- Email: [email protected]
Renal prognosis of toxigenic Escherichia coli infection
review
http://www.kidney-international.org & 2009 International Society of Nephrology
Renal prognosis of toxigenic Escherichia coli infection Amit X. Garg1,2, Marina Salvadori3, Louise M. Moist1,2, Rita S. Suri1,2 and William F. Clark1,2 1
Division of Nephrology, Department of Medicine, University of Western Ontario, London, Ontario, Canada; 2Department of Epidemiology and Biostatistics, University of Western Ontario, London, Ontario, Canada and 3Division of Infectious Diseases, Department of Pediatrics, University of Western Ontario, London, Ontario, Canada
There is a spectrum of acute illness for individuals who ingest toxigenic Escherichia coli, ranging from no symptoms to self-limited gastroenteritis, to full acute hemolytic uremic syndrome. Knowledge of the long-term prognosis of acute toxigenic E. coli ingestion, based on the presentation of acute illness, is important for patient counseling and follow-up. Here, we consider subsequent 5-year renal outcomes for groups of individuals who presented across a spectrum of acute illness during a municipal water outbreak of E. coli O157:H7 in Walkerton, Canada. Kidney International (2009) 75 (Suppl 112), S38–S41; doi:10.1038/ki.2008.617 KEYWORDS: HUS; E. coli O157:H7; renal failure; hypertension; proteinuria
Correspondence: Amit X. Garg, Kidney Clinical Research Unit, Department of Medicine and Epidemiology, University of Western Ontario, Room ELL-101, Westminster Tower, London Health Sciences Centre, 800 Commissioners Road East, London, Ontario, Canada N6A 4G5. E-mail: [email protected] S38
Many strains of toxigenic Escherichia coli are implicated in diseases, with E. coli O157:H7 being the most common. The main source of infection is animals, particularly cattle, to which these bacteria are non-pathogenic. Bacteria residing in the gastrointestinal tract of animals are shed in the feces. Although contaminated meat is a major source of transmission, situational outbreaks of contaminated lettuce, alfalfa sprouts, unpasteurized milk, and apple cider have also lead to extensive morbidity.1 In the case of Walkerton, municipal water became contaminated with E. coli O157:H7. At the time of the outbreak, heavy rainfall contributed to the transport of livestock fecal contaminants into inadequately chlorinated drinking water, supplied from a shallow well. There is a spectrum of acute illness for the person who ingests toxigenic E. coli. This ranges from no symptoms to self-limited gastroenteritis, to full acute hemolytic uremic syndrome (HUS) (Figure 1). At the time of the outbreak, over 2300 people became ill with acute gastroenteritis, and there were 28 cases of HUS. Who becomes ill after toxigenic E. coli ingestion is influenced by a number of factors, including the amount and type of bacteria, and a patient’s predisposition to illness. When caring for someone with acute kidney injury from this infection, the focus is on adequate hydration, avoidance of nephrotoxins, and dialysis when it is required. Better awareness and supportive care result in most patients surviving the acute illness. Serum creatinine values frequently return to the laboratory reference range. However, acute kidney injury across many settings, even when transient, is increasingly recognized as having important long-term consequences.2 Similar to other renal insults, it has been suggested that HUS can lead to a critical reduction in nephron number, with unsustainable remnant single-nephron hyperfiltration and progressive renal disease.3 This then raises the question: Are patients who survive acute toxigenic E. coli infections at risk of longer term renal sequelae? Here, we consider subsequent renal outcomes for groups of individuals who presented across the spectrum of acute illness during the Walkerton outbreak.4–9 Participants were followed for a period of 5 years after the outbreak. E. coli O157:H7 INGESTION WITHOUT ANY SYMPTOMS
Over 95% participants in the Walkerton Health Study, or a thousand individuals, ingested contaminated municipal water without any gastrointestinal or other symptoms at Kidney International (2009) 75 (Suppl 112), S38–S41
review
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
1
3
2
Figure 1 | Spectrum of acute illness for individuals who ingest toxigenic Escherichia coli. After ingestion there may be a spectrum of acute illness from (1) no symptoms to (2) symptoms of selflimited gastroenteritis, to (3) full hemolytic uremic syndrome.
the time of the outbreak. Of course, when a person has no symptoms, one is never sure whether toxigenic E. coli were even ingested. Nonetheless, even if ingested, it seems biologically implausible that an individual without any symptoms of acute illness has appreciable silent renal injury which puts them at risk of future chronic sequelae. Consistent with this assertion, follow-up results for individuals without any symptoms at the time of the outbreak were similar to population reference standards.10 For this reason, participants without any symptoms at the time of the outbreak were treated as a reference or the control group, to which other groups with acute symptoms were compared. E. coli O157:H7 INGESTION WITH HUS
Of the 28 childhood cases of HUS during the outbreak, 1 child had proteinuria, which predated the outbreak, and 1 died at the time of the outbreak. Of the remaining 26 children, 7 were lost to follow-up and 19 participated in the Walkerton Health Study. Two physicians independently confirmed the presence of HUS during the outbreak. Briefly, in the absence of another medical explanation, children with all the following criteria at the time of the outbreak were defined as having HUS: (1) hemoglobin p10 g/100 ml (100 g/l) with evidence of destruction of erythrocytes on blood smear, (2) platelets p150 109/l, and (3) serum creatinine X95th percentile for age and sex. All these children with HUS were known to be previously healthy, Kidney International (2009) 75 (Suppl 112), S38–S41
without medical conditions (such as diabetes, hypertension, renal disease, recurrent infections, or failure to thrive) predating the outbreak. At the time of acute illness, 16 (84%) had stools positive for E. coli O157:H7 bacteria or verotoxin. The average lowest hemoglobin was 6.1 g/100 ml (s.d. 1.4, range 4.5–9.5) (61 g/l, range 45–95) and some patients were subsequently transfused, the average lowest platelet count was 56 109/l (s.d. 38, range 12–150), and the average highest serum creatinine was 2.8 mg/100 ml (s.d. 2.8, range 0.5–9.7) (246 mmol/l, s.d. 248, range 42–861). No child developed convulsions, coma, stroke, or severe colitis requiring surgery. Seven anuric children were treated with peritoneal dialysis for an average of 14 days (range 10–19). During the acute illness, one child was treated with antihypertensive therapy. All 19 children recovered in the months after acute HUS; none required long-term dialysis and all had a serum creatinine that returned to the lab-defined normal range. Two of these children were treated with antihypertensive therapy (an angiotensin-converting enzyme inhibitor) for variable amounts of time in follow-up; one child for the first 3 years, whereas another child remained on therapy until the 5-year assessment. What is the long-term renal prognosis of these children? An accurate estimate of prognosis remains critical for patient counseling, follow-up, and monitoring. To guide this estimate and data collection in the Walkerton Health Study, we conducted a systematic review of 49 studies of 3476 children with a mean follow-up of 4.4 years (range 1–22) from 18 countries, 1950–2001.11 The key finding of this review was that the long-term renal prognosis of diarrheaassociated HUS was highly variable. No definitive conclusion about a precise estimate could be drawn. The proportion of HUS survivors with renal sequelae in follow-up, defined as any one of a glomerular filtration rate (GFR) o80 ml/min per 1.73 m2, hypertension, or proteinuria, ranged anywhere from 0 to 64%. Some studies cited minimal sequelae in the 5 years after HUS,12–15 whereas other studies described survivors with hypertension, proteinuria, or a reduced GFR;16–20 in some cases, large proportions of children presented with such sequelae.21–25 Interpreting the studies remains a challenge because toxigenic E. coli was often not established as the cause of HUS, and many times large proportions of children were lost to follow-up. Different definitions were used for outcomes of proteinuria, GFR, and hypertension, and measurements were performed at variable times of follow-up. The majority of studies did not state whether observed sequelae were persistent after acute HUS, or whether they developed after apparent recovery. Finally, earlier studies did not use a healthy control group to establish outcomes attributable to HUS. Fortunately, when we followed the outcomes of children from Walkerton who recovered from HUS for 5 years after the outbreak, the results were very encouraging.9 Although more children who recovered from HUS presented with microalbuminuria than controls (20 vs 3%; relative risk: 6.0 (95% CI: 1.1–32.8)), no child had evidence of overt S39
review
proteinuria. There were no differences between donors and controls in blood pressure, or GFR when it was estimated with serum creatinine. GFR estimated with cystatin C was lower after HUS compared to controls (100 vs 110 ml/min per 1.73 m2, P ¼ 0.02), but no child had a GFR o80 ml/min per 1.73 m2. Reasons for differences between our results and other studies are likely multifactorial, and may relate to differences in the severity of acute disease, advances in health care, and possibly the use of angiotensin-converting enzyme inhibitor therapy in two children during follow-up. Unlike many other cohorts, serum creatinine values of HUS children in our study returned to the laboratory-defined normal reference range soon after their acute illness. E. coli O157:H7 INGESTION WITH SELF-LIMITED GASTROENTERITIS
No study had previously econsidered whether bacterial gastroenteritis, in the absence of recognized HUS, leads to clinically important long-term renal sequelae. We examined the outcomes of adults and children in the Walkerton Health Study separately, given that sequelae such as hypertension are defined in a very different manner among these two groups. The 1958 adults with no known history of hypertension before the outbreak were considered first.7 Of the participants, 675 were asymptomatic during the outbreak, 909 had moderate symptoms of acute self-limited gastroenteritis, and 374 had severe symptoms, which necessitated medical attention. Cultures were submitted in 40% of those with severe gastroenteritis, of which 14% were positive for E. coli O157:H7. Compared with those with moderate symptoms, participants with severe symptoms were more likely to describe bloody diarrhea (18 vs 40%), prolonged diarrhea (53 vs 70%), and fever (33 vs 50%). Four years after the outbreak, hypertension was diagnosed in 27.0% of those who were asymptomatic, and 32.3 and 35.9% of those with moderate and severe acute gastroenteritis, respectively (trend P ¼ 0.009). Compared with asymptomatic participants, the adjusted relative risk of hypertension was 1.15 (95% CI: 0.97–1.35) in those with moderate gastroenteritis, and 1.28 (95% CI: 1.04–1.56) in those with severe gastroenteritis. A similar graded association was observed for developing reduced kidney function, defined by the presence of an estimated GFR below 60 ml/min per 1.73m2 (trend P ¼ 0.03). No association was observed between gastroenteritis and the subsequent risk of albuminuria. In summary, acute bacterial gastroenteritis necessitating medical attention was associated with a higher risk of renal sequelae 4 years later. We then considered the renal outcomes of children and adolescents.8 Similar to adults, these outcomes were high blood pressure (defined by being greater than the 95th percentile expected for age, sex, and height), reduced kidney function, and microalbuminuria. Of the 951 participants, 313 were asymptomatic during the outbreak, 305 had moderate symptoms of acute gastroenteritis, and 333 had severe S40
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
symptoms, which necessitated medical attention. Compared with those with moderate symptoms, participants with severe symptoms were more likely to describe bloody diarrhea (17 vs 38%), prolonged diarrhea (51 vs 73%), and fever (45 vs 66%). Of those with severe symptoms during the outbreak, 87% described abdominal pain, 69% visited an emergency room, and 3% were admitted to hospital. A total of 70 participants with severe symptoms had a stool positive for E. coli O157:H7 (bacteria or Shiga toxin). Four years after the outbreak, there were no differences in mean systolic blood pressure between those who had no, moderate, or severe symptoms of acute gastroenteritis during the outbreak (109, 110, and 107 mm Hg). Similarly, there were no group differences in mean diastolic blood pressure, estimated GFR, or random urine albumin-to-creatinine ratio, or in the adjusted relative risk of high blood pressure, a GFR o80 ml/min per 1.73 m2, or microalbuminuria. In other words, patients who presented to medical attention with gastroenteritis during this E. coli O157:H7 outbreak had an absence of renal sequelae 4 years later. So, why did the outcomes between adults and children differ? It might be hypothesized that adults diagnosed with gastroenteritis had more severe disease than children. However, both groups with gastroenteritis had comparable presenting symptoms and rates of hospital admission, making it less likely that this was an important mitigating factor. Is it possible that the adult analyses were partly confounded? Unlike children, some adult participants may have had unmeasured co-morbidities predating the outbreak, which made them more susceptible to an acute gastrointestinal infection during the outbreak. Such individuals could then show higher rates of renal disease in follow-up, exaggerating any true association between acute gastroenteritis and subsequent sequelae. Finally, it remains conceivable that children, unlike adults, have more renal reserve, and were able to compensate after an insult of toxigenic E. coli more readily than adults. If this is true, longer follow-up of this group of children will elucidate chronic renal sequelae of E. coli gastroenteritis that were not evident 4 years after the outbreak. In conclusion, large water-borne infectious outbreaks are extremely rare within Western nations. The support of the local community for the Walkerton Health Study allowed us to make important observations about long-term renal sequelae. Time will tell how the results truly play out. For example, ongoing follow-up of the group of children with HUS will clarify the clinical relevance of microalbuminuria and mild decrements in GFR 5 years after recovery. Finally, it remains to be seen whether the observed prognosis ultimately generalizes to other patients affected by toxigenic E. coli infection. DISCLOSURE
LMM has received consulting fees from Amgen. WFC has received lecture fees from Amgen and AstraZeneca and is the recipient of research grant support from the Ontario Ministry of Health, the Canadian Institute of Health Research, and the Kidney Foundation of Canada. The remaining authors have declared no financial interests. Kidney International (2009) 75 (Suppl 112), S38–S41
review
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
ACKNOWLEDGMENTS
This research was funded by the Kidney Foundation of Canada and the Ontario Ministry of Health. Dr Garg was supported by Clinician Scientist Award from the Canadian Institutes of Health Research.
11.
12. 13.
REFERENCES 1. Khanna R, Waechter L, Sargeant J et al. Environmental prevention of human disease from verocytotoxin-producing Escherichia coli. Nephrol Dial Transplant 2008; 23: 1819–1822. 2. Liano F, Felipe C, Tenorio MT et al. Long-term outcome of acute tubular necrosis: a contribution to its natural history. Kidney Int 2007; 71: 679–686. 3. Brenner BM, Lawler EV, Mackenzie HS. The hyperfiltration theory: a paradigm shift in nephrology. Kidney Int 1996; 49: 1774–1777. 4. Garg AX, Clark WF, Salvadori M et al. Microalbuminuria three years after recovery from Escherichia coli O157 hemolytic uremic syndrome due to municipal water contamination. Kidney Int 2005; 67: 1476–1482. 5. Garg AX, Macnab J, Clark W et al. Long-term health sequelae following E. coli and campylobacter contamination of municipal water. Population sampling and assessing non-participation biases. Can J Public Health 2005; 96: 125–130. 6. Garg AX, Marshall J, Salvadori M, et al., Walkerton Health Study Investigators. A gradient of acute gastroenteritis was characterized, to assess risk of long-term health sequelae after drinking bacterial-contaminated water. J Clin Epidemiol 2006; 59: 421–428. 7. Garg AX, Moist L, Matsell D et al. Risk of hypertension and reduced kidney function after acute gastroenteritis from bacteria-contaminated drinking water. CMAJ 2005; 173: 261–268. 8. Garg AX, Clark WF, Salvadori M et al. Absence of renal sequelae after childhood Escherichia coli O157:H7 gastroenteritis. Kidney Int 2006; 70: 807–812. 9. Garg AX, Salvadori M, Okell JM et al. Albuminuria and estimated GFR 5 years after Escherichia coli O157 hemolytic uremic syndrome: an update. Am J Kidney Dis 2008; 51: 435–444. 10. Jones CA, Francis ME, Eberhardt MS et al. Microalbuminuria in the US population: third National Health and Nutrition Examination Survey. Am J Kidney Dis 2002; 39: 445–459.
Kidney International (2009) 75 (Suppl 112), S38–S41
14.
15.
16. 17. 18.
19.
20.
21. 22. 23. 24.
25.
Garg AX, Suri RS, Barrowman N et al. Long-term renal prognosis of diarrhea-associated hemolytic uremic syndrome: a systematic review, meta-analysis, and meta-regression. JAMA 2003; 290: 1360–1370. Ogborn MR, Hamiwka L, Orrbine E et al. Renal function in Inuit survivors of epidemic hemolytic-uremic syndrome. Pediatr Nephrol 1998; 12: 485–488. McLean MM, Jones CH, Sutherland DA. Haemolytic-uraemic syndrome. A report of an outbreak. Arch Dis Child 1966; 41: 76–81. Brichard B, Vermylen C, Heremans MT et al. Plasma infusion as treatment for 33 children with haemolytic uraemic syndrome: a good therapy? Acta Clin Belg 1993; 48: 156–163. Vermylen C, Bodart E, Ninane J et al. The haemolytic uraemic syndrome of childhood: is living in Belgium a factor of good prognosis? Acta Clin Belg 1988; 43: 101–108. Miyazaki S. The etiology and clinical features of hemolytic uremic syndrome. Rinsho Ketsueki 1994; 35: 341–345. Riella MC, George CR, Hickman RO et al. Renal microangiopathy of the hemolytic-uremic syndrome in childhood. Nephron 1976; 17: 188–203. Guyot C, Daniel MD, Roze JC et al. Prognosis of hemolytic uremic syndrome in children. Importance of extrarenal involvement. Arch Fr Pediatr 1986; 43: 253–258. Rizzoni G, Claris-Appiani A, Edefonti A et al. Plasma infusion for hemolytic-uremic syndrome in children: results of a multicenter controlled trial. J Pediatr 1988; 112: 284–290. Brandt JR, Joseph MW, Fouser LS et al. Cholelithiasis following Escherichia coli O157:H7-associated hemolytic uremic syndrome. Pediatr Nephrol 1998; 12: 222–225. Sorrenti LY, Lewy PR. The hemolytic-uremic syndrome: experience at a center in the Midwest. Am J Dis Child 1978; 132: 59–62. Campos A, Kim Y, Miller K et al. Radionuclide studies of the kidney in children with hemolytic-uremic syndrome. Radiology 1982; 145: 811–813. Diekmann L. Treatment of the hemolytic-uremic syndrome with streptokinase and heparin (author’s translation). Klin Padiatr 1980; 192: 430–435. Loirat C, Beaufils F, Sonsino E et al. Treatment of childhood hemolytic-uremic syndrome with urokinase. Cooperative controlled trial. Arch Fr Pediatr 1984; 41: 15–19. Henning PH, Tham EB, Martin AA et al. Haemolytic-uraemic syndrome outbreak caused by Escherichia coli O111:H-: clinical outcomes. Med J Aust 1998; 168: 552–555.
S41
http://www.kidney-international.org & 2009 International Society of Nephrology
Renal prognosis of toxigenic Escherichia coli infection Amit X. Garg1,2, Marina Salvadori3, Louise M. Moist1,2, Rita S. Suri1,2 and William F. Clark1,2 1
Division of Nephrology, Department of Medicine, University of Western Ontario, London, Ontario, Canada; 2Department of Epidemiology and Biostatistics, University of Western Ontario, London, Ontario, Canada and 3Division of Infectious Diseases, Department of Pediatrics, University of Western Ontario, London, Ontario, Canada
There is a spectrum of acute illness for individuals who ingest toxigenic Escherichia coli, ranging from no symptoms to self-limited gastroenteritis, to full acute hemolytic uremic syndrome. Knowledge of the long-term prognosis of acute toxigenic E. coli ingestion, based on the presentation of acute illness, is important for patient counseling and follow-up. Here, we consider subsequent 5-year renal outcomes for groups of individuals who presented across a spectrum of acute illness during a municipal water outbreak of E. coli O157:H7 in Walkerton, Canada. Kidney International (2009) 75 (Suppl 112), S38–S41; doi:10.1038/ki.2008.617 KEYWORDS: HUS; E. coli O157:H7; renal failure; hypertension; proteinuria
Correspondence: Amit X. Garg, Kidney Clinical Research Unit, Department of Medicine and Epidemiology, University of Western Ontario, Room ELL-101, Westminster Tower, London Health Sciences Centre, 800 Commissioners Road East, London, Ontario, Canada N6A 4G5. E-mail: [email protected] S38
Many strains of toxigenic Escherichia coli are implicated in diseases, with E. coli O157:H7 being the most common. The main source of infection is animals, particularly cattle, to which these bacteria are non-pathogenic. Bacteria residing in the gastrointestinal tract of animals are shed in the feces. Although contaminated meat is a major source of transmission, situational outbreaks of contaminated lettuce, alfalfa sprouts, unpasteurized milk, and apple cider have also lead to extensive morbidity.1 In the case of Walkerton, municipal water became contaminated with E. coli O157:H7. At the time of the outbreak, heavy rainfall contributed to the transport of livestock fecal contaminants into inadequately chlorinated drinking water, supplied from a shallow well. There is a spectrum of acute illness for the person who ingests toxigenic E. coli. This ranges from no symptoms to self-limited gastroenteritis, to full acute hemolytic uremic syndrome (HUS) (Figure 1). At the time of the outbreak, over 2300 people became ill with acute gastroenteritis, and there were 28 cases of HUS. Who becomes ill after toxigenic E. coli ingestion is influenced by a number of factors, including the amount and type of bacteria, and a patient’s predisposition to illness. When caring for someone with acute kidney injury from this infection, the focus is on adequate hydration, avoidance of nephrotoxins, and dialysis when it is required. Better awareness and supportive care result in most patients surviving the acute illness. Serum creatinine values frequently return to the laboratory reference range. However, acute kidney injury across many settings, even when transient, is increasingly recognized as having important long-term consequences.2 Similar to other renal insults, it has been suggested that HUS can lead to a critical reduction in nephron number, with unsustainable remnant single-nephron hyperfiltration and progressive renal disease.3 This then raises the question: Are patients who survive acute toxigenic E. coli infections at risk of longer term renal sequelae? Here, we consider subsequent renal outcomes for groups of individuals who presented across the spectrum of acute illness during the Walkerton outbreak.4–9 Participants were followed for a period of 5 years after the outbreak. E. coli O157:H7 INGESTION WITHOUT ANY SYMPTOMS
Over 95% participants in the Walkerton Health Study, or a thousand individuals, ingested contaminated municipal water without any gastrointestinal or other symptoms at Kidney International (2009) 75 (Suppl 112), S38–S41
review
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
1
3
2
Figure 1 | Spectrum of acute illness for individuals who ingest toxigenic Escherichia coli. After ingestion there may be a spectrum of acute illness from (1) no symptoms to (2) symptoms of selflimited gastroenteritis, to (3) full hemolytic uremic syndrome.
the time of the outbreak. Of course, when a person has no symptoms, one is never sure whether toxigenic E. coli were even ingested. Nonetheless, even if ingested, it seems biologically implausible that an individual without any symptoms of acute illness has appreciable silent renal injury which puts them at risk of future chronic sequelae. Consistent with this assertion, follow-up results for individuals without any symptoms at the time of the outbreak were similar to population reference standards.10 For this reason, participants without any symptoms at the time of the outbreak were treated as a reference or the control group, to which other groups with acute symptoms were compared. E. coli O157:H7 INGESTION WITH HUS
Of the 28 childhood cases of HUS during the outbreak, 1 child had proteinuria, which predated the outbreak, and 1 died at the time of the outbreak. Of the remaining 26 children, 7 were lost to follow-up and 19 participated in the Walkerton Health Study. Two physicians independently confirmed the presence of HUS during the outbreak. Briefly, in the absence of another medical explanation, children with all the following criteria at the time of the outbreak were defined as having HUS: (1) hemoglobin p10 g/100 ml (100 g/l) with evidence of destruction of erythrocytes on blood smear, (2) platelets p150 109/l, and (3) serum creatinine X95th percentile for age and sex. All these children with HUS were known to be previously healthy, Kidney International (2009) 75 (Suppl 112), S38–S41
without medical conditions (such as diabetes, hypertension, renal disease, recurrent infections, or failure to thrive) predating the outbreak. At the time of acute illness, 16 (84%) had stools positive for E. coli O157:H7 bacteria or verotoxin. The average lowest hemoglobin was 6.1 g/100 ml (s.d. 1.4, range 4.5–9.5) (61 g/l, range 45–95) and some patients were subsequently transfused, the average lowest platelet count was 56 109/l (s.d. 38, range 12–150), and the average highest serum creatinine was 2.8 mg/100 ml (s.d. 2.8, range 0.5–9.7) (246 mmol/l, s.d. 248, range 42–861). No child developed convulsions, coma, stroke, or severe colitis requiring surgery. Seven anuric children were treated with peritoneal dialysis for an average of 14 days (range 10–19). During the acute illness, one child was treated with antihypertensive therapy. All 19 children recovered in the months after acute HUS; none required long-term dialysis and all had a serum creatinine that returned to the lab-defined normal range. Two of these children were treated with antihypertensive therapy (an angiotensin-converting enzyme inhibitor) for variable amounts of time in follow-up; one child for the first 3 years, whereas another child remained on therapy until the 5-year assessment. What is the long-term renal prognosis of these children? An accurate estimate of prognosis remains critical for patient counseling, follow-up, and monitoring. To guide this estimate and data collection in the Walkerton Health Study, we conducted a systematic review of 49 studies of 3476 children with a mean follow-up of 4.4 years (range 1–22) from 18 countries, 1950–2001.11 The key finding of this review was that the long-term renal prognosis of diarrheaassociated HUS was highly variable. No definitive conclusion about a precise estimate could be drawn. The proportion of HUS survivors with renal sequelae in follow-up, defined as any one of a glomerular filtration rate (GFR) o80 ml/min per 1.73 m2, hypertension, or proteinuria, ranged anywhere from 0 to 64%. Some studies cited minimal sequelae in the 5 years after HUS,12–15 whereas other studies described survivors with hypertension, proteinuria, or a reduced GFR;16–20 in some cases, large proportions of children presented with such sequelae.21–25 Interpreting the studies remains a challenge because toxigenic E. coli was often not established as the cause of HUS, and many times large proportions of children were lost to follow-up. Different definitions were used for outcomes of proteinuria, GFR, and hypertension, and measurements were performed at variable times of follow-up. The majority of studies did not state whether observed sequelae were persistent after acute HUS, or whether they developed after apparent recovery. Finally, earlier studies did not use a healthy control group to establish outcomes attributable to HUS. Fortunately, when we followed the outcomes of children from Walkerton who recovered from HUS for 5 years after the outbreak, the results were very encouraging.9 Although more children who recovered from HUS presented with microalbuminuria than controls (20 vs 3%; relative risk: 6.0 (95% CI: 1.1–32.8)), no child had evidence of overt S39
review
proteinuria. There were no differences between donors and controls in blood pressure, or GFR when it was estimated with serum creatinine. GFR estimated with cystatin C was lower after HUS compared to controls (100 vs 110 ml/min per 1.73 m2, P ¼ 0.02), but no child had a GFR o80 ml/min per 1.73 m2. Reasons for differences between our results and other studies are likely multifactorial, and may relate to differences in the severity of acute disease, advances in health care, and possibly the use of angiotensin-converting enzyme inhibitor therapy in two children during follow-up. Unlike many other cohorts, serum creatinine values of HUS children in our study returned to the laboratory-defined normal reference range soon after their acute illness. E. coli O157:H7 INGESTION WITH SELF-LIMITED GASTROENTERITIS
No study had previously econsidered whether bacterial gastroenteritis, in the absence of recognized HUS, leads to clinically important long-term renal sequelae. We examined the outcomes of adults and children in the Walkerton Health Study separately, given that sequelae such as hypertension are defined in a very different manner among these two groups. The 1958 adults with no known history of hypertension before the outbreak were considered first.7 Of the participants, 675 were asymptomatic during the outbreak, 909 had moderate symptoms of acute self-limited gastroenteritis, and 374 had severe symptoms, which necessitated medical attention. Cultures were submitted in 40% of those with severe gastroenteritis, of which 14% were positive for E. coli O157:H7. Compared with those with moderate symptoms, participants with severe symptoms were more likely to describe bloody diarrhea (18 vs 40%), prolonged diarrhea (53 vs 70%), and fever (33 vs 50%). Four years after the outbreak, hypertension was diagnosed in 27.0% of those who were asymptomatic, and 32.3 and 35.9% of those with moderate and severe acute gastroenteritis, respectively (trend P ¼ 0.009). Compared with asymptomatic participants, the adjusted relative risk of hypertension was 1.15 (95% CI: 0.97–1.35) in those with moderate gastroenteritis, and 1.28 (95% CI: 1.04–1.56) in those with severe gastroenteritis. A similar graded association was observed for developing reduced kidney function, defined by the presence of an estimated GFR below 60 ml/min per 1.73m2 (trend P ¼ 0.03). No association was observed between gastroenteritis and the subsequent risk of albuminuria. In summary, acute bacterial gastroenteritis necessitating medical attention was associated with a higher risk of renal sequelae 4 years later. We then considered the renal outcomes of children and adolescents.8 Similar to adults, these outcomes were high blood pressure (defined by being greater than the 95th percentile expected for age, sex, and height), reduced kidney function, and microalbuminuria. Of the 951 participants, 313 were asymptomatic during the outbreak, 305 had moderate symptoms of acute gastroenteritis, and 333 had severe S40
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
symptoms, which necessitated medical attention. Compared with those with moderate symptoms, participants with severe symptoms were more likely to describe bloody diarrhea (17 vs 38%), prolonged diarrhea (51 vs 73%), and fever (45 vs 66%). Of those with severe symptoms during the outbreak, 87% described abdominal pain, 69% visited an emergency room, and 3% were admitted to hospital. A total of 70 participants with severe symptoms had a stool positive for E. coli O157:H7 (bacteria or Shiga toxin). Four years after the outbreak, there were no differences in mean systolic blood pressure between those who had no, moderate, or severe symptoms of acute gastroenteritis during the outbreak (109, 110, and 107 mm Hg). Similarly, there were no group differences in mean diastolic blood pressure, estimated GFR, or random urine albumin-to-creatinine ratio, or in the adjusted relative risk of high blood pressure, a GFR o80 ml/min per 1.73 m2, or microalbuminuria. In other words, patients who presented to medical attention with gastroenteritis during this E. coli O157:H7 outbreak had an absence of renal sequelae 4 years later. So, why did the outcomes between adults and children differ? It might be hypothesized that adults diagnosed with gastroenteritis had more severe disease than children. However, both groups with gastroenteritis had comparable presenting symptoms and rates of hospital admission, making it less likely that this was an important mitigating factor. Is it possible that the adult analyses were partly confounded? Unlike children, some adult participants may have had unmeasured co-morbidities predating the outbreak, which made them more susceptible to an acute gastrointestinal infection during the outbreak. Such individuals could then show higher rates of renal disease in follow-up, exaggerating any true association between acute gastroenteritis and subsequent sequelae. Finally, it remains conceivable that children, unlike adults, have more renal reserve, and were able to compensate after an insult of toxigenic E. coli more readily than adults. If this is true, longer follow-up of this group of children will elucidate chronic renal sequelae of E. coli gastroenteritis that were not evident 4 years after the outbreak. In conclusion, large water-borne infectious outbreaks are extremely rare within Western nations. The support of the local community for the Walkerton Health Study allowed us to make important observations about long-term renal sequelae. Time will tell how the results truly play out. For example, ongoing follow-up of the group of children with HUS will clarify the clinical relevance of microalbuminuria and mild decrements in GFR 5 years after recovery. Finally, it remains to be seen whether the observed prognosis ultimately generalizes to other patients affected by toxigenic E. coli infection. DISCLOSURE
LMM has received consulting fees from Amgen. WFC has received lecture fees from Amgen and AstraZeneca and is the recipient of research grant support from the Ontario Ministry of Health, the Canadian Institute of Health Research, and the Kidney Foundation of Canada. The remaining authors have declared no financial interests. Kidney International (2009) 75 (Suppl 112), S38–S41
review
AX Garg et al.: Renal prognosis of toxigenic E. coli infection
ACKNOWLEDGMENTS
This research was funded by the Kidney Foundation of Canada and the Ontario Ministry of Health. Dr Garg was supported by Clinician Scientist Award from the Canadian Institutes of Health Research.
11.
12. 13.
REFERENCES 1. Khanna R, Waechter L, Sargeant J et al. Environmental prevention of human disease from verocytotoxin-producing Escherichia coli. Nephrol Dial Transplant 2008; 23: 1819–1822. 2. Liano F, Felipe C, Tenorio MT et al. Long-term outcome of acute tubular necrosis: a contribution to its natural history. Kidney Int 2007; 71: 679–686. 3. Brenner BM, Lawler EV, Mackenzie HS. The hyperfiltration theory: a paradigm shift in nephrology. Kidney Int 1996; 49: 1774–1777. 4. Garg AX, Clark WF, Salvadori M et al. Microalbuminuria three years after recovery from Escherichia coli O157 hemolytic uremic syndrome due to municipal water contamination. Kidney Int 2005; 67: 1476–1482. 5. Garg AX, Macnab J, Clark W et al. Long-term health sequelae following E. coli and campylobacter contamination of municipal water. Population sampling and assessing non-participation biases. Can J Public Health 2005; 96: 125–130. 6. Garg AX, Marshall J, Salvadori M, et al., Walkerton Health Study Investigators. A gradient of acute gastroenteritis was characterized, to assess risk of long-term health sequelae after drinking bacterial-contaminated water. J Clin Epidemiol 2006; 59: 421–428. 7. Garg AX, Moist L, Matsell D et al. Risk of hypertension and reduced kidney function after acute gastroenteritis from bacteria-contaminated drinking water. CMAJ 2005; 173: 261–268. 8. Garg AX, Clark WF, Salvadori M et al. Absence of renal sequelae after childhood Escherichia coli O157:H7 gastroenteritis. Kidney Int 2006; 70: 807–812. 9. Garg AX, Salvadori M, Okell JM et al. Albuminuria and estimated GFR 5 years after Escherichia coli O157 hemolytic uremic syndrome: an update. Am J Kidney Dis 2008; 51: 435–444. 10. Jones CA, Francis ME, Eberhardt MS et al. Microalbuminuria in the US population: third National Health and Nutrition Examination Survey. Am J Kidney Dis 2002; 39: 445–459.
Kidney International (2009) 75 (Suppl 112), S38–S41
14.
15.
16. 17. 18.
19.
20.
21. 22. 23. 24.
25.
Garg AX, Suri RS, Barrowman N et al. Long-term renal prognosis of diarrhea-associated hemolytic uremic syndrome: a systematic review, meta-analysis, and meta-regression. JAMA 2003; 290: 1360–1370. Ogborn MR, Hamiwka L, Orrbine E et al. Renal function in Inuit survivors of epidemic hemolytic-uremic syndrome. Pediatr Nephrol 1998; 12: 485–488. McLean MM, Jones CH, Sutherland DA. Haemolytic-uraemic syndrome. A report of an outbreak. Arch Dis Child 1966; 41: 76–81. Brichard B, Vermylen C, Heremans MT et al. Plasma infusion as treatment for 33 children with haemolytic uraemic syndrome: a good therapy? Acta Clin Belg 1993; 48: 156–163. Vermylen C, Bodart E, Ninane J et al. The haemolytic uraemic syndrome of childhood: is living in Belgium a factor of good prognosis? Acta Clin Belg 1988; 43: 101–108. Miyazaki S. The etiology and clinical features of hemolytic uremic syndrome. Rinsho Ketsueki 1994; 35: 341–345. Riella MC, George CR, Hickman RO et al. Renal microangiopathy of the hemolytic-uremic syndrome in childhood. Nephron 1976; 17: 188–203. Guyot C, Daniel MD, Roze JC et al. Prognosis of hemolytic uremic syndrome in children. Importance of extrarenal involvement. Arch Fr Pediatr 1986; 43: 253–258. Rizzoni G, Claris-Appiani A, Edefonti A et al. Plasma infusion for hemolytic-uremic syndrome in children: results of a multicenter controlled trial. J Pediatr 1988; 112: 284–290. Brandt JR, Joseph MW, Fouser LS et al. Cholelithiasis following Escherichia coli O157:H7-associated hemolytic uremic syndrome. Pediatr Nephrol 1998; 12: 222–225. Sorrenti LY, Lewy PR. The hemolytic-uremic syndrome: experience at a center in the Midwest. Am J Dis Child 1978; 132: 59–62. Campos A, Kim Y, Miller K et al. Radionuclide studies of the kidney in children with hemolytic-uremic syndrome. Radiology 1982; 145: 811–813. Diekmann L. Treatment of the hemolytic-uremic syndrome with streptokinase and heparin (author’s translation). Klin Padiatr 1980; 192: 430–435. Loirat C, Beaufils F, Sonsino E et al. Treatment of childhood hemolytic-uremic syndrome with urokinase. Cooperative controlled trial. Arch Fr Pediatr 1984; 41: 15–19. Henning PH, Tham EB, Martin AA et al. Haemolytic-uraemic syndrome outbreak caused by Escherichia coli O111:H-: clinical outcomes. Med J Aust 1998; 168: 552–555.
S41