematuria,      Glomerular,    Micro-     Received for publication: January 13, 1993;.
           scopy.                                    Accepted: May 2, 1994
MEHTA ET AL.                                           RBC MORPHOLOGY & HEMATURIA SHE

secutive children with significant hema-       resembled circulating RBCs in peripheral
turia(8) were studied blindly by light         blood indicated non glomerular hematuria
microscopy of unstained standard urine         (Fig. 2).
sediment (LM), light microscopy of sedi-
                                                   In each case, a final diagnosis was made
ment using Wright's stain (WS) by modifi-
                                               by standard clinical, biochemical, bacterio-
cation of method of Chang(5) and phase
                                               logical, immunological, radiological and
contrast microscopy (PCM) of unstained
                                               histopathological investigations and corre-
                                               lated with the type of hematuria. Twelve
    Fresh urine samples with specific gravi-   cases of acute glomerulonephritis (AGN)
ty (SG) 1010 were examined within 60-90        were followed up with serial urine examina-
min of voiding. A minimum of 100 RBCs          tions for 4 weeks to detect changes in uri-
were examined by each technique. If more       nary erythrocyte count and percentage of
than 20% of red cells showed dysmorphism,      dysmorphic red cells in urine. The specifici-
glomerular basis of hematuria was suspect-     ty and sensitivity of the three methods were
ed (Fig. 1). Eumorphic red cells which         compared statistically.
INDIAN PEDIATRICS                                                 VOLUME 31-SEPTEMBER 1994

                                                detected by each method were compared
    Out of 87 children between 2 days-15
                                                statistically (Tables I & II) and it was found
 ears of age with significant hematuria, 50
                                                that PCM had a sensitivity of 91.2%; WS of
 were boys and 37 girls. In 45, glomerular
                                                82.3% and LM of 66.6% in diagnosis of
 bleeding was diagnosed by clinical bio-
                                                glomerular bleeding. Nonglomerular hema-
 chemical and immunological parameters
                                                turia was detected in 39/42 (93%) cases by
 and in 13, renal histopathology was avail-
                                                each method.
 able for diagnosis (Table I); 42 had nonglo-
 merular causes as diagnosed by standard            Using standard error of difference
 clinical, biochemical, bacteriological and     between two proportions, it was noted that
 radiological investigations (Table II). To     there was no significant difference between
 evaluate efficacy of each method, number       percentage of dysmorphism obtained by
 of glomerular (dysmorphic RBCs) and non-       PCM and WS (76 ± 17.5 vs 65 ± 16) but
 glomerular (eumorphic RBCs) hematuria          there were significant differences between

MEHTA ET AL.                                                     RBC MORPHOLOGY & HEMATURIA SITE

                  TABLE I-Correlation of Urinary Dysmorphic ROCs with Glomerular Disease.

Etiology                                               No.                    Dysmorphic RBCs by

                                                                     PCM                  WS        LM

Acute glomerulonephritis                               30                28               25         20
Lupus nephritis                                        5-                4                4             3
Hemolytic uremic syndrome                               2                1                1             1
Henoch Schonlein Purpura                               2-                2                2             1
Membranoproliferative                                  2-                2                1             1
Chronic glnmerulnn'pbriti, IgA                         4-                4                4             4
Mesangioproliferative glomerulonephritis             (1 each)
Benign hematuria (normal biopsy)
Total                                                 45               41               37            30
                                                                    (91.2%)          (82.8% )      (66.7%)
* Biopsied.
  High statistical significance when PCM compared with LM p < 0.001
  High statistical significance when PCM compared with WS p < 0.001
  High statistical significance when WS compared with LM p < 0.001

                   TABLE II-Correlation of Nonglomerular Diseases with Eumorphic ROCs.
Etiology                                       No.                  Eumorphic RBCs in urine using

                                                                     PCM             WS            LM

UTI                                             18                  17               16             17

Postsurgical                                    8                    8                8             8
Calculi                                         6                    5                6             5
Hypercalciuria                                  3                    2                2             3
Renal tuberculosis                              1                    1                1             0
Leukemia, Scurvy, Hemophilia,                   6                    6                6             6
Pyonephrosis, Cyclophosphamide              (one each)
induced viral infection
Total                                           42                  39               39             39
                                                                   (93%)            (93%)       (93%)

results obtained when PCM and WS were                       by LM only in 9/45 cases of glomerular
compared with LM (49 ± 20) (p <0.001).                      hematuria. In 13/25 cases there were no
RBC casts were seen by PCM in 20/45 and                     casts/proteinuria noted.
INDIAN PEDIATRICS                                                     VOLUME 31-SEPTEMBER 1994

False Negative Cases                            ferentiating glomerular from nonglomerular
                                                hematuria (Table III).
   Of the false negative cases 4 were by
PCM, 8 by WS and 15 by LM. Out of               Discussion
30 patients of acute glomerulonephritis,            There are a few reports of RBC dysmor-
eumorphic RBCs were seen in 2 by PCM in         phism used for detecting the site of hema-
5 by WS and in 10 cases by LM. One case         turia in pediatric patients(2,3,4,7,ll), but
each of lupus and hemolytic uremic syn-         none from India. In our country with limited
drome (HUS) revealed eumorphic RBCs by          resources and lack of facilities for advanced
PCM, 3 additional cases by WS and 5 by          investigations it is imperative to have a sim-
LM with false -ve results were noted            ple test to decide which case ;of hematuria
(Table I).                                      should be subjected to the elaborate, expen-
                                                sive and invasive tests. From our study we
   False Positive Cases: There were 3 false     could predict the site of hematuria in 92%
positive cases by each method. Non glomer-      patients using PCM, 88% using Wright's
ular      disorders     with      significant   stain and 80% using plain light microscopy,
dysmorphism were noted in cases of              but there was greater variation in the sensi-
calculi, urinary tract infection (UTI),         tivity of detection of glomerular hematuria
idiopathic    hypercalciuria and renal          (91%) by PCM, 82% by WS and 67% by
tuberculosis, which is also reported by         light microscopy).
other authors(2,5,10). Progression of
                                                    Twenty five of the forty five patients
dysmorphism was seen in 10/12 cases of
                                                with glomerular disease did not show casts
acute glomerulonephritis on weekly follow
                                                and in 13 out of these 25, there was no pro-
up over a 4 week period. The data was ana-
                                                teinuria. Hence, 13/45 patients with glomer-
lyzed to decide the predictive value of the
                                                ular disease might have been subjected to
percentage of dysmorphic RBCs for diag-
                                                urologic/radiologic procedures such as IVP/
nosis of glomerular hematuria. Table III
                                                cystoscopy. However, on basis of red cell
compares the yield of 20%, 40% and 80%
                                                morphology it is possible to channelize this
of dysmorphic RBCs by PCM. Specificity
                                                group towards specific investigations and if
was 100% when 40% and 80% were taken
                                                required renal biopsy; otherwise cases of
as cut off points, whilst the sensitivity was
                                                IgA nephropathy, hereditary nephritis, etc.
88.8% and 42.2%, respectively. At 20% of
                                                may be missed.
dysmorphism, the sensitivity (91.1%) and
specificity (92.86%) were suitable for dif-         High sensitivity of PCM in detection of
                                                glomerular hematuria in the range of 92-
TABLE III-Sensitivity and Predictive Value of   97% has been reported by various work-
           Percentage of Dysmorphic RBCs        ers(l,2,3,4,10) and compares well with
           on Phase Contrast Microscopy         91.2% observed by us (Table IV). Chang
                                                reported 90-100% sensitivity using Wright's
Dysmorphic    Sensitivity (%) Specificity (%)
 RBCs(%)                                        Stain(5), but we observed only 82% sensi-
                                                tivity with Wright's stain in detection of
     20           91.1            92.86         glomerular hematuria.
     40           88.8           100               The controversial issue is the percentage
                                                of dysmorphic RBCs for detection of
     80           42.2           100

glomerular hematuria. The percentage of           tions should include urine culture, calcium
dysmorphic RBCs required for the diagnosis        excretion studies, renal ultrasound, voiding
of glomerular hematuria has varied widely         cystourethrogram and urologic consultation
in different studies ranging from 80% to as       for nonglomerular diseases. The most sensi-
low as 10% (Table IV); and most of the            tive method to detect glomerular/nonglo-
studies have been in adults(2,3).                 merular hematuria at present is phase con-
                                                  trast microscopy, but the cost and need for
     In our study the presence of more than
                                                  training makes it exclusive and Wright's
40% dysmorphic RBCs has a 100% predic-
                                                  stain which is easily accessible may be used
tive value for glomerular hematuria but is
                                                  for routine purpose of detection of dysmor-
likely to miss 11.2% cases of glomerular he-
                                                  phic red cells in urine.
maturia that have lesser degrees of dysmor-
phism. On the other hand using a cut off          Acknowledgement
value of 20% of dysmorphism has an in-
                                                     We are grateful to Dr. S.M. Merchant,
creased sensitivity, with slight fall in speci-
                                                  Director, Research Department, Bai Jerbai
ficity to 93%. Twenty per cent dysmorphic
                                                  Wadia Hospital for Children and Research
RBCs could, therefore, be used as a cut off
                                                  Centre for giving permission to use phase
value to screen for glomerular hematuria.
                                                  contrast microscope for this study.
    In conclusion, this study clearly shows
that whenever hematuria is not accompa-
nied by typical clinical features of glomeru-       1. Birch DF, Fairley, KF. Hematuria:
lonephritis (edema, hypertension) and if               Glomerular or nonglomerular. Lancet
there are no casts in urine, dysmorphic red            1979, 2: 845-846.
cells indicate glomerular hematuria if 20%          2. Rizzoni G, Braggion F, Zacchello G.
or more of urinary red cells are dysmorphic.           Evaluation of glomerular and nonglome-
In case of asymptomatic isolated hematuria,            rular hematuria by phase contrast micro-
if dysmorphic RBCs are detected and ultra-             scopy. J Pediatr 1983,103: 370-374.
sonography has ruled out calculous disease,         3. Stapleton FB. Morphology of urinary red
further investigations (which may include              blood cells-A simple guide in localizing
kidney biopsy) for diagnosis of glomerular             the site of hematuria. Pediatr Clin North
disease are required. On the other hand if             Am 1987, 34: 561-569.
eumorphic RBCs are found then investiga-            4. De Santo NG, Nuzzi F, Capodicasa G,

INDIAN PEDIATRICS                                                     VOLUME 31- SEPTEMBER 1994

     Lama G, Caputo G, Rosati P, Giordani C.           localize the site of hematuria. Arch Dis
     Phase contrast microscopy of urine sedi-          Child 1991, 66: 338-340.
     ment for diagnosis of glomerular and non-    8.   Kincaid Smith P. Hematuria and exercise
     glomerular bleeding-data in children and          related hematuria. Br Med J 1982, 285:
     adults with normal creatinine clearance.          1595-1596.
     Nephron 1987, 45: 35-39.
                                                  9.   Zernicke. How I discovered phase con-
5.   Chang BS. Red cell morphology as a diag-          trast? Science, 1955, 121: 345-349.
     nostic aid in hematuria. JAMA 1984, 252:
                                                 10.   Fassett RG, Horgan BA, Mathew TH.
                                                       Detection of glomerular bleeding by
6.   Hauglustaine D, Bollens W, Michielson             phase contrast microscopy. Lancet 1982,
     P. Detection of glomerular bleeding using         1: 1432-1434.
     a simple method for light microscopy.       11.   Rath B, Turner C, Hartley B, Chantler C.
     Lancet 1982, 2: 761.                              What makes red cells dysmorphic in
7.   Rath B, Tuner C, Hartley B, Chantler C.           glomerular hematuria? Pediatr Nephrol
     An evaluation of light microscopy to              1992, 6: 424-427.


To top