From bioactive substances to research on breast-feeding promotion.

UNLABELLED: Despite known health benefits, exclusive breast-feeding for at least 4 months is uncommon in many countries. In Mexico, most mothers initiate breast-feeding but few breast-feed exclusively. OBJECTIVE: The objective was to examine the effectiveness of home visits by lay peer counselors to increase exclusive breast-feeding among mothers in a periurban area of Mexico. METHODS: An ethnographic assessment conducted in 1994 that identified key maternal beliefs, practices, and needs was used to guide educational strategies. Lay counselors were recruited from the same community and trained by La Leche League. From March 1995 through September 1996, pregnant women were identified by community census and invited to participate. Women were enrolled into a randomized, controlled study of 3 groups: no intervention (control), 3 visits, and 6 visits during pregnancy and early postpartum. Data collection was performed by a social worker apart from the counselors. Exclusive breast-feeding was defined by WHO criteria. RESULTS: The study enrolled 130 women; 52 were in the 3-visit group, 44 in the 6-visit group, and 34 in the control group. Study groups did not differ in the maternal characteristics or initiation of breast-feeding (96%). At 3 months postpartum, exclusive breast-feeding was practiced by only 12% of controls vs. 52% in the 3-visit group and 67% in the 6-visit group (P < 0.001, log rank test). In the first 3 months, significantly (P = 0.037) fewer intervention than control infants had an episode of diarrhea (11% vs. 26%, respectively). Intervention effectiveness was independent of maternal factors or birth hospital. CONCLUSIONS: This unique experimental study demonstrated a dramatic increase in exclusive breast-feeding and a significant reduction in infant illness in an urban community through well-designed maternal support including early intervention and repeated contact.

Serum antibody as a marker of protection against natural rotavirus infection and disease.

To determine whether naturally acquired serum IgA and IgG antibodies were associated with protection against rotavirus infection and illness, a cohort of 200 Mexican infants was monitored weekly for rotavirus excretion and diarrhea from birth to age 2 years. Serum samples collected during the first week after birth and every 4 months were tested for anti-rotavirus IgA and IgG. Children with an IgA titer >1:800 had a lower risk of rotavirus infection (adjusted relative risk [aRR], 0.21; P<.001) and diarrhea (aRR, 0. 16; P=.01) and were protected completely against moderate-to-severe diarrhea. However, children with an IgG titer >1:6400 were protected against rotavirus infection (aRR, 0.51; P<.001) but not against rotavirus diarrhea. Protective antibody titers were achieved after 2 consecutive symptomatic or asymptomatic rotavirus infections. These findings indicate that serum anti-rotavirus antibody, especially IgA, was a marker of protection against rotavirus infection and moderate-to-severe diarrhea.

Rapid ethnographic assessment of breastfeeding practices in periurban Mexico City.

Before carrying out a breastfeeding promotion programme in a periurban area of Mexico City, we conducted a rapid ethnographic study to determine the factors associated with absence of exclusive breastfeeding. The responses to pilot interviews were used to develop a standardized questionnaire regarding reasons for infant feeding choice, sources of advice, and barriers to breastfeeding. We interviewed a random sample of 150 mothers with a child < 5 years of age; 136 (91%) of them had initiated breastfeeding; but only 2% exclusively breastfed up to 4 months. The mothers consistently stated that the child's nutrition, health, growth, and hygiene were the main reasons for the type of feeding selected; cost, comfort, and the husband's opinion were less important. Physicians were ranked as the most important source of advice. Reduction or cessation of breastfeeding occurred on the doctor's advice (68%); or when the mothers encountered local folk illnesses such as "coraje" (52%) or "susto" (54%), which are associated with anger or fright; or had "not enough milk" (62%) or "bad milk" (56%); or because of illness of the mother (56%) or child (43%). During childhood illnesses and conditions, breastfeeding was reduced and the use of supplementary foods was increased. This study emphasizes the importance of cultural values in infant feeding choices, defines specific barriers to breastfeeding, and provides a basis for interventions to promote exclusive breastfeeding in the study population.

PIP: Prior to initiating a community-based intervention program to promote exclusive breast feeding in San Pedro Martir, Mexico, a 2-month (1994) rapid ethnographic assessment was conducted. 150 mothers whose youngest child was under 5 years of age were interviewed. 136 mothers (91%) had breast-fed their infant, for a median duration of 6 months, but only 2% exclusively breast-fed for up to 4 months. Mothers consistently described breast feeding as the best nutrition for their infant. However, the dominant feeding pattern was mixed breast and bottle-feeding. Formula, tea, and water were introduced during the first postpartum day. By the end of the third month, 63% of mothers had introduced solid food to promote growth. It was common practice to reduce breast feeding and increase feeding of supplementary foods when a child was ill. Physicians were the most respected source of knowledge on breast feeding. 42% of mothers reported that, at some point when they were breast feeding, a doctor had advised them to stop and half these mothers complied. The data collected in this rapid survey were used to guide a peer counseling program to promote exclusive breast feeding in the community.

Efficacy of home-based peer counselling to promote exclusive breastfeeding: a randomised controlled trial.

BACKGROUND: Exclusive breastfeeding is recommended worldwide but not commonly practised. We undertook a randomised controlled study of the efficacy of home-based peer counselling to increase the proportion of exclusive breastfeeding among mothers and infants residing in periurban Mexico City. METHODS: Two intervention groups with different counselling frequencies, six visits (44) and three visits (52), were compared with a control group (34) that had no intervention. From March, 1995, to September, 1996, 170 pregnant women were identified by census and invited to participate in the study. Home visits were made during pregnancy and early post partum by peer counsellors recruited from the same community and trained by La Leche League. Data were collected by independent interview. Exclusive breastfeeding was defined by WHO criteria. FINDINGS: 130 women participated in the study. Only 12 women refused participation. Study groups did not differ in baseline factors. At 3 months post partum, exclusive breastfeeding was practised by 67% of six-visit, 50% of three-visit, and 12% of control mothers (intervention groups vs controls, p<0.001; six-visit vs three-visit, p=0.02). Duration of breastfeeding was significantly (p=0.02) longer in intervention groups than in controls, and fewer intervention than control infants had an episode of diarrhoea (12% vs 26%, p=0.03). INTERPRETATION: This is the first reported community-based randomised trial of breastfeeding promotion. Early and repeated contact with peer counsellors was associated with a significant increase in breastfeeding exclusivity and duration. The two-fold decrease in diarrhoea demonstrates the importance of breastfeeding promotion to infant health.

A prospective study of astrovirus diarrhea of infancy in Mexico City.

AIM: To describe the epidemiologic and clinical characteristics of astrovirus-associated diarrhea in a cohort of young children from a periurban community in Mexico City. METHODS: From November, 1988, through December, 1991, a total of 214 children were enrolled in a longitudinal study of diarrhea and monitored from birth to 18 months of age. A stool specimen was collected during each episode of diarrhea. Specimens from a total of 510 diarrhea episodes were tested for astrovirus by enzyme immunoassay and examined for other enteric pathogens. The antigenic types of astrovirus were determined by a typing enzyme immunoassay. RESULTS: Astrovirus was detected in 26 (5%) of 510 diarrhea episodes, with an incidence rate of 0.1 episode/child year; the highest rate was in children 13 to 18 months of age. Astrovirus-associated diarrhea was characterized by a median of 4 stools (range, 2 to 10) during the first 24 h, a median duration of 3 days (range, 1 to 21), vomiting (20%), and fever (7%). No cases of dehydration or repeat symptomatic infections were observed. Coinfection with another pathogen was detected in 11 of the 26 episodes (42%). Serotype 2 (35%) was most common, followed by serotypes 4 (15%), 3 (11%), and 1 and 5 (4% each); 31% were nontypable. Astrovirus-associated diarrhea was less severe, as measured by the number of stools (4.3 +/- 1.9), than diarrhea caused by rotavirus (7.1 +/- 2.8) or when coinfections occurred (5.5 +/- 1.6; P = 0.008). CONCLUSIONS: Astrovirus was associated with 5% of the episodes of diarrhea in this cohort of young Mexican children and presented as a mild secretory diarrhea. Five predominant antigenic types were detected with type 2 being the most common.

Role of human-milk lactadherin in protection against symptomatic rotavirus infection.

BACKGROUND: Human milk contains a 46 kDa mucin-associated glycoprotein, lactadherin, which binds specifically to rotavirus and inhibits its replication. This study tested the hypothesis that lactadherin protects against symptoms of rotavirus infection. METHODS: 200 infants in Mexico City were recruited at birth and monitored by regular stool EIA for rotavirus, serology, and recording of feeding and stool patterns. Milk samples were obtained from the mothers weekly until 4 weeks post partum then monthly. The sample taken immediately before an infant's episode of rotavirus infection was assayed for lactadherin, butyrophilin, mucin, and secretory IgA. An infection was defined as symptomatic if diarrhoea occurred in the 5 days before or after detection of the virus. FINDINGS: 31 infants developed rotavirus infection; 15 were symptomatic and 16 had no symptoms. The median concentration of lactadherin in the milk samples (obtained 4-41 days [median 13] before the infection) was 48.4 (range 5.6-180) microg/mL in the asymptomatic group and 29-2 (6.2-103-4) microg/mL in the symptomatic group. Although these medians did not differ significantly, in logistic regression analysis adjusted for age at infection and secretory IgA concentration there was a significant difference between the groups (p=0O01). No association between symptom status and concentrations of butyrophilin, mucin, or secretory IgA was found. INTERPRETATION: Protection against rotavirus by human milk is associated with the glycoprotein lactadherin. This association is independent of products of the secretory immune system.

Rotavirus infection in infants as protection against subsequent infections.

BACKGROUND: Rotavirus is the leading cause of severe diarrhea in infants. To provide a base line for assessing the efficacy of rotavirus vaccines, we evaluated the protection that is conferred by natural rotavirus infection. METHODS: We monitored 200 Mexican infants from birth to two years of age by weekly home visits and stool collections. A physician assessed the severity of any episodes of diarrhea and collected additional stool specimens for testing by enzyme immunoassay and typing of strains. Serum collected during the first week of life and every four months thereafter was tested for antirotavirus IgA and IgG. RESULTS: A total of 316 rotavirus infections were detected on the basis of the fecal excretion of virus (56 percent) or a serologic response (77 percent), of which 52 percent were first and 48 percent repeated infections. Children with one, two, or three previous infections had progressively lower risks of both subsequent rotavirus infection (adjusted relative risk, 0.62, 0.40, and 0.34, respectively) and diarrhea (adjusted relative risk, 0.23, 0.17, and 0.08) than children who had no previous infections. No child had moderate-to-severe diarrhea after two infections, whether symptomatic or asymptomatic. Subsequent infections were significantly less severe than first infections (P=0.024), and second infections were more likely to be caused by another G type (P=0.054). CONCLUSION: In infants, natural rotavirus infection confers protection against subsequent infection. This protection increases with each new infection and reduces the severity of the diarrhea.

Asymptomatic Shigella infections in a cohort of Mexican children younger than two years of age.

The proportion of Shigella infections that occur asymptomatically in young children has not been established. A community-based cohort study of 367 infants was followed prospectively by weekly home visits from January, 1990, through December, 1991. Stool samples were collected weekly and when diarrhea occurred and were tested for Shigella and other enteropathogens. There were 2925 child months of observation and 65 episodes of Shigella infection. There were 3.1 episodes/100 child months during the warm season (May through September) and 0.97 episode/100 child months during the cold season. Shigella infections were rare during the first 6 months of life but increased with age (P < 0.0001). Overall 55% of detected infections were asymptomatic. The proportion of infections that were asymptomatic increased as age increased (P < 0.01). Symptom status was not significantly associated with Shigella species or season. All isolates from symptomatic and asymptomatic children had the 120- to 140-megadalton virulence plasmid. We conclude that infections with virulent strains of Shigella are commonly asymptomatic in Mexican children during the first 2 years of life.

PIP: During January 1990-December 1991, each week, field workers visited the home of 367 children aged 0-24 months from a periurban area southwest of Mexico City (San Pedro Martir and San Andres Totoltepec, Tlalpan) and collected stool specimens from them to determine whether Shigella infections are often asymptomatic. The crude incidence rate of diarrhea, regardless of etiology, was 29 episodes/100 child months during the warmer and rainy months (May-September), while it was 21 episodes/100 child months for the rest of the year (October-April) (relative risk [RR] =1.38). 53 of all children (l4%) had 65 Shigella infections. The overall monthly incidence of symptomatic and asymptomatic Shigella infection was higher during May-September than October-April (3.13 vs. 0.97 episodes/100 children; RR = 3.22). 55% of all Shigella infections (36) were asymptomatic. 32% developed secretory-type diarrhea and 13% had blood present in the stool. The incidence of Shigella infections grew as did the age (0.4-8.2 episodes/100 child months for 0-6 month olds to 18-24 month olds; p 0.0001). The proportion of asymptomatic Shigella infections also increased with age (33% for 0-6 month olds, 40% for 7-12 month olds, 46% for 13-18 month olds, and 78% for 18-24 month olds; p 0.01). Shigella sonnei, S. flexneri, and S. boydii were the only species detected. The 120-140 megadalton virulence plasmid was present in all isolates from asymptomatic and symptomatic children. Mixed infections were rather common in both asymptomatic (47%) and symptomatic (45%) children. Among infants aged less than 12 year months, breast feeding infants were less likely to be infected with Shigella than nonbreast feeding infants (RR = 2.41). On the other hand, among children aged 12-24 months, nonbreast feeding was associated with a lower risk of Shigella infection (RR = 0.69). These findings show that Shigella infections in Mexican children aged 0-24 months range from asymptomatic infections to secretory diarrhea to bloody diarrhea.

Prevalence of stress hyperglycemia among patients attending a pediatric emergency department.

OBJECTIVE: To determine the prevalence and clinical characteristics associated with stress hyperglycemia among children and adolescents attending a pediatric emergency department. DESIGN: Patients who required a venipuncture for evaluation of an acute illness or injury from October 1992 to March 1993 in an urban pediatric emergency department were enrolled and screened prospectively for hyperglycemia (glucose level > or = 8.3 mmol/L; > or = 150 mg/dl). Data were collected regarding demographic characteristics, history, clinical findings, and admission status. RESULTS: A total of 926 patients ranging in age from 3 days to 21 years were enrolled. Blood glucose values ranged from 1.94 mmol/L (35 mg/L) to 14.65 mmol/L (264 mg/dl); 35 patients (3.8%) had hyperglycemia. The prevalence of stress hyperglycemia was significantly increased among patients if they (1) had temperatures greater than 39.5 degrees C (9.3%) versus normal temperatures (2.8%) (p < 0.001), (2) had been admitted to a critical care unit of the hospital (24.1%) or to any hospital unit (4.4%) versus not having been admitted (2.6%) (p < 0.001), and (3) had received fluids intravenously (6.0%) versus having received no fluids intravenously (2.7%) (p = 0.014). CONCLUSIONS: Stress hyperglycemia is a frequent clinical occurrence in a pediatric emergency department. It does not appear to be associated with a particular diagnostic category but is significantly associated with severity of illness as measured by elevated temperature, hospital admission, and hydration status.

Outbreaks of astrovirus gastroenteritis in day care centers.

OBJECTIVE: This study evaluated astrovirus as a cause of diarrhea outbreaks among infants and toddlers in day care centers. DESIGN: Stool specimens were collected weekly during four periods (from January 1986 through December 1991) from children 6 to 30 months of age who were enrolled in prospective studies of diarrhea in day care centers. All diarrheal stool specimens were tested for bacterial enteropathogens, rotavirus, enteric adenovirus, and Giardia lamblia. A total of 1365 stool specimens from 70 outbreaks in which no etiologic agent was identified and from another 11 outbreaks with a known cause were tested for astrovirus, by means of a monoclonal antibody-based enzyme immunoassay. Confirmatory testing was performed by reverse transcriptase-polymerase chain reaction with primers designed to produce an 89 base-pair product. RESULTS: Astrovirus was detected in 6 (7%) of the 81 outbreaks. Of 217 children tested, 73 (34%) were infected with astrovirus; infections in 35 (48%) were symptomatic and in 38 (52%) asymptomatic. The six outbreaks lasted 11 to 44 days (median 22 days). Astrovirus excretion was detected for a duration of 2 to 30 days, with excretion occurring from 1 to 8 days (median 2 days) before diarrhea began to 1 to 20 days (median 2 days) after diarrhea ceased. Younger children (< or = 12 months) were at greater risk than older children (p = 0.011) of becoming infected with astrovirus during an outbreak and were more likely (p = 0.015) to have symptoms when infected. Of 24 specimens with astrovirus by enzyme immunoassay, 20 (83%) were confirmed to have the virus by reverse transcriptase-polymerase chain reaction. CONCLUSION: Astrovirus was an important cause of outbreaks of diarrhea among children attending day care centers, more frequently infected younger children, and often produced asymptomatic infections.

Multiply resistant nontyphoidal Salmonella gastroenteritis in children.

From January, 1990, to December 31, 1990, 75 children with multiply resistant Salmonella gastroenteritis were studied at the Children's Hospital "Ricardo Gutierrez" of Buenos Aires. These children ranged from 1 month to 15 years of age. Infection was community-acquired in 20 (26.6%), nosocomially acquired in 50 (66.7%) and undetermined in 5. Thirty-nine (52%) had grossly bloody stools. Fever occurred at some point in the clinical course in 61 children (81.3%) with a duration of 1 to 33 days (mean, 6.7 days). The duration of diarrhea (1 to 69 days) was longer in those who developed complications (P < 0.001). Six (8%) developed enterocolitis (2 with bowel perforation), 1 had a pulmonary abscess and 8 (11.4%) had bacteremia; 4 children died (5.3%). Salmonella typhimurium was the most common serovar (85.3%). Ninety percent minimum inhibitory concentration studies demonstrated that all strains were resistant to ampicillin (> 128 micrograms/ml), cephalothin (> 128 micrograms/ml), cefuroxime (> 128 micrograms/ml), nalidixic acid (> 256 micrograms/ml), rifampin (> 256 micrograms/ml), gentamicin (> 256 micrograms/ml) and tobramycin (256 micrograms/ml); 77.3% of strains were resistant to ceftazidime (32 micrograms/ml), 97.6% to netilmicin (> 256 micrograms/ml), 92.8% to amikacin (256 micrograms/ml), 24.4% to isepamicin (32 micrograms/ml), 5.3% to chloramphenicol (4 micrograms/ml) and 2.7% to cefoxitin (2 micrograms/ml). The 90% minimum inhibitory concentration of cefotaxime and ceftazidime was reduced by the addition of clavulanate. Aggressive multiply resistant Salmonella strains are a major pediatric problem in Buenos Aires.

Child day care increases the risk of clinic visits for acute diarrhea and diarrhea due to rotavirus.

A case-control study of risk factors for acute diarrhea was conducted among children under 3 years of age attending a health maintenance organization clinic in Houston, Texas. During a 19-month period from September 1985 through March 1987, 339 children with diarrhea and 363 age- and season-stratified controls were enrolled. A total of 90% of cases were under age 2 years. Compared with children cared for at home, the risk of clinic visits for diarrhea was significantly greater for children receiving child day care and was similar for those attending day care centers (odds ratio (OR) = 2.4, 95% confidence interval (CI) 1.6-3.7), day care homes (OR = 2.0, 95% CI 1.3-3.1), mother's day out (OR = 1.8, 95% CI 0.8-4.2), or when cared for by a relative (OR = 2.0, 95% CI 1.2-3.2). Rotavirus-positive diarrhea was also significantly greater in child day care (OR = 2.4, 95% CI 1.1-5.1). The day care-associated risk was highest during the first month of enrollment (OR = 3.1, 95% CI 1.8-5.4). In this population, where 40% of children receive child care, 19% of the clinic visits for acute diarrhea were attributable to child care. These data indicate that child day care increases the risk of acute diarrhea, whether in a center or in a home, and the risk is highest in the first month of enrollment.

Concentration of milk secretory immunoglobulin A against Shigella virulence plasmid-associated antigens as a predictor of symptom status in Shigella-infected breast-fed infants.

We conducted a prospective, community-based study of healthy breast-fed Mexican infants to determine the protective effects of anti-Shigella secretory IgA antibodies in milk. Milk samples were collected monthly, and stool culture specimens were obtained weekly and at the time of episodes of diarrhea. Nineteen breast-fed infants were found to have Shigella flexneri, Shigella boydii, or Shigella sonnei in stool samples. Ages of the 10 infants with symptomatic infection and the nine with asymptomatic infection did not differ significantly. Milk samples collected up to 12 weeks before infection were evaluated by enzyme-linked immunosorbent assay for secretory IgA antibodies against lipopolysaccharides of S. flexneri, S. boydii serotype 2, S. sonnei, and virulence plasmid-associated antigens. The geometric mean titers of anti-Shigella antibodies to virulence plasmid-associated antigens in milk received before infection were eightfold higher in infants who remained well than in those in whom diarrhea developed. The significance of milk secretory IgA directed against lipopolysaccharide was less clear. We conclude that human milk protects infants against symptomatic shigella infection when it contains high concentrations of secretory IgA against virulence plasmid-associated antigens.

Protection against infection with Giardia lamblia by breast-feeding in a cohort of Mexican infants.

To determine whether breast-feeding protects infants against symptomatic and asymptomatic infection by Giardia lamblia, we followed 197 infants in a poor area of Mexico City from birth to 18 months of age; symptoms and feeding status were recorded weekly. Stool specimens were collected every 1 to 2 weeks and tested for Giardia by enzyme-linked immunosorbent assay. A mean of 1.0 Giardia infection per child-year was detected; 94 infants had a total of 139 infections; 17% of infections were symptomatic. Ninety-one percent of infants were breast fed from birth and 38% were breast fed at 1 year of age. Lack of breast-feeding was a significant risk factor for first Giardia infection at all ages. The adjusted incidence rate ratio for first Giardia infection for none versus complete breast-feeding was 5.0 (confidence interval (CI) 1.5 to 16.9; p = 0.009), and for none versus any breast-feeding, 1.8 (CI 1.1 to 2.8; p = 0.013). Symptomatic Giardia infection was also associated with lack of breast-feeding (none vs any: incidence rate ratio = 2.5; CI 0.9 to 6.8; p = 0.077), but breast-feeding did not protect against chronic carriage of Giardia. Other significant risk factors for Giardia infection were presence of animals in the household (p = 0.005) and the use of water or nonmilk liquid for infant feedings (p = 0.035). We conclude that breast-feeding protects infants against Giardia by mechanisms that include preventing the establishment of infection.

PIP: To determine whether breast feeding protects infants against symptomatic and asymptomatic infection by Giardia lamblia, the authors followed 197 infants in a poor area of Mexico City from birth to 18 months of age. Symptoms and feeding status were recorded weekly. Stool specimens were collected every 1-2 weeks and tested for Giardia by an enzyme-linked immunosorbent assay. A mean of 1.0 Giardia infection/child-year was detected; 94 infants had a total of 139 infections and 17% were symptomatic. 91% of the infants were breast fed from birth and 38% were breast fed at 1 year of age. Lack of breast feeding was a significant risk factor for 1st Giardia infection across all ages. The adjusted incidence rate ratio for 1st Giardia infection for none vs. complete breast feeding was 5.0 (confidence interval [CI] 1.5-16.9; p=0.009), and for none vs. any breast feeding, 1.8 (CI 1.1-2.8; p=0.013). Symptomatic Giardia infection was also associated with a lack of breast feeding (none vs. any; incidence rate ratio=2.5; CI 0.9-6.8; p=0.077), but breast feeding did not protect against chronic carrying of the infection. Other significant risk factors for Giardia infection were: presence of animals in household (p=0.005) and the use of water or nonmilk liquid for infant feedings (p-0.035). The authors conclude that breast feeding feeding helps protect infants against Giarda by mechanisms which include the prevention of infection invasion at the outset.

Association between severity of gastrointestinal prodrome and long-term prognosis in classic hemolytic-uremic syndrome.

To determine whether severity of the prodromal gastrointestinal illness is associated with the course and complications of the extraintestinal manifestations of hemolytic-uremic syndrome, we conducted a retrospective review of children (n = 509) hospitalized with hemolytic-uremic syndrome. Those who came to the hospital with colitis and rectal prolapse associated with hemolytic-uremic syndrome (group I, n = 40) were compared with an equal number of time-matched children with hemolytic-uremic syndrome but without prolapse (group II). Children in group I had evidence of more severe colitis than children in group II had, as indicated by increased frequency of bloody diarrhea (p less than 0.001) and longer duration of diarrhea (p less than 0.001). However, they also had more severe extraintestinal manifestations during hemolytic-uremic syndrome, including edema (p less than 0.0001), severe thrombocytopenia (p less than 0.0001), prolonged anuria (p less than 0.001), and seizures (p = 0.036). Long-term prognosis for recovery of renal function was worse for group I than group II. Within group II, patients with bloody diarrhea had milder extraintestinal illness than those with prolapse but more severe extraintestinal illness than those with watery diarrhea. Analysis of Kaplan-Meier survival curves demonstrated a better prognosis for return of normal renal function in the children with watery diarrhea but without prolapse (p = 0.009) than in children with bloody diarrhea or prolapse. These data demonstrate that the severity of the gastrointestinal prodrome reflects the severity of the extraintestinal acute microangiopathic process and the resulting long-term outcome. Widespread vascular damage, often followed by permanent sequelae, is characteristic of patients with the most severe colitis.