Vibrio vulnificus occurs worldwide in warm seas Vibrio septicemia

Document Sample
Vibrio vulnificus occurs worldwide in warm seas Vibrio  septicemia Powered By Docstoc
					Vibrio vulnificus Bacteremia Associated with
Chronic Lymphocytic Leukemia,
Hypogammaglobulinemia, and Hepatic
Cirrhosis: Relation to Host and Exposure
Factors in 252 V. Vulnificus Infections
Reported in Louisiana

ABSTRACT: Background: Vibrio vulnificus infection in               (9.4%). The prevalence of preexisting conditions was
persons with B-chronic lymphocytic leukemia (B-CLL)                significantly greater in cases with positive blood cultures
or hypogammaglobulinemia has been reported infre-                  than in cases with positive wound or stool cultures.
quently. Patient and Methods: A woman with B-CLL,                  Exposure data in 116 cases revealed crab consumption
hypogammaglobulinemia, and hepatic cirrhosis died of               without raw oyster consumption or seawater exposure in
V. vulnificus bacteremia after eating cooked shrimp and            3.4%. Conclusion: The present patient had several condi-
crabs. We reviewed host and exposure data in 252 cases             tions associated with increased risk of V. vulnificus infec-
of V. vulnificus infection reported in Louisiana during            tion and bacteremia, especially hepatic cirrhosis, but her
the interval of 1980 through 2004. Results: V. vulnificus          route of exposure to V. vulnificus was unusual. KEY
was isolated from blood in 122 cases (48.8%). Preexist-            INDEXING TERMS: IgG; Bacteremia; Liver disease;
ing conditions in 138 cases included liver disease                 Malignancy. [Am J Med Sci 2006;332(4):216–220.]
(41.3%), malignancy (13.8%), and immunosuppression

                                                                   or hypogammaglobulinemia may also have in-
V    ibrio vulnificus occurs worldwide in warm seas,
     and infections with this spiral bacterium are
therefore common in the states adjacent to the Gulf
                                                                   creased susceptibility to V. vulnificus infection, but
                                                                   there are few descriptions of such cases.19 –21
of Mexico.1,2 Many infections are associated with the                 We report the case of a woman with B-CLL, hy-
consumption of raw or undercooked seafood or occur                 pogammaglobulinemia, hepatitis C, and hepatic cir-
when V. vulnificus enters superficial wounds that                  rhosis who died of V. vulnificus bacteremia. We also
contact seawater.1–9 Host factors that increase the                reviewed host and exposure factors in 252 cases of V.
risk of V. vulnificus infections include liver dis-                vulnificus infection reported in Louisiana during the
ease,5,10 –12 iron overload disorders,2,5,13–15 human              interval of 1980 through 2004. The pertinence of
immunodeficiency virus infection,7,11,16 and chronic               these and other observations to the risk of V. vulni-
renal insufficiency requiring hemodialysis.17,18 Per-              ficus infection in the present case is discussed.
sons with B-chronic lymphocytic leukemia (B-CLL)
                                                                                      Patients and Methods

   From the Southern Iron Disorders Center, and the Department     Case Report
of Medicine, University of Alabama at Birmingham, Birming-           A 73-year-old white woman from Louisiana was diagnosed with
ham, Alabama (JCB); and the Infectious Disease Epidemiology        B-CLL; flow cytometry demonstrated a large subpopulation of
Section, Louisiana Office of Public Health, Louisiana Department   blood mononuclear cells that coexpressed surface positivity for
of Health and Hospitals, New Orleans, Louisiana (RCR).             CD19, CD20, CD5, human leukocyte antigen (HLA)-DR, and
   Submitted January 27, 2006; accepted in revised form June 27,   kappa light chains. She was treated with fludarabine and cyclo-
2006. This work was supported in part by Southern Iron Disorders   phosphamide in September 2003 when she had a blood lympho-
Center and the Louisiana Department of Health and Hospitals.       cyte count of 155,000/mm3, marked lymphadenopathy, and
  Correspondence: Dr. James C. Barton, Suite G105, 2022 Brook-     splenomegaly; there was resolution of these abnormalities.
wood Medical Center Drive, Birmingham, AL 35029 (E-mail:             The patient moved to Alabama and was referred for further                                               evaluation and management. She reported that she consumed

216                                                                                              October 2006 Volume 332 Number 4
                                                                                                                 Barton and Ratard

Table 1. Ages of 173 Persons with Vibrio vulnificus Infections       seafood shop in October 2004; her daughter reported that they did
Reported in Louisiana during the Interval 1980 through 2004a         not consume raw or partially cooked seafood. Approximately 36
                                                                     hours later, the patient developed malaise, and diarrhea. Within
           Men, % (n) of     Women, % (n) of       Total, % (n) of   the next 16 hours while traveling, she developed fever, chills,
Age, y      122 cases           51 cases             173 cases       prostration, and multiple large fluid-filled bullae on her extrem-
                                                                     ities. She died of septic shock in an Alabama hospital emergency
 0–4             0                   0                    0          department; there was no apparent superficial wound. Blood
 5–14         0.8 (1)             2.0 (1)              1.2 (2)       cultures were positive for V. vulnificus. Inspection of the seafood
15–24         9.0 (11)               0                 6.4 (11)      shop by personnel of the Louisiana Department of Health and
25–34        13.1 (16)           11.7 (6)             12.7 (22)      Hospitals revealed the use of food handling and storage practices
35–44        13.1 (16)           11.7 (6)             12.7 (22)      that permitted contamination of cooked shellfish with fresh sea-
45–54        31.1 (28)           23.5 (12)            23.1 (40)      water drippings from raw shellfish.
55–64        16.4 (20)           17.6 (9)             16.8 (29)
 65          24.6 (30)           33.3 (17)            27.2 (47)      Review of Vibrio vulnificus Infections Reported in Louisiana
  Age data were available on 173 of the 252 cases (68.7%) reported     We reviewed reports of V. vulnificus infections and preexisting
in Louisiana during the interval 1980 through 2004.                  conditions from the Louisiana Department of Health and Hospi-
                                                                     tals. There were 252 cases of V. vulnificus infection reported
                                                                     during the interval of 1980 through 2004, including that of the
ethanol infrequently. She developed diffuse interstitial pneumo-     present patient. Age and gender data were available in 173 cases;
nitis; a causative organism was not identified by cultures, bron-    most cases occurred in adults ( 90%) and 44.0% occurred in
choscopy, and lung biopsy. She had severe hypogammaglobuline-        persons 55 years of age or older (Table 1). There was a predom-
mia with serum IgG 135 mg/dL (reference range, 700 –1600 mg/         inance of men (70.5% of 173 cases) (Table 1). Month of infection
dL), IgA 5 mg/dL (reference range, 70 – 400 mg/dL), and IgM 3        data were available for 157 cases reported during the interval of
mg/dL (reference range, 40 –230 mg/dL). Serum creatinine was         1987 through 2004; more than 87% of cases were reported in May
0.4 mg/dL; blood urea nitrogen was 10 mg/dL. She had elevated        through November, with a peak number of cases in August
serum concentrations of alanine and aspartate aminotransferase,      (Figure 1). The number of reported V. vulnificus infections in-
positive hepatitis C antibody, and serum HCV viral RNA greater       creased from an average of 6 per year to 12 per year over the
than 5,000,000 copies/mL. Liver biopsy revealed early micronodu-     interval of 1980 through 2004 (Figure 2). The trends of these data
lar cirrhosis and mild steatosis consistent with chronic hepatitis   are described by the equation y 0.3838x 4.41 (R2 0.4517),
C, moderate chronic triaditis, and a monotypic lymphocytic infil-    where x is years and y is number of case reports (Figure 2). In
trate consistent with B-CLL; there was no increase in stainable      contrast, the number of reports of V parahaemoliticus infections
iron. HLA typing revealed positivity for A*02,*68; B*14,*27. Over    remained stable (approximately 10 cases per year) over the same
9 months, the patient had mild, stable lymphocytosis and throm-      interval; the trends of these data are described by the equation y
bocytopenia. Pneumonitis resolved with intravenous infusions of      0.0223x 9.19 (R2 0.0028) (data not shown). In the 252 cases,
IgG (400 mg/kg monthly) and oral antibiotics. In July 2004, nadir    V. vulnificus was isolated from blood in 122 cases (48.4%), from
serum IgG measurements before her IgG infusion revealed total        wounds in 85 cases (33.7%), from stools in 44 cases (17.5%), and
IgG level 695 mg/dL with subnormal IgG1 and IgG3 levels of 416       from peritoneal fluid in one case (0.4%).
mg/dL (reference range, 422–1292 mg/dL) and 25 mg/dL (refer-           Reports of preexisting conditions were available for 138 cases of
ence range, 41–129 mg/dL), respectively.                             V. vulnificus infection. Liver disease and malignancy were among
  The patient revisited Louisiana but did not receive IgG infu-      the three conditions most frequently reported; immunosuppres-
sions thereafter. She and her children consumed a “take-out”         sion was the seventh most frequently reported condition (Table
meal of boiled shrimp and crabs purchased from a waterside           2). The prevalence of reports of preexisting conditions was signif-

Figure 1. Months of 157 Vibrio vulnificus infections reported in Louisiana during the interval 1987 through 2004.

THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES                                                                                       217
Vibrio vulnificus and Leukemia

Figure 2. Trend of 252 Vibrio vulnificus infections reported in Louisiana during the interval 1980 through 2004.

icantly greater in cases associated with positive blood cultures         the United States and in other geographic areas,
than in cases in which wound or stool cultures were positive.
   Exposure data were available for 116 cases. Consumption of raw
                                                                         many of which are also contiguous to warm seas.1,2
oysters was reported in 55 cases (47.4%); exposure to seawater was       In the southeastern United States, onset of illness
reported in 51 cases (44.0%). Consumption of crab was reported in        with V. vulnificus infection usually occurs in March
16 cases (13.8%). In 4 of these 16 cases, consumption of crab was        through November.4,5 In the present case series,
reported without accompanying reports of consumption of raw oys-
ters or exposure to seawater. In three of these four cases, V. vulni-    most infections occurred in a similar interval. Forty-
ficus was isolated from blood. Possible explanations for these four      four percent of the 173 present cases for whom age
cases and the present case include consumption of inadequately           data were available were 55 years of age or older.
cooked crabs, or contamination of fully cooked crabs with seawater       This is consistent with previous reports that the
or fresh shellfish during preparation or consumption.
                                                                         median age of persons who develop V. vulnificus
                                                                         septicemia or wound infections in this geographic
                                                                         area is greater than 60 years.18,22 These observa-
  Infections with V. vulnificus are reported fre-                        tions are also consistent with the month of disease
quently in states adjacent to the Gulf of Mexico in                      onset and age of the present patient. The pattern of

    Table 2. Preexisting Conditions and Positive Cultures in 138 Cases of Vibrio vulnificus Infections Reported in Louisiana 1980
                                                            through 2004

                                                   Positive Culture           Positive Culture            Positive Culture
    Preexisting Condition Reporteda                   of Bloodb                  of Woundb                    of Stoolb               Totalb

Yes                                                      82                         41                          26                     113
No                                                        8                         16                           9                      25
Total cases                                              90                         57                          35                     138
Cases with pre-existing condition, %c                    91.1                       71.9                        74.3                    82
  Percentages of cases in which preexisting conditions were reported included: liver disease, 41.3%; alcoholism, 26.1%; malignancy, 13.8%;
peptic ulcer, 12.3%; renal failure, 12.3%; antacid use, 11.6%; immunosuppression, 9.4%; high-dose steroid use, 5.1%; insulin-dependent
diabetes mellitus, 4.3%; gastritis, 4.3%; and splenectomy 2.2%. Precise definitions of the preexisting conditions were not employed on
report forms.
  Two or more sites of positive cultures were reported in some cases.
  The prevalence of reports of preexisting conditions was significantly higher in cases associated with positive blood cultures than in cases
with positive wound or stool cultures (P 0.0057, Pearson’s 3 3 2 analysis).

218                                                                                                      October 2006 Volume 332 Number 4
                                                                                                     Barton and Ratard

reporting observed in the present study indicates          hypogammaglobulinemia also occur in patients with
that the frequency of V. vulnificus infections is in-      B-CLL,31–34 including some that are attributable to
creasing, and that increased numbers of reports are        antileukemia chemotherapy.35 The present data from
not due to increased surveillance. Most infections         Louisiana demonstrate that malignancies were re-
are attributable to ingestion of uncooked shellfish        ported as preexisting conditions in 14% of cases, al-
(typically raw oysters), or to exposure of superficial     though more specific definition of these malignancies
wounds, eyes, or ears to salty or brackish water or to     was not available.
drippings from raw seafoods.1–3,5–7,9,12,13,23 Although       In the present analysis of previous V. vulnificus
exposure data were available in only 46% of the            infections, the preexisting condition with the greatest
present 252 Louisiana cases, the present observa-          prevalence was liver disease. However, a potential
tions are consistent with these reports. The usual         limitation of the present study is that reports of pre-
primary site of infection in patients who develop          existing conditions were available in only 55% of cases.
bullae and bacteremia is the gastrointestinal tract,       In previous reports, cirrhosis and other liver disorders
as in the present patient. It is likely that her infec-    were associated with a significantly increased risk for
tion was caused by the ingestion of undercooked            V. vulnificus infections,10,12,36,37 and V. vulnificus in-
shrimp and crabs, or ingestion of uncooked seawater        fections have been reported to occur in persons with
that contaminated cooked shrimp and crabs. V. vulni-       hepatitis C.7,38 Low neutrophil phagocytosis activity
ficus infections associated with bacteremia are se-        and high serum ferritin level were independent pre-
vere and typically occur only in persons with under-       dictors of the survival of V. vulnificus in the whole
lying diseases, like those described in the present        blood of patients with chronic liver disease.37 Abnor-
review and in other reports.2,4,5,10,14,15,17,18,21        mal neutrophil function, including reduced phagocyto-
   Immunoglobulin G antibodies that react with V.          sis by neutrophils, also occurs in B-CLL,31,32,34 al-
vulnificus or its antigenic determinants are detect-       though serum ferritin concentrations in persons with
able in the sera of infected patients and in persons       B-CLL are typically normal.39
without known exposure to V. vulnificus.24 –27 The            We conclude that the present patient had several
results of studies of the prevalences of anti-V. vulni-    preexisting conditions associated with increased
ficus antibody in persons with presumed exposure           risk of V. vulnificus infection and bacteremia, espe-
to V. vulnificus (shellfish industry workers) and in       cially hepatic cirrhosis. Her exposure to V. vulnifi-
persons with minimal exposure (Seventh Day Ad-             cus by consumption of inadequately cooked crab or
ventists) are conflicting.26,27 Because antibodies that    crab contaminated with drippings from seawater
react with the capsular polysaccharides of V. vulni-       during preparation or consumption was unusual.
ficus are demonstrable in persons without a history        Physicians should warn at-risk patients to avoid
of V. vulnificus infection,26,27 some persons in the       consumption of uncooked shellfish harvested in
general population may have exposure to V. vulnifi-        warm seas. Contact with or ingestion of seawater in
cus without apparent infection, or have cross-react-       natural bodies of seawater or as a result of contam-
ing antibodies. A patient with hypogammaglobu-             ination of thoroughly cooked food by seawater or
linemia due to nephrotic syndrome had V. vulnificus        shellfish drippings may also result in infection. Ad-
infection.21 Altogether, these results suggest that        visories that persons with risk factors for V. vulni-
the synthesis of specific or cross-reacting IgG anti-      ficus infection should not eat uncooked shellfish may
body in humans may provide some immunity                   also be effective when displayed at all points of sale
against V. vulnificus. In the present data from Lou-       of uncooked oysters or other shellfish.1,5
isiana, immunosuppression was reported as a pre-
existing condition in 9% of cases, and splenectomy                                 References
was reported in 2% of cases. This suggests that there
was an immunologic abnormality in these cases,             1. Shapiro RL, Altekruse S, Hutwagner L, et al. The role of
although more specific definition of these preexist-          Gulf Coast oysters harvested in warmer months in Vibrio
ing conditions was not available.                             vulnificus infections in the United States, 1988-1996. Vibrio
                                                              Working Group. J Infect Dis 1998;178:752–9.
   Vibrio vulnificus infections have occurred in pa-       2. Bullen JJ. Bacterial infections in hemochromatosis. In: Bar-
tients with B-CLL,19,20 including a 68-year-old man           ton JC, Edwards CQ, eds. Hemochromatosis: Genetics, Patho-
who had hepatic cirrhosis.20 The present patient had          physiology, Diagnosis, and Treatment. Cambridge, UK: Cam-
subnormal levels of IgG typical of a high proportion of       bridge University Press; 2000. p. 381–6.
patients with B-CLL.28 –30 It is probable that her se-     3. Blake PA. Vibrios on the half shell: what the walrus and the
rum or plasma IgG levels were subnormal at the time           carpenter didn’t know. Ann Intern Med 1983;99:558–9.
her infection occurred due to failure to receive monthly   4. Bonner JR, Coker AS, Berryman CR, et al. Spectrum of
                                                              Vibrio infections in a Gulf Coast community. Ann Intern Med
IgG infusion for 3 months. However, we did not mea-
                                                              1983; 99:464–9.
sure her serum antibodies against V. vulnificus, nor       5. Hlady WG, Klontz KC. The epidemiology of Vibrio infec-
were we able to obtain information regarding anti-V.          tions in Florida, 1981-1993. J Infect Dis 1996; 173:1176–83.
vulnificus specificity and titers in commercially avail-   6. Colodner R, Raz R, Meir I, et al. Identification of the
able IgG preparations. Immunity deficits other than           emerging pathogen Vibrio vulnificus biotype 3 by commer-

THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES                                                                          219
Vibrio vulnificus and Leukemia

      cially available phenotypic methods. J Clin Microbiol 2004;             onstration of antibodies by indirect immunofluorescence.
      42:4137–40.                                                             Arch Intern Med 1983;143:837–8.
 7.   Ulusarac O, Carter E. Varied clinical presentations of            25.   Musher DM, Hansen MV, Goree A, et al. Emergence of
      Vibrio vulnificus infections: a report of four unusual cases            bactericidal and opsonizing antibody to Vibrio vulnificus fol-
      and review of the literature. South Med J 2004;97:163–8.                lowing bacterial infection. J Clin Microbiol 1986;23:411–5.
 8.   Oliver JD. Wound infections caused by Vibrio vulnificus and       26.   Lefkowitz A, Fout GS, Losonsky G, et al. A serosurvey of
      other marine bacteria. Epidemiol Infect 2005;133:383–91.                pathogens associated with shellfish: prevalence of antibodies
 9.   Vibrio illnesses after Hurricane Katrina–multiple states, Au-           to Vibrio species and Norwalk virus in the Chesapeake Bay
      gust-September 2005. MMWR Morb Mortal Wkly Rep 2005;                    region. Am J Epidemiol 1992;135:369–80.
      54:928–31.                                                        27.   Fiore A, Hayat U, Wasserman SS, et al. Antibodies that
10.   Vollberg CM, Herrera JL. Vibrio vulnificus infection: an
                                                                              react with the capsular polysaccharide of Vibrio vulnificus
      important cause of septicemia in patients with cirrhosis.
                                                                              are detectable in infected patients, and in persons without
      South Med J 1997;90:1040–2.
                                                                              known exposure to the organism. Diagn Microbiol Infect Dis
11.   Chen Y, Satoh T, Tokunaga O. Vibrio vulnificus infection
      in patients with liver disease: report of five autopsy cases.           1996 March;24:165–7.
                                                                        28.   Aittoniemi J, Miettinen A, Laine S, et al. Opsonising
      Virchows Arch 2002 July;441:88–92.
12.   Haq SM, Dayal HH. Chronic liver disease and consumption                 immunoglobulins and mannan-binding lectin in chronic lym-
      of raw oysters: a potentially lethal combination–a review of            phocytic leukemia. Leuk Lymphoma 1999;34:381–5.
      Vibrio vulnificus septicemia. Am J Gastroenterol 2005 May;        29.   Rozman C, Montserrat E, Vinolas N. Serum immuno-
      100:1195–9.                                                             globulins in B-chronic lymphocytic leukemia: natural history
13.   Tefany FJ, Lee S, Shumack S. Oysters, iron overload and                 and prognostic significance. Cancer 1988;61:279–83.
      Vibrio vulnificus septicaemia. Australas J Dermatol 1990;31:      30.   Tsiodras S, Samonis G, Keating MJ, et al. Infection and
      27–31.                                                                  immunity in chronic lymphocytic leukemia. Mayo Clin Proc
14.   Bullen JJ, Spalding PB, Ward CG, et al. Hemochromato-                   2000;75:1039–54.
      sis, iron and septicemia caused by Vibrio vulnificus. Arch        31.   Mittelman M, Fishman P, Djaldetti M. Reduced phagocyto-
      Intern Med 1991; 151:1606–9.                                            sis promoting activity of saliva from patients with chronic lym-
15.   Gerhard GS, Levin KA, Price GJ, et al. Vibrio vulnificus                phocytic leukemia. Biomed Pharmacother 1986;40:308–11.
      septicemia in a patient with the hemochromatosis HFE              32.   Nagler A, Obedeanu N, Tatarsky I, et al. Chemilumines-
      C282Y mutation. Arch Pathol Lab Med 2001;125:1107–9.                    cence response of polymorphonuclear leucocytes to serum-
16.   Chin KP, Lowe MA, Tong MJ, et al. Vibrio vulnificus                     opsonized zymosan–a predictive parameter for infections and
      infection after raw oyster ingestion in a patient with liver            prognosis in multiple myeloma patients. Clin Lab Haematol
      disease and acquired immune deficiency syndrome-related                 1988;10:365–75.
      complex. Gastroenterology 1987;92:796–9.                          33.   Morra E, Nosari A, Montillo M. Infectious complications in
17.   Gholami P, Lew SQ, Klontz KC. Raw shellfish consump-                    chronic lymphocytic leukaemia. Hematol Cell Ther 1999 Au-
      tion among renal disease patients: a risk factor for severe             gust;41:145–51.
      Vibrio vulnificus infection. Am J Prev Med 1998;15:243–5.         34.   Itala M, Vainio O, Remes K. Functional abnormalities in
18.   Barton JC, Coghlan ME, Reymann MT, et al. Vibrio                        granulocytes predict susceptibility to bacterial infections in
      vulnificus infection in a hemodialysis patient receiving intra-         chronic lymphocytic leukaemia. Eur J Haematol 1996;57:46–53.
      venous iron therapy. Clin Infect Dis 2003;37:e63–e67.             35.   Kipps TJ. Chronic lymphocytic leukemia and related dis-
19.   Holmberg SD. Vibrios and aeromonas. Infect Dis Clin North
                                                                              eases. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ,
      Am 1988;2:655–76.
                                                                              Seligsohn U, eds. Williams Hematology. New York: McGraw-
20.   Bock T, Christensen N, Eriksen NH, et al. The first fatal
                                                                              Hill; 2001. p. 1163–94.
      case of Vibrio vulnificus infection in Denmark. APMIS 1994;
                                                                        36.   Chang JJ, Sheen IS, Peng SM, et al. Vibrio vulnificus
21.   Wang SM, Liu CC, Chiou YY, et al. Vibrio vulnificus                     infection: report of 8 cases and review of cases in Taiwan.
      infection complicated by acute respiratory distress syndrome            Changgeng Yi Xue Za Zhi 1994;17:339–46.
                                                                        37.   Hor LI, Chang TT, Wang ST. Survival of Vibrio vulnificus
      in a child with nephrotic syndrome. Pediatr Pulmonol 2000;
      29:400–3.                                                               in whole blood from patients with chronic liver diseases:
22.   Klontz KC, Lieb S, Schreiber M, et al. Syndromes of                     association with phagocytosis by neutrophils and serum fer-
      Vibrio vulnificus infections: clinical and epidemiologic fea-           ritin levels. J Infect Dis 1999;179:275–8.
      tures in Florida cases, 1981-1987. Ann Intern Med 1988;109:       38.   Osawa M, Tachibana M, Arita M, et al. [An autopsied case
      318–23.                                                                 of septicemia due to Vibrio vulnificus]. Kansenshogaku
23.   Taylor R, McDonald M, Russ G, et al. Vibrio alginolyticus               Zasshi 2002;76:63–6.
      peritonitis associated with ambulatory peritoneal dialysis. Br    39.   Moroz C, Bessler H, Lurie Y, et al. New monoclonal anti-
      Med J (Clin Res Ed) 1981;283:275.                                       body enzymoassay for the specific measurement of placental
24.   Pollak SJ, Parrish EF III, Barrett TJ, et al. Vibrio vulni-             ferritin isotype in hematologic malignancies. Exp Hematol
      ficus septicemia. Isolation of organism from stool and dem-             1987;15:258–62.

220                                                                                                       October 2006 Volume 332 Number 4