- HUS =
- hemolytic uremic syndrome •
- ECH =
- Egleston Children's Hospital •
- HSCH =
- Hughes Spalding Children's Hospital •
- SRCH =
- Scottish Rite Children's Hospital •
- CDC =
- Centers for Disease Control and Prevention •
- CT =
- computed tomography •
- BUN =
- blood urea nitrogen •
- PT =
- prothrombin time •
- PTT =
- partial thromboplastin time •
- CSF =
- cerebrospinal fluid •
- DIC =
- disseminated intravascular coagulopathy
Hemolytic uremic syndrome (HUS), the most common cause of acute renal failure in childhood, is characterized by acute renal failure, microangiopathic hemolytic anemia, and thrombocytopenia. The majority of HUS cases occur after infectious diarrhea, and most of these cases are associated with Escherichia coli O157:H7 infection.1 2 However, atypical cases of HUS also occur in the absence of infectious diarrhea, although less commonly.3 Among 117 children <18 years of age identified with HUS in the state of Minnesota from 1979 through 1988, 16 patients had no preceding diarrheal illness, and 6 had a respiratory illness prodrome.4 Invasive infection with Streptococcus pneumoniae has rarely been associated with atypical HUS cases.5-14 However, there are no published data on the prevalence of invasive pneumococcal infections among patients with HUS or on the prevalence of HUS cases associated with invasive S pneumoniae infection.
We report 7 cases of S pneumoniae-associated HUS managed at three Atlanta children's hospitals over a period of 3 years. Recognition of these cases encouraged us to review our recent experience and determine the frequency of S pneumoniae as a cause of HUS at our institutions. An ongoing pneumococcal disease surveillance project in the metropolitan Atlanta area enabled us to determine the incidence of HUS associated with systemic pneumococcal infection.
MATERIALS AND METHODS
Medical records from the three children's hospitals in the metropolitan Atlanta area (Egleston Children's Hospital [ECH], Hughes Spalding Children's Hospital [HSCH], and Scottish Rite Children's Hospital [SRCH]) were searched for all patients with HUS from January 1994 to December 1996 usingInternational Classification of Diseases, Ninth Revisioncoding for nonimmune hemolytic anemia (283.1) and HUS (283.11). Case databases of pediatric nephrologists serving these three hospitals were also reviewed to determine the existence of HUS cases not captured by this chart search. Medical records of all patients were reviewed to determine if illness met criteria for HUS: acute onset anemia with microangiopathic changes (ie, schistocytes, burr cells, or helmet cells) on peripheral blood smear and renal injury evidenced by either hematuria, proteinuria, or elevated creatinine level (ie, 1.0 mg/dL in a child aged <13 years or 50% increase over baseline).16Cases were classified as associated with invasive pneumococcal infection if S pneumoniae was isolated from a normally sterile body fluid within 1 week before or after the onset of signs of HUS.
Pneumococcal isolates from these patients were serotyped and antimicrobial susceptibility testing was performed as part of a prospective, laboratory-based surveillance for invasive pneumococcal infections that has been conducted in the eight-county metropolitan Atlanta area (population 2.34 million) since January 1994.17 Microbiology laboratories of all 28 acute care hospitals and one major reference laboratory in the metropolitan area submit all pneumococcal isolates obtained from normally sterile sites to the laboratories of the Centers for Disease Control and Prevention (CDC). For patients who lived outside of the surveillance area, isolates were obtained, if available, from hospital laboratories where the patient was initially evaluated. Submitted isolates were confirmed to be pneumococci at CDC by susceptibility to optochin and bile solubility. Isolates available to the CDC were serotyped on the basis of capsular swelling with type-specific antisera (quellung reaction).18 Antimicrobial susceptibility testing was performed by the broth microdilution method17 for isolates obtained by the CDC. Minimal inhibitory concentrations from isolates not available to the CDC were obtained from the hospital laboratories where isolates were initially identified.
Statistical analysis was performed using standard statistical software (Sigma-Stat, Jandel Scientific, San Rafael, CA, and Epi-Info Version 6, CDC, Atlanta, GA). Patient data did not satisfy criteria for a normal population and were therefore analyzed using nonparametric techniques.
Seven patients with HUS associated with invasive pneumococcal infection were admitted to the three metropolitan Atlanta children's hospitals during 1994 to 1996 (Table 1). The median age for these patients was 16 months, and all were <24 months old. Two patients had pneumococcal meningitis, and 5 had clinical and radiographic evidence of pneumonia. Three of the 5 patients with pneumonia had an associated empyema. Three of the patients manifested all components of HUS at the time of admission, while 4 progressed to HUS within 72 hours. None had underlying conditions associated with increased risk of invasive pneumococcal infection (eg, sickle cell disease, splenectomy). Peritoneal dialysis was used for management of acute renal failure in all patients. Continuous veno-venous hemofiltration was used initially in 1 patient who later was managed with peritoneal dialysis. Median duration of dialysis was 25 days. Two patients died and 1 patient survived with severe neurologic sequelae (see “Selected Case Reports”). Of the 5 survivors, none required chronic dialysis or renal transplant.
Of 5 isolates available for serotyping, 2 were serotype 14, 2 were 23F, and 1 was 6B (Table 2). Two of the 7 patients were infected with strains which were resistant to penicillin and cefotaxime (minimal inhibitory concentrations ≥2 μg/mL). Four of 7 patients received oral antibiotic therapy before admission, and 3 of these 4 patients were infected with pneumococci resistant to their oral antibiotic regimens (Table 2). Both patients with meningitis had cephalosporin-resistant strains of S pneumoniae; 1 of these patients died and the other had significant neurologic impairment after recovery. Four patients lived within the metropolitan Atlanta surveillance area. These cases represented 0.6% of the 618 invasive pneumococcal infections identified among children <2 years old during 1994 to 1996. The annual incidence of HUS associated with invasive pneumococcal infection based on the CDC surveillance data during this period in Atlanta was 1.7 cases per 100 000 children <2 years old.
Selected Case Reports
A 5-month-old black female, previously healthy, developed an upper respiratory infection and fever, lethargy, and loose stools. In the emergency room, she was obtunded with grunting respirations, pallor, and tachycardia that rapidly progressed to cardiopulmonary arrest requiring extensive resuscitative measures and inotropic support. A chest radiograph showed complete opacification of the left hemithorax, and a chest tube was placed for treatment of empyema. Lumbar puncture was unremarkable. Antibiotics were initiated. Stool culture was negative for pathogens including E coli O157:H7. Blood cultures subsequently grew S pneumoniae (Table 2).
She developed anuria on the second day of admission with a blood urea nitrogen 66 mg/dL and creatinine 2.4 mg/dL. HUS was diagnosed based on findings of microangiopathic hemolytic anemia, thrombocytopenia, and renal failure. A Tenckhoff catheter was placed and peritoneal dialysis was started. She developed severe neurologic injury attributed to her initial arrest. Computed tomography (CT) revealed diffuse cerebral atrophy suggestive of hypoxic ischemic encephalopathy. The patient was transferred to a hospital near her home in Florida for further management with persistent requirement for peritoneal dialysis. She died 1 week later. No autopsy was performed.
A 13-month-old white male with otitis media and an upper respiratory infection was hospitalized for increased lethargy and persistent fever. Initial platelet count, blood urea nitrogen (BUN), creatinine, prothrombin time (PT), and partial thromboplastin time (PTT) were normal. A lumbar puncture revealed a cerebrospinal fluid (CSF) glucose of 29 mg/dL, protein of 535 mg/dL, and CSF pleocytosis with large numbers of Gram-positive diplococci on gram stain. Upon transfer to ECH, the patient was tachypneic and had a nonconjugate gaze. He was intubated and received antibiotics. Cranial CT was normal. Both CSF and blood cultures were subsequently positive for S pneumoniae (Table 2).
Three days after admission he developed oliguria. BUN increased to 103 mg/dL and creatinine to 2.1 mg/dL. Hemolytic anemia with schistocytosis and a markedly decreased platelet count (11 000/mm3) ensued. PT and PTT remained normal. A catheter was placed for peritoneal dialysis. Oliguric renal failure and thrombocytopenia persisted despite therapy. The patient developed seizures, mental status changes, and ventricular tachycardia. Cranial CT showed a large subdural hematoma that required craniotomy. Postoperatively, he developed severely increased intracranial pressures unresponsive to multiple measures including pentobarbital coma. Nine days after admission the patient was declared brain dead and life support was discontinued.
A 16-month-old white male developed persistent rhinorrhea and otitis media that did not respond to cefprozil, and he was later admitted to a local hospital with fever and lethargy. A lumbar puncture showed CSF pleocytosis, glucose <2 mg/dL, protein 542 mg/dL, and Gram-positive diplococci. Initial PT was 14.2, PTT was 47.6, fibrinogen was 565 mg/dL and D-dimers were present at 8000 to 16 000 ng/dL. BUN was 20 mg/dL and creatinine 0.2 mg/dL. Antibiotics were given. A CT scan showed mild to moderate hydrocephalus and left mastoid and ethmoid opacification. He was intubated on transfer to the ECH intensive care unit due to progressive neurologic deterioration. He underwent a left mastoidectomy and placement of a tympanostomy tube. Gross pus and clinical evidence of osteomyelitis were noted at surgery. Blood cultures were positive for highly cephalosporin resistant S pneumoniae (Table 2). Two days after admission, he developed acute renal failure with azotemia, hemolytic anemia, and thrombocytopenia, and required peritoneal dialysis for 21 days. He was discharged with significant neurologic deficits, but was alert and active.
Comparison of HUS Cases With and Without Pneumococcal Infection
During 1994 through 1996, 30 children with illness meeting CDC case definition criteria for HUS16 were admitted to the three children's hospitals in Atlanta. Thus, the 7 patients with HUS associated with invasive pneumococcal infection represented 23% of all the HUS cases. All 23 patients without pneumococcal infection had some form of diarrhea, and 18 of 23 patients had bloody diarrhea. Stool was not routinely collected and cultured for E coli O157:H7, and this pathogen was only isolated from 1 patient. HUS was associated with cyclosporine or ganciclovir toxicity in 2 patients. Although 14 (47%) of all HUS cases occurred during the summer months (June through August), cases with and without pneumococcal infection occurred year-round (Figure).
Children with S pneumoniae-associated HUS were significantly younger than the other HUS patients (16 months compared with 5 years,P < .0001 by Mann-Whitney rank sum test). Although all children with HUS associated with pneumococcal infection required dialysis, only 11 (48%) of those with HUS due to other conditions required dialysis (Fisher exact P = .03). One patient with acute lymphocytic leukemia after a bone marrow transplant for non-Hodgkin's lymphoma developed HUS associated with cyclosporine use and died. However, the difference in survival between patients withS pneumoniae-associated HUS (71%) and other patients with HUS (96%) was not statistically significant (Fisher's exact test,P = .12). The median length of stay of the S pneumoniae-associated cases of HUS was 41 days compared to the other HUS cases which was 12 (Mann-Whitney rank sum testP = .008). These differences persisted when nondiarrheal HUS cases were excluded.
These 7 cases represent the largest case series of HUS associated with invasive S pneumoniae infection reported to date. Based on our findings, factors that may help identify HUS patients with invasive pneumococcal infection include age <2 years and illness requiring dialysis for management of renal insufficiency. Pneumococcal infection-associated HUS occurred in all seasons in Atlanta. Cases ofE coli O157:H7 infections with and without HUS tend to occur during the summer months19 20; therefore, HUS occurring during the winter months in a child with no history suggestive of diarrhea-associated HUS should increase suspicion of invasive pneumococcal infection. Our finding of 3 patients with HUS and pneumococcal empyema was surprising, and the significance of this observation is not apparent. Empyema among children with pneumonia is unusual, although an apparent increase in pneumococcal empyema cases during the early 1990s has been recently reported from a children's hospital in Cincinnati.21
The pathogenic mechanism of HUS caused by S pneumoniae has been proposed by Novak and Martin.10 Neuraminidase produced by pneumococci remove N-acetylneuraminic acid from cell membrane surfaces exposing the Thomsen-Friedenreich antigen, present on erythrocytes, platelets, and glomerular capillary walls11 15 to circulating blood elements. Antigen-antibody interaction causes damage to the surface of cells expressing this antigen, resulting in hemolysis, thrombocytopenia, and renal microangiopathy and the clinical manifestations of HUS. All clinical isolates of S pneumoniae produce neuraminidase, and indirect evidence suggests that pneumococcal neuraminidase may play a role in the pathogenesis of invasive infection.22 Among patients with pneumococcal meningitis, elevated levels of freeN-acetylneuraminic acid were associated with both coma and bacteremia.23
Three pneumococcal serotypes were represented among our patients with HUS (6B, 14, 23F). These serotypes are common among children with invasive pneumococcal disease, accounting for nearly half of all blood and CSF isolates among preschool aged children in the United States.24 All 7 patients in our report and almost all children with pneumococci-associated HUS previously reported in the medical literature were <2 years old.5-15 The incidence of invasive pneumococcal infection is highest among children <2 years old25; however, age-related variability in antibody production or antigen expression may also contribute to tendency for pneumococcal associated HUS to occur in very young children. Only 2 (28%) of our HUS patients had penicillin- and cephalosporin-resistant strains; this is similar to the prevalence of resistant strains among all young children with invasive pneumococcal disease in metropolitan Atlanta.26 Although the number of cases is small, our experience suggests that antibiotic resistance does not necessarily correlate with development of HUS.
The treatment of S pneumoniae-associated HUS is aimed at supportive care and treatment of the underlying infection. Transfusion with blood products containing plasma can theoretically worsen the manifestations of HUS due to the presence of plasma anti-Thomsen-Friedenreich antigen IgM; therefore, when tranfusion is indicated, use of washed blood products has been recommended.10 Prior antibiotic use inShigella-induced HUS has been suggested to play a role in the progression of enteritis to HUS by causing a release ofShigella endotoxin from dying bacteria.27 Our data are not sufficient to judge whether antibiotic therapy plays a role in pneumococcal-associated HUS.
Renal function in our patients recovered with supportive therapy including dialysis during anuria. Previous reports28-30have evaluated the long-term outcome of renal function with HUS associated with E coli O157:H7. Seigler et al28found that duration of anuria was the best predictor of disease (hypertension, proteinuria, or low creatinine clearance) at follow-up. The majority of the patients with E coli O157:H7 associated HUS recovered without sequelae. The authors also suggested that renal function should be periodically evaluated for an extended period of time. No data exist regarding long-term renal function in HUS associated with S pneumoniae.
Patients with HUS associated with pneumococcal infection were more likely to undergo dialysis and had longer hospital stays than other patients with HUS, suggesting that S pneumoniae causes a more severe form of the illness. The relatively high prevalence ofS pneumoniae infections among patients evaluated for HUS at the three Atlanta children's hospitals (23%) may reflect referral patterns for more critically ill patients. The actual incidence of invasive pneumococcal infections among children with HUS in metropolitan Atlanta may have been greater if blood cultures were not obtained or were delayed until after initiation of antimicrobial therapy in HUS patients for whom systemic bacterial infection was not initially suspected.
Our method of case collection was likely to detect more severe HUS cases and could have missed milder HUS cases not identified or not requiring consultation. Although some children are admitted to other hospitals in the metropolitan Atlanta area, pediatric nephrologists only admit patients to these three children's hospitals, and it is unlikely that children with HUS would be admitted to other facilities. It is also possible that some of these cases actually represented acute renal failure associated with pneumococcal septic shock and disseminated intravascular coagulopathy (DIC); however, several findings suggest otherwise. All patients met CDC case definition criteria for HUS, and each patient demonstrated schistocytosis. These patients did not consistently demonstrate other findings of DIC such as elevated D-dimers, PT, PTT, or decreased fibrinogen levels. Most of the patients did not manifest findings of shock severe enough to induce renal failure. The observed thrombocytopenia was prolonged, in contrast to its relatively short-lived occurrence in typical cases of DIC. Finally, progression of anuric renal failure requiring dialysis in every patient would seem more consistent with HUS than with acute tubular necrosis associated with sepsis.
In conclusion, HUS is an uncommon complication of invasive S pneumoniae infection. Future surveillance for HUS should include collection of data to determine the true prevalence of invasive pneumococcal infection among these patients. Clinicians managing patients with invasive pneumococcal disease should maintain vigilance for the development of HUS and the possibility of altered drug clearance resulting from acute renal failure. This is particularly relevant in light of the emergence of drug-resistant pneumococcal strains and recent recommendations that vancomycin be included in empiric therapy for suspected pneumococcal meningitis and invasive infections in critically ill patients.31
We thank the physians and nursing staff of each children's hospital for their meticulous care of these patients; Dr Monica Farley and Ms Wendy Baughman for directing the metropolitan Atlanta pneumococcal surveillance; Dr John Elliott, Dr Richard Facklam, Alma Franklin, and Delois Jackson for serotyping and performing susceptibility testing of isolates; Dr Margarette Kolczak and Ms Katherine Deaver-Robinson for assistance with database management; and the Georgia Department of Human Resources Epidemiology and Prevention Branch for its support of the active surveillance program.
- Received June 4, 1997.
- Accepted August 28, 1997.
Reprint requests to (G.R.C.) Division of Critical Care, Egleston Children's Hospital, 1405 Clifton Road NE, Atlanta, GA 30322.
- Siegler RL
- ↵Centers for Disease Control and Prevention. Case definitions for infectious conditions under national public health surveillance. MMWR. 1997;46(RR-10):17
- ↵Facklam RR, Washington JA II. Streptococcus and related catalase-negative gram-positive cocci. In: Balows A, Hausler WJ Jr, Herrmann KL, Isenberg HD, Shadomy HJ, eds. Manual of Clinical Microbiology. 5th ed. Washington, DC: American Society for Microbiology; 1991:238–257
- ↵Butler JC, Breiman RF, Lipman HB, Hofmann J, Facklam RR. Serotype distribution of Streptococcus pneumoniae infections among preschool children in the United States, 1978–1994: implications for development of a conjugate vaccine. J Infect Dis. 1995;171:171:885–889
- ↵Cetron MS, Breiman RF, Jorgensen JH et al. Geographic variation of drug-resistant Streptococcus pneumoniae: population-based surveillance from nine locations, 1995–1996. In: Abstracts of the 97th General Meeting of the American Society for Microbiology; May 4–8, 1997; Miami, FL. Washington, DC: American Society for Microbiology; C-371a, 185
- Siegler RL,
- Pavia AT,
- Christofferson RD,
- Milligan MK
- American Academy of Pediatrics, Committee on Infectious Diseases
- Copyright © 1998 American Academy of Pediatrics