Original Article
The Epidemiology and Clinical Aspects of the Hemolytic Uremic Syndrome in Minnesota
N Engl J Med 1990; 323:1161-1167October 25, 1990
- Abstract
- Abstract
Background.
The frequency of the hemolytic uremic syndrome, characterized by microangiopathic hemolytic anemia, thrombocytopenia, and renal failure, is increasing. Although Escherichia coli serotype 0157:H7 has been implicated as a causative agent, more information is needed about the basic epidemiology and clinical aspects of this syndrome.
Methods.
We conducted a retrospective population-based study of hemolytic uremic syndrome in Minnesota residents less than 18 years of age from 1979 through 1988 to assess trends in disease occurrence, describe the clinical illness, and identify predictors of disease severity and outcome. We also conducted a case–control study of patients with onsets of illness from 1986 through 1988 to examine risk factors.
Results.
One hundred seventeen patients were identified. The mean annual incidence increased from 0.5 case per 100,000 child-years among children less than 18 in 1979 (6 cases) to 2.0 cases per 100,000 in 1988 (26 cases) (P = 0.000004). E. coli 0157:H7 was isolated from 13 of 28 patients (46 percent) who had stool specimens submitted for testing. For those who presented with typical hemolytic uremic syndrome, an elevated polymorphonuclear-leukocyte count on hospital admission, a shorter duration of prodrome, and the presence of bloody diarrhea were predictive of severe disease. In the case–control study, the patients were more likely to attend large day-care centers (more than 50 children) than were the controls (odds ratio, 10.2; P = 0.03), suggesting that day-care attendance may be a risk factor. On the basis of the population-attributable risk, however, this factor could account for no more than 16 percent of the cases.
Conclusions.
This study provides evidence for an increase in the incidence of hemolytic uremic syndrome, which is probably related to an increased incidence of E. coli 0157:H7 infections. Hemolytic uremic syndrome has become an important pediatric and public health problem. (N Engl J Med 1990; 323:1161–7.)
Media in This Article
Figure 1
Number of Cases of Hemolytic Uremic Syndrome in Minnesota Children, 1979 through 1988.Figure 2
Annual Age-Specific Incidence Rates of Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.Article Activity
- Article
THE hemolytic uremic syndrome is a leading cause of acute renal failure in childhood.1 The classic triad of microangiopathic hemolytic anemia, thrombocytopenia, and renal failure was first reported in 1955,2 and subsequently progress has been made in understanding the epidemiology and pathophysiology of this syndrome. Distinct subgroups of the syndrome have been identified,3 , 4 and serotypes of Escherichia coli that produce a cytotoxin active on Vero cells (verotoxin-producing serotypes) have been implicated as important etiologic agents in the more common form of hemolytic uremic syndrome.5 6 7 8 9 E. coli 0157:H7, the most frequently isolated serotype of verotoxin-producing E. coli in the United States, is capable of causing a broad spectrum of illness, including nonbloody diarrhea, bloody diarrhea, hemolytic uremic syndrome, and thrombotic thrombocytopenic purpura.10
A recently published population-based study from King County, Washington, documented an increase in the incidence of hemolytic uremic syndrome in that area over the past 15 years.11 Other population-based studies have been limited in their ability to ascertain temporal trends in the syndrome.12 , 13 For a better understanding of its epidemiology in Minnesota, we conducted a 10-year retrospective record review of all cases in Minnesota requiring hospitalization. The results were used to assess trends in the incidence of disease, describe demographic features, characterize the clinical illness, and identify possible predictors of disease severity. To evaluate risk factors for hemolytic uremic syndrome, we also conducted a case–control study involving patients with onsets of illness from January 1986 through December 1988. Our findings indicate that hemolytic uremic syndrome is emerging as an important public health and pediatric infectious disease problem.
Methods
Epidemiologic Investigation
All 165 hospitals in Minnesota and 5 referral hospitals in contiguous states were asked to provide identifying information for patients with a discharge diagnosis of hemolytic uremic syndrome (International Classification of Diseases, ninth revision, section 283.1) or thrombotic thrombocytopenic purpura (section 446.6) during the 10-year period from January 1979 through December 1988. A list of patients was received from all the hospitals that were contacted. Age-specific population data for the study period were obtained from the Minnesota Center for Health Statistics.
We then reviewed the medical records of all patients who were less than 18 years of age. The following data were recorded for each patient: demographic information, presenting symptoms, use of antimicrobial therapy during the prodromal illness, laboratory data on admission (blood urea nitrogen level, serum creatinine level, hematocrit, platelet count, white-cell count, and differential cell count), results of stool culture for routine pathogens (if performed), the need for transfusions (type and amount of transfusion), the need for dialysis, the length of the hospital stay, complications during hospitalization, and outcome. If a stool specimen was submitted for E. coli 0157:H7 testing, the results were obtained from the Division of Public Health Laboratories of the Minnesota Department of Health. The isolates of E. coli 0157:H7 were also tested for verotoxin (Shiga-like toxin) production at the Centers for Disease Control Enteric Bacteriology Laboratory.
We also identified all Minnesota residents less than 18 years of age who died between 1979 and 1988 and who had hemolytic uremic syndrome or thrombotic thrombocytopenic purpura on their death certificates. In addition, we reviewed data on recipients of kidney transplants at the University of Minnesota Kidney Transplant Center.
A case was defined as an illness in a Minnesota resident younger than 18 with the following characteristics: microangiopathic hemolytic anemia (hematocrit-volume fraction less than 0.30, with evidence of red-cell fragmentation), thrombocytopenia (platelet count less than 150×109 per liter), and evidence of renal involvement with at least two of the following: blood urea nitrogen level higher than 7.1 mmol per liter, serum creatinine level higher than 71 μmol per liter, or abnormal urinary sediment, as evidenced by cylindruria, proteinuria, or hematuria.
To evaluate potential diagnostic bias during the study period, we examined trends over time in admission laboratory values for blood urea nitrogen, serum creatinine, hematocrit, and platelet count, the proportion of all patients requiring dialysis, the number of patients classified as having severe disease, and the number of patients classified as having a poor outcome (as defined below).
Clinical Evaluation
The morbidity associated with hemolytic uremic syndrome was assessed in two ways. First, the patients were classified as having a poor long-term outcome if they died before or after they were discharged or if any of the following conditions were present at or after hospital discharge: blindness, hemiparesis, chronic renal failure, or the need for anticonvulsant therapy, kidney transplantation, or colonic resection. A good long-term outcome was defined as the absence of any of the preceding conditions. Second, the patients were classified as having mild or severe disease. Severe disease was defined as that ending in death or meeting two of the following three criteria: hospitalization for more than two weeks, dialysis for more than 10 days, and the presence of at least one severe complication. Severe complications included all the poor-outcome variables noted above, as well as shock, cardiac arrest, coagulopathy, and the need for mechanical ventilation.
Several factors were evaluated as predictors of outcome and disease severity, including age at onset; time from the onset of symptoms to hospitalization (duration of prodrome); type of presentation of hemolytic uremic syndrome — typical, or classic (with a diarrheal prodrome), or atypical (without a diarrheal prodrome); presence or absence of bloody diarrhea for those with a typical presentation; use of antimicrobial agents during the prodrome; and polymorphonuclear-leukocyte count in the peripheral blood on admission.
Case–Control Study
All patients who became ill from January 1986 through December 1988 and who were less than six years of age at the time of onset were included in the case–control study. Two controls per case, matched for date of birth and residence status (rural or urban), were randomly selected from all Minnesota birth certificates with the use of a random-numbers table. When either of the first two controls could not be enrolled, alternative controls were randomly selected with the same method. The parents of patients and controls were asked about their child's day-care use (frequency of attendance and size of the facility), consumption of raw milk, exposure to farm animals, and approximate pattern of hamburger consumption in the month before the onset of illness in the patients. The parents were also asked about the number of siblings in the household and their ages, illness in family members in the month before or after the patient's illness (for patients only), and family income.
Statistical Analysis
Summary rates were adjusted for age to the 1980 U.S. population. Trends in incidence, the need for dialysis, long-term complications, and severity of the clinical course were evaluated with the chi-square test for trend. Nonparametric methods (Kruskal—Wallis) were used to test differences over time in laboratory values on admission. Standard univariate methods (chi-square test and Student's t-test) were used to examine demographic variables and variables considered possible predictors of outcome or severe disease. Pearson's correlation coefficients were calculated to assess the interrelatedness of certain predictor variables.14 Stepwise logistic-regression analysis was then used to evaluate the effect of certain predictor variables on outcome and disease severity.15 Independent variables entered into the model included polymorphonuclear-leukocyte count, age, type of presentation (typical or atypical), previous use of antimicrobial drugs, and duration of prodrome for all patients. A second model included only the patients with a typical presentation; the factors entered included all those in the first model and the presence of bloody diarrhea.
The MantelHaenszel odds-ratio estimate for matched multiple controls and the MantelHaenszel test of the null hypothesis (two-sided test) were used in univariate analyses of case–control data. Conditional logistic-regression analysis was used to preserve matching when indicator variables were analyzed.
Results
Epidemiologic Investigation
A review of hospital records identified 117 cases of hemolytic uremic syndrome in Minnesota residents younger than 18 years of age during the study period. No additional cases were identified by a review of state death-certificate records or records from the University of Minnesota Kidney Transplant Center. The mean age of the patients was 3.8 years (range, 2 months to 17 years); 90 (77 percent) were under 5. No significant differences were noted in sex or residence status; 113 (97 percent) of the patients were white, a proportion similar to that in the population of Minnesota as a whole (Table 1Table 1
Demographic Characteristics of 117 Patients with Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.).The number of cases of hemolytic uremic syndrome per year increased substantially during the study period (Fig. 1Figure 1
Number of Cases of Hemolytic Uremic Syndrome in Minnesota Children, 1979 through 1988.). The annual incidence among children less than 18 years of age increased from 0.5 case per 100,000 child-years in 1979 (6 cases) to 2.0 cases per 100,000 child-years in 1988 (26 cases) (P = 0.000004). Most of the increase occurred in children less than five years of age; in this group, the incidence increased from 1.6 cases to 5.8 cases per 100,000 child-years (P = 0.0002) (Fig. 2Figure 2Annual Age-Specific Incidence Rates of Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.). The majority of these cases were sporadic, with only two clusters identified. The first cluster involved two children attending the same day-care facility in 1986, and the second involved three children attending the same day-care facility in 1988. A seasonal pattern of occurrence was noted: 86 (74 percent) of the cases occurred from June through October (Fig. 3Figure 3
Number of Cases of Hemolytic Uremic Syndrome in Minnesota, According to the Month of Onset, 1979 through 1988.).All patients were eventually hospitalized at a total of 16 hospitals; 91 (78 percent) were hospitalized at 1 of 4 hospitals. Although the proportion of patients referred to a second hospital decreased during the study period, the referral pattern did not change. The mean laboratory values on admission did not change over time, except for blood urea nitrogen levels, which dropped slightly. There were no temporal trends in the annual proportion of patients with severe disease, poor outcomes, or a need for dialysis.
Clinical Evaluation
Clinical Characteristics
The presenting symptoms are shown in Table 2Table 2
Presenting Symptoms of 117 Patients with Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.. A total of 101 patients (86 percent) had typical hemolytic uremic syndrome, and 69 of these (68 percent) had bloody diarrhea. Of the 16 patients who had atypical hemolytic uremic syndrome, 6 presented with a respiratory prodrome and 10 had no distinct prodrome. The mean duration of hospitalization was 15.4 days (range, 2 to 60). Fifty-five patients (47 percent) required dialysis for a mean of 12 days (range, 1 to 60 days). One hundred four patients (89 percent) received erythrocyte transfusions, 36 (31 percent) received platelets, and 27 (23 percent) received fresh-frozen plasma. Twenty patients (17 percent) had received antimicrobial agents during their prodromal illness and before hospitalization; most were treated for other conditions, such as otitis or pharyngitis.Twenty-seven patients (23 percent) had serious complications during their hospital stay or after they were discharged (Table 3Table 3
Complications According to System in 27 Patients with Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.). Six of the eight patients who underwent laparotomy had the surgery before the diagnosis of hemolytic uremic syndrome was made; all presented with bloody diarrhea and were initially given the diagnosis of acute abdomen or possible appendicitis. Only one child had evidence of intestinal perforation before surgery. The onsets of illness in these patients were distributed over the study period, with three occurring early in the period (1980 through 1981) and five occurring in 1986 or later.Four children died, yielding a case fatality rate of 3.4 percent. Three of them died early in the clinical course; all three had a diarrheal prodrome, two had severe neurologic involvement, and one had overwhelming sepsis. The fourth child, who had an atypical presentation and a prolonged clinical course, died of neurologic sequelae four months after the onset of hemolytic uremic syndrome.
Microbiologic Findings
Seventy-five patients (64 percent) had stool cultures for routine enteric pathogens (salmonella, shigella, and campylobacter); salmonella was isolated from the stool of one 17-month-old child. E. coli 0157:H7 was isolated from 13 of the 28 patients (46 percent) who had stools submitted for examination for this pathogen. No correlation was noted between the time from the onset of diarrhea to the collection of the stool specimen and culture results. Of 11 isolates analyzed for verotoxin production, 8 produced verotoxin Types I and II and 3 produced only verotoxin Type II. No data on antimicrobial sensitivity testing were available for these isolates.
Predictors of Disease Severity
Thirty-three patients (28 percent) were classified as having severe disease according to the criteria outlined in Methods. Univariate and multivariate analyses of all cases revealed that the polymorphonuclear-leukocyte count on admission was highly correlated with the severity of disease (P = 0.00005) (Table 4Table 4
Polymorphonuclear-Leukocyte Count on Admission as a Predictor of Outcome and Disease Severity in 113 Patients with Hemolytic Uremic Syndrome in Minnesota, 1979 through 1988.*) and previous use of antimicrobial agents correlated with mild disease (1 of 33 with severe disease received antimicrobial agents, as compared with 19 of 84 with mild disease) (odds ratio, 9.4; 95 percent confidence interval, 1.2 to 199.0; P = 0.01). The results were similar for that subgroup of patients who presented with diarrhea. For that subgroup, univariate analysis demonstrated that a shorter duration of prodrome, 4.8 days as compared with 6.6 days (P = 0.03) and the presence of bloody diarrhea (odds ratio, 3.1; 95 percent confidence interval, 1.0 to 10.4; P = 0.03) were also associated with severe disease. Twenty-five of 30 patients (83 percent) with severe disease had bloody diarrhea, as compared with 44 of 71 patients (62 percent) with mild disease. Age did not correlate with severe disease. Analysis with Pearson's correlation coefficient showed that the polymorphonuclear-leukocyte count, presence of bloody diarrhea, and shorter duration of prodromal illness were highly correlated with each other (P<0.001). Antimicrobial treatment during the prodromal illness did not correlate with these variables.Predictors of Poor Outcome
Seventeen children (14 percent) were classified as having a poor outcome. Univariate analysis demonstrated that the polymorphonuclear-leukocyte count on admission correlated with outcome (P = 0.03) (Table 4) and pretreatment with antimicrobial agents had a protective effect (odds ratio, undefined; upper bound of 95 percent confidence interval, 1.3; P = 0.04); none of the 17 patients with a poor outcome had received antimicrobial agents previously, as compared with 20 of the 100 patients who had a good outcome. Multivariate analysis of the subgroup of patients who presented with diarrhea confirmed that elevated polymorphonuclear-leukocyte counts on admission predicted a poor outcome (P = 0.01); a regression model that used all cases demonstrated the same findings (P = 0.003).
Case–Control Study
Of 49 eligible patients, 45 (92 percent) participated in the case–control study; the parents of 1 child refused, and we were unable to contact the parents of the remaining 3. Of the controls, 82 (91 percent) were the first two selected, and 8 (9 percent) were the third or fourth selected. The median age of the participants was 2.4 years, and 21 (47 percent) were urban residents. All patients and 94 percent of the controls were white. The median family income was similar for patients and controls ($35,000 per year). Rates of day-care attendance were similar for patients and controls: 20 patients (44 percent) and 40 controls (44 percent) attended day-care centers. The patients, however, were more likely to attend large day-care centers (with 50 or more children) than were controls (7 of 45 as compared with 2 of 90) (odds ratio, 10.2; 95 percent confidence interval, 1.2 to 87.8; P = 0.03 by logistic regression). This analysis included three children in whom hemolytic uremic syndrome developed after an outbreak of E. coli 0157:H7 infection in one large day-care center. When the data were analyzed including only the index child for that cluster, a trend toward statistical significance was still present (odds ratio, 7.1; 95 percent confidence interval, 0.8 to 64.8; P = 0.08). This finding was associated with children in the Twin Cities metropolitan area, since none of the patients or controls living in rural areas attended any child-care facility that was not in a private home. The proportion of disease accounted for by attendance at a large day-care center (the population-attributable risk) was 16 percent. There were no significant differences between patients and controls in the history of raw-milk consumption, contact with farms or farm animals, number of siblings in day care, total number of hours spent in day care per week, or approximate pattern of hamburger consumption.
Discussion
Our results document a substantial increase in the incidence of hemolytic uremic syndrome in Minnesota from 1979 through 1988, with the highest rate of disease occurring in children less than five years of age. The trend noted in this study persisted through 1989: 21 cases of hemolytic uremic syndrome were reported to the Minnesota Department of Health during that year, yielding an annual incidence of 1.6 cases per 100,000 child-years. These cases occurred despite extensive media publicity in the state about risk factors for E. coli 0157:H7 infection (such as eating undercooked hamburger). The importance of hemolytic uremic syndrome can be illustrated by comparing its current incidence with the incidence of Reye's syndrome in Minnesota from 1980 through 1984 (before the association of salicylate use in patients with influenza or chickenpox was recognized). During that time, the Minnesota Department of Health conducted statewide active surveillance for Reye's syndrome as part of a Centers for Disease Control multicenter study. The mean annual incidence identified through active surveillance was 0.98 case per 100,000 child-years (an average of 11 cases per year), which is lower than the current incidence of hemolytic uremic syndrome in our state.
Most of the increase identified in the incidence of hemolytic uremic syndrome can be attributed to disease in young children. Thus, age appears to be a major risk factor for hemolytic uremic syndrome. It is likely that other host factors, such as preexisting immunity to precipitating infectious agents or to toxins, are important in determining the risk of disease.16 , 17
We do not believe that the increase in the incidence of hemolytic uremic syndrome noted here is an artifact resulting from increased awareness of this condition by physicians, for the following reasons. First, the percentage of all patients requiring dialysis and the percentage of patients with the syndrome who were classified as having severe disease or poor outcomes were stable during the 10-year study period, suggesting that milder cases were not being recognized with increased frequency later in the study period. Second, no significant change in laboratory values on admission was noted over time. Finally, referral patterns remained relatively stable, with most cases being referred to tertiary care centers in which the diagnosis of hemolytic uremic syndrome was well recognized even in 1979, thus limiting the potential for detection bias.
The increasing incidence of hemolytic uremic syndrome noted here coincides with the emergence of verotoxin-producing strains of E. coli as important pathogens capable of causing a broad spectrum of disease in humans, including hemolytic uremic syndrome. Evidence of infection with E. coli 0157:H7 was noted in 46 percent of the patients who had stool cultures performed. It is likely that a higher percentage of patients was actually infected with E. coli 0157:H7, given the variability in isolating the pathogen from the stool of patients identified in outbreak settings.18 The striking seasonality of cases is similar to the seasonal pattern of E. coli 0157:H7 infection in Minnesota, indicating that cases of hemolytic uremic syndrome serve as a community measure for the incidence of E. coli 0157:H7.
We were able to examine potential predictors of the severity of illness and outcome. Children with elevated polymorphonuclear-leukocyte counts on admission, especially those exceeding 15×109 per liter, were found to be at higher risk for severe disease and poor outcome in both univariate and multivariate analyses. A similar finding was noted by Walters et al. in 79 children with hemolytic uremic syndrome; however, the relation was confined to those with a diarrheal prodrome.19 In addition, for patients presenting with diarrhea, elevated polymorphonuclear-leukocyte counts, a shorter duration of diarrheal prodrome, and the presence of bloody diarrhea were associated with severe disease in univariate analysis. Further analysis demonstrated that these three factors were highly correlated. It is possible that these findings reflect more invasive disease or a higher infectious dose of verotoxin-producing E. coli, thus leading to a higher level of circulating toxin (which may be important in the pathogenesis of hemolytic uremic syndrome). Identifying these predictors of disease severity may be useful for clinicians who are considering certain treatments (such as intravenous immune globulin or plasmapheresis).4 , 20
We also found that antimicrobial treatment during the prodromal illness was associated with a mild clinical course and a good outcome, independent of other significant predictive variables. Other authors have reported a detrimental effect of antimicrobial use before hospitalization for cases of hemolytic uremic syndrome associated with infection with shigella21 or with E. coli 0157:H7.9 , 22 Also, laboratory evidence suggests that some antimicrobial agents (such as trimethoprim–sulfamethoxazole) enhance the release of verotoxin from E. coli. 23 No conclusions about antimicrobial use can be drawn from this study, however, because of the small samples involved and the limitations of obtaining such data from reviewing medical records.
Several other clinical findings in this study are of particular interest. First, six children in this series underwent laparotomy before the diagnosis of hemolytic uremic syndrome was made. Although the colitis associated with the syndrome can be impressive, it resolves without intervention in almost all cases. Infection with E. coli 0157:H7 should be considered in the differential diagnosis of any child presenting with bloody diarrhea. Second, even though the case fatality rate in this series was low, our results demonstrate the substantial morbidity associated with the hemolytic uremic syndrome.
In the case–control study, the patients were significantly more likely to attend large day-care centers (more than 50 children) than the controls, suggesting that exposure in the day-care setting may be a risk factor for some children. This hypothesis is compatible with the recent recognition of multiple day-care-center—related outbreaks of E. coli 0157:H7 infection in Minnesota.24 However, attendance at large day-care centers could potentially account for only 16 percent of the cases, suggesting that foodborne transmission of E. coli 0157:H7 is the most important means of infection.
Hamburger is a major vehicle that has been associated with foodborne outbreaks of E. coli 0157:H7 infection,25 26 27 indicating that cattle are an important reservoir for this pathogen. Similar data from the evaluation of sporadic cases also support this hypothesis,28 although it is difficult to obtain reliable food histories of the consumption of inadequately cooked hamburger, and such an association may thus be missed. The increase in the incidence of hemolytic uremic syndrome demonstrated here and the apparent increase in E. coli 0157:H7 infections nationally over the past several years suggest that an epizootic infection may be occurring in the animal reservoir.
In conclusion, hemolytic uremic syndrome and E. coli 0157:H7 infection are emerging as important clinical and public health problems. Hemolytic uremic syndrome is associated with substantial morbidity, and its incidence is increasing, particularly in young children. Prevention of some cases may be possible through early recognition of outbreaks of E. coli 0157:H7 infection and through consumer education (such as teaching adults not to serve raw milk or undercooked hamburger to young children). Hemolytic uremic syndrome is currently a reportable condition in Minnesota; to our knowledge, Washington is the only other state that requires cases to be reported. We believe that other states should consider making it reportable as an index of E. coli 0157:H7 infection to monitor treads in disease occurrence and identify outbreaks in a timely fashion.
We are indebted to Alfred J. Fish, M.D., and Edward A. Belongia, M.D., for their help and support of this study; to Joan H. Rambeck, R.N., M.P.H., and Joseph A. Mariotti, B.A., M.T., for their assistance with review of the medical records; to Ms. Janice E. Wiehle for assistance in the preparation of the manuscript; to Chap T. Le, Ph.D., for statistical advice; and to the medical directors of all the participating hospitals for their assistance in identifying patients and their attending physicians.
Source Information
From the Acute Disease Epidemiology Section, Minnesota Department of Health, 717 Delaware St. S.E., Minneapolis, MN 55440, where reprint requests should be addressed to Dr. MacDonald.
- References
References
1
Gianantonio
CrossRef | Web of Science | Medline2
Gasser
Medline3
Kaplan
Full Text | Web of Science | Medline4
Levin
Medline5
Karmali
CrossRef | Web of Science | Medline6
Spika
CrossRef | Web of Science | Medline7
Outbreak of gastrointestinal disease — Ontario . Can Dis Wkly Rep 1987; 13–2:5–8.
8
Neill
Web of Science | Medline9
Carter
Full Text | Web of Science | Medline10
Griffin
Web of Science | Medline11
Tarr
Web of Science | Medline12
Rogers
Web of Science | Medline13
Kinney
CrossRef | Web of Science | Medline14
Nie NH, Hull CH, Jenkins JG, et al. SPSS: statistical package for the social sciences. 2nd ed. New York: McGraw-Hill, 1975:280–8.
15
Breslow NE, Day NE. Statistical methods in cancer research. Vol. 1. The analysis of case–control studies. Lyon, France: International Agency for Research on Cancer, 1980. (IARC scientific publications no. 32.)
16
Ashkenazi
CrossRef | Web of Science | Medline17
Milford
Web of Science | Medline18
Wells
Web of Science | Medline19
Walters
CrossRef | Web of Science | Medline20
Fong
Web of Science | Medline21
Butler
CrossRef | Web of Science | Medline22
Pavia AT, Nichols CR, Green D, Tauxe RV, Mottice S, Wells J. Hemolytic uremic syndrome following an outbreak of Escherichia coli 0157:H7 infection: leukocytosis as a predictor and sulfonamide exposure as a risk factor. In: Program and abstracts of the 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, October 23–26, 1988. Washington, D.C.: American Society for Microbiology, 1988. abstract.
23
Karch H, Goroncy-Bermes P, Opferkuch W, Kroll H-P, O'Brien A. Subinhibitory concentrations of antibiotics modulate amount of Shiga-like toxin produced by Escherichia coli. In: Adam D, Hahn H, Opferkuch W, eds. The influence of antibiotics on the host-parasite relationship II. Berlin: Springer-Verlag, 1985:239–45.
24
Belongia EA, MacDonald KL, Blaser MT, et al. Outbreak of Escherichia coli 0157:H7 in a day care center: patterns of transmission and control methods. In: Program and abstracts of the 29th Interscience Conference on Antimicrobial Agents and Chemotherapy, Houston, September 17–20, 1989. Washington, D.C.: American Society for Microbiology, 1989:1. abstract.
25
Riley
Full Text | Web of Science | Medline26
Belongia EA, MacDonald KL, White KE, Korlath JA, Lobato M, Osterholm MT. Outbreak of Escherichia coli 0157:H7 colitis associated with precooked hamburgers. In: Program and abstracts of the 29th Interscience Conference on Antimicrobial Agents and Chemotherapy, Houston, September 17–20, 1989. Washington, D.C.: American Society for Microbiology, 1989:1. abstract.
27
Ryan
CrossRef | Web of Science | Medline28
MacDonald
CrossRef | Web of Science | Medline
- Citing Articles (42)
Citing Articles
1
William F. Clark. (2012) Thrombotic Microangiopathy: Current Knowledge and Outcomes With Plasma Exchange. Seminars in Dialysisno-no
CrossRef2
B. Borgatta, N. Kmet-Lunaček, J. Rello. (2012) E. coli O104:H4 outbreak and haemolytic–uraemic syndrome. Medicina Intensiva
CrossRef3
Markus J. Kemper. (2011) Outbreak of hemolytic uremic syndrome caused by E. coli O104:H4 in Germany: a pediatric perspective. Pediatric Nephrology
CrossRef4
Gia Oh, Keith K. Lau. (2011) Characteristics of children with sporadic hemolytic uremic syndrome in a single Northern California center. International Urology and Nephrology
CrossRef5
D Provan, AC Newland, PK MacCallum. 2011. Acquired disorders affecting megakaryocytes and platelets. , 523-545.
CrossRef6
Veit Sturm, Marcel N. Menke, Klara Landau, Guido F. Laube, Thomas J. Neuhaus. (2010) Ocular involvement in paediatric haemolytic uraemic syndrome. Acta Ophthalmologica 88:7, 804-807
CrossRef7
Ajay P Sharma, Guido Filler, Prabo Dwight, William F Clark. (2010) Chronic renal disease is more prevalent in patients with hemolytic uremic syndrome who had a positive history of diarrhea. Kidney International 78:6, 598-604
CrossRef8
Elizabeth J Elliott, Greta F Ridley, Elisabeth M Hodson, Jonathan C Craig, Elizabeth J Elliott. 2010. Interventions for preventing haemolytic uraemic syndrome/thrombotic thrombocytopenic purpura. .
CrossRef9
Mini Michael, Elizabeth J Elliott, Greta F Ridley, Elisabeth M Hodson, Jonathan C Craig, Mini Michael. 2009. Interventions for haemolytic uraemic syndrome and thrombotic thrombocytopenic purpura. .
CrossRef10
Amit X. Garg, Marina Salvadori, Justin M. Okell, Heather R. Thiessen-Philbrook, Rita S. Suri, Guido Filler, Louise Moist, Douglas Matsell, William F. Clark. (2008) Albuminuria and Estimated GFR 5 Years After Escherichia coli O157 Hemolytic Uremic Syndrome: An Update. American Journal of Kidney Diseases 51:3, 435-444
CrossRef11
Saravanan Balamuthusamy, Rohit Arora. (2007) Hematologic Adverse Effects of Clopidogrel. American Journal of Therapeutics 14:1, 106-112
CrossRef12
G. Z. PANOS, G. I. BETSI, M. E. FALAGAS. (2006) Systematic review: are antibiotics detrimental or beneficial for the treatment of patients with Escherichia coli O157:H7 infection?. Alimentary Pharmacology and Therapeutics 24:5, 731-742
CrossRef13
Phillip I Tarr, Carrie A Gordon, Wayne L Chandler. (2005) Shiga-toxin-producing Escherichia coli and haemolytic uraemic syndrome. The Lancet 365:9464, 1073-1086
CrossRef14
David G. Renter, Jan M. Sargeant. (2002) Enterohemorrhagic Escherichia coli O157: epidemiology and ecology in bovine production environments. Animal Health Research Reviews 3:02, 83-94
CrossRef15
Quan Shu, Harsharnjit S Gill. (2002) Immune protection mediated by the probiotic Lactobacillus rhamnosus HN001 (DR20â¢) against Escherichia coli O157:H7 infection in mice. FEMS Immunology & Medical Microbiology 34:1, 59-64
CrossRef16
E Elliott, G Ridley, E Hodson, J Craig, Jonathan Craig. 2002. Interventions for haemolytic uraemic syndrome and thrombotic thrombocytopenic purpura. .
CrossRef17
Elizabeth J Elliott, Greta F Ridley, Elisabeth M Hodson, Jonathan C Craig, Elizabeth J Elliott. 2002. Interventions for preventing haemolytic uraemic syndrome/thrombotic thrombocytopenic purpura. .
CrossRef18
Gary W. Procop. (2001) GASTROINTESTINAL INFECTIONS. Infectious Disease Clinics of North America 15:4, 1073-1108
CrossRef19
Urai Chaisri, Michio Nagata, Hisao Kurazono, Hiroshi Horie, Pongsri Tongtawe, Hideo Hayashi, Teruo Watanabe, Pramuan Tapchaisri, Manas Chongsa-nguan, Wanpen Chaicumpa. (2001) Localization of Shiga toxins of enterohaemorrhagic Escherichia coli in kidneys of paediatric and geriatric patients with fatal haemolytic uraemic syndrome. Microbial Pathogenesis 31:2, 59-67
CrossRef20
FRAN??OIS PROULX, ERNEST G. SEIDMAN, AND, DIANA KARPMAN. (2001) Pathogenesis of Shiga Toxin-Associated Hemolytic Uremic Syndrome. Pediatric Research 50:2, 163-171
CrossRef21
Andre A. Kaplan. (2001) Apheresis for Renal Disease. Therapeutic Apheresis and Dialysis 5:2, 134-141
CrossRef22
Subhas Banerjee, J.Thomas LaMont. (2000) Treatment of gastrointestinal infections. Gastroenterology 118:2, S48-S67
CrossRef23
Kazuaki Yoshimura, Jun Fujii, Hatsumi Taniguchi, Shin-ichi Yoshida. (1999) Chemotherapy for enterohemorrhagic Escherichia coli O157:H infection in a mouse model. FEMS Immunology & Medical Microbiology 26:2, 101-108
CrossRef24
Andre A. Kaplan. (1999) Therapeutic Apheresis for Renal Disorders. Therapeutic Apheresis and Dialysis 3:1, 25-30
CrossRef25
Paul S Mead, Patricia M Griffin. (1998) Escherichia coli O157:H7. The Lancet 352:9135, 1207-1212
CrossRef26
James N. George, Ronald O. Gilcher, James W. Smith, Linda Chandler, Deanna Duvall, Cherrie Ellis. (1998) Thrombotic thrombocytopenic purpura-hemolytic uremic syndrome: Diagnosis and management. Journal of Clinical Apheresis 13:3, 120-125
CrossRef27
Manuel E. Babaian, Dominic C. Chow, Gregory N. Soloway, Howard J. Taubin, Ingram M. Roberts. (1997) Colonic Ischemic Stricture Presenting as a Late Complication of the Hemolytic Uremic Syndrome. Journal of Clinical Gastroenterology 25:3, 555-556
CrossRef28
R PARKER. (1997) ETIOLOGY AND TREATMENT OF ACQUIRED COAGULOPATHIES IN THE CRITICALLY ILL ADULT AND CHILD. Critical Care Clinics 13:3, 591-609
CrossRef29
Cabot, Richard C.Scully, Robert E., Mark, Eugene J., McNeely, William F., Ebeling, Sally H.Phillips, Lucy D., Seifter, Julian L.Nickeleit, Volker. (1997) Case 17-1997. New England Journal of Medicine 336:22, 1587-1594
Full Text30
Tarr, Phillip I., Fouser, Laurie S., Stapleton, Ann E., Wilson, Richard A., Kim, Harold H., Vary, James C. Jr., Clausen, Carla R., . (1996) Hemolytic–Uremic Syndrome in a Six-Year-Old Girl after a Urinary Tract Infection with Shiga-Toxin–Producing Escherichia coli O103:H2. New England Journal of Medicine 335:9, 635-638
Full Text31
Andre A. Kaplan. (1996) Plasma Exchange in Renal Disease. Seminars in Dialysis 9:1, 61-70
CrossRef32
Boyce, Thomas G., Swerdlow, David L., Griffin, Patricia M., . (1995) Escherichia coli O157:H7 and the Hemolytic–Uremic Syndrome. New England Journal of Medicine 333:6, 364-368
Full Text33
C. M. Pennings, H. G. Lenard, R. C. Seitz, H. Karch. (1994) Haemolytic anaemia in association withEscherichia coli 0157 infection in two sisters. European Journal of Pediatrics 153:9, 656-658
CrossRef34
Shunji Akashi, Kohsuke Joh, Takeshi Mori, Atsutoshi Tsuji, Hiroshi Ito, Hiroyuki Hoshi, Takashi Hayakawa, Jiro Ihara, Tsuneyasu Abe, Masayuki Hatori, Taizo Nakamura, S. Akashi. (1994) A severe outbreak of haemorrhagic colitis and haemolytic uraemic syndrome associated withEscherichia coli 0157: H7 in Japan. European Journal of Pediatrics 153:9, 650-655
CrossRef35
Nestor Perez, Ricardo Rahman, Javier Zalba, Norma Bibiloni, Patricia Lasarte, Francisco Spizzirri. (1994) Haemolytic uraemic syndrome and unpasteurized milk. Acta Paediatrica 83:2, 142-142
CrossRef36
Tae Takeda, Hiroshi Nakao, Tatsuhiro Yamanaka, Takashi Igarashi, Yoshifumi Takeda, Noboru Kobayashi. (1993) High prevalence of serum antibodies to Verotoxins 1 and 2 among healthy adults in Japan. Journal of Infection 27:2, 211-213
CrossRef37
A. P. Burnens, P. Boss, F. Ørskov, I. Ørskov, U. B. Schaad, F. Müller, R. Heinzle, J. Nicolet. (1992) Occurrence and phenotypic properties of verotoxin producingEscherichia coli in sporadic cases of gastroenteritis. European Journal of Clinical Microbiology & Infectious Diseases 11:7, 631-634
CrossRef38
Piero Ruggenenti, Giuseppe Remuzzi. (1991) Thrombotic microangiopathies. Critical Reviews in Oncology/Hematology 11:4, 243-265
CrossRef39
I. Kaye Wachsmuth, Patricia M. Griffin, Joy G. Wells. (1991) Escherichia Coli O157: H7, a Cause of Hemorrhagic Colitis and Hemolytic Uremic Syndrome. Pediatrics International 33:5, 603-612
CrossRef40
Guerrant, Richard L., Bobak, David A., . (1991) Bacterial and Protozoal Gastroenteritis. New England Journal of Medicine 325:5, 327-340
Full Text41
Cabot, Richard C.Scully, Robert E., Mark, Eugene J., McNeely, William F., McNeely, Betty U., Kuter, David J.Ellman, LeonardMcCluskey, Robert T.. (1991) Case 30-1991. New England Journal of Medicine 325:4, 265-273
Full Text42
(1991) Epidemiology of the Hemolytic Uremic Syndrome. New England Journal of Medicine 324:15, 1065-1066
Full Text
- Letters
