Prolonged episodes of acute diarrhea reduce growth and increase risk of persistent diarrhea in children

Abstract

Background & aims: Prolonged episodes of acute diarrhea (ProD; duration 7-13 days) or persistent diarrhea (PD; duration ≥14 days) are important causes of undernutrition, yet the epidemiology and nutritional impact of ProD are poorly understood.

Methods: We conducted a 10-year cohort study of 414 children from a Brazilian shantytown who were followed from birth; data were collected on diarrhea, enteric pathogens, and anthropometry.

Results: During 1276 child-years of observation, we recorded 3257 diarrheal episodes. ProD was twice as common as PD (12% and 5% of episodes, respectively); ProD and PD together accounted for 50% of all days with diarrhea. ProD was more common in infants whose mothers had not completed primary school (relative risk [RR], 2.1; 95% confidence interval: 1.02-2.78). Early weaning was associated with earlier onset of ProD (Spearman ρ = 0.309; P = .005). Infants with ProD were twice as likely to develop PD in later childhood (log rank, P = .002) compared with infants with only acute diarrhea (AD; duration <7 days), even after controlling for confounders. Children's growth was more severely stunted before their first episode of ProD, compared with AD (mean height-for-age Z score (HAZ) -0.81 vs -0.51, respectively, P < .05, unpaired t test). Following ProD, HAZ (ΔHAZ = -0.232) and weight-for-age (ΔWAZ = -0.26) significantly decreased (P < .005 in paired t tests). ProD was associated with Cryptosporidium and Shigella infections.

Conclusions: ProD accounts for significant morbidity and identifies children at risk of a vicious cycle of diarrhea and malnutrition. Further studies are needed to address the recognition and control of ProD and its consequences in resource-limited settings and assess its role in PD pathogenesis.

Figure 1

A. Diarrhea attack rates of acute (<7 days), prolonged (≥7 and <14 days), and persistent (≥14 days) episodes per child-year, by age, among 414 children in Gonçalves Dias in Fortaleza, Brazil, from August 1989 through March 2000. B. Days with diarrhea per child-year, by age. Episodes and days of diarrhea for prolonged and persistent diarrhea peaked at 6-12 and 12–24 months of age, respectively.

Figure 1

A. Diarrhea attack rates of acute (<7 days), prolonged (≥7 and <14 days), and persistent (≥14 days) episodes per child-year, by age, among 414 children in Gonçalves Dias in Fortaleza, Brazil, from August 1989 through March 2000. B. Days with diarrhea per child-year, by age. Episodes and days of diarrhea for prolonged and persistent diarrhea peaked at 6-12 and 12–24 months of age, respectively.

Figure 2

Proportions of total diarrhea morbidity accounted for by acute (<7 days), prolonged (≥7 and <14 days), and persistent (≥14 days) episodes and by days of illness.

Figure 3

Proportional hazard curves of time to first persistent diarrheal (PD, duration ≥14 days) episode in children with (n=112) and without (n=294) a prolonged episode of diarrhea (ProD, duration ≥7 and <14 days) in the first year of life. Eight children who developed PD prior to their first episode of ProD were excluded. Children who experienced at least one ProD episode of diarrhea before age 1 were nearly twice (15.5% (s.e.=2.5%) versus 29.9% (s.e.=5.0) as likely to experience a PD episode by age 2 years. Cox Regression showed (OR 2.2, 95% CI; [1.32-3.54], P=0.002) a similar overall increased risk of PD by age 6 years. ProD before age 1 year remained a significant (P =0.025) independent predictor of later PD even after controlling for household crowding, type of sanitation, maternal education, age at weaning from exclusive breastfeeding, and birth date.

Figure 4

Impact of acute (<7 days), prolonged (≥7 and <14 days), and persistent (≥14 days) diarrhea on anthropometry. Nutritional Z-scores were compared 3 months before and after children's first acute (n=308), prolonged (n=145), and persistent (n=62) episodes. A. Weight-for-height Z (WHZ) scores declined significantly following persistent episodes, but not acute or prolonged episodes. B. Weight-for-age Z (WAZ) scores decreased with all episodes types and mean WAZ prior to acute diarrhea was significantly greater than mean WAZ prior to persistent diarrhea. C. Height-for-age Z (HAZ) scores declined following acute and prolonged episodes, but not persistent episodes. HAZ prior to acute diarrhea was greater than HAZ prior to prolonged or persistent diarrhea. (Error bars indicate SEM; *P<0.005, paired t-test of Z scores before and after diarrhea; §P<0.05, unpaired t-test of Z scores prior to acute vs. prolonged, acute vs. persistent, or prolonged vs. persistent episodes.)