High Prevalence and Increased Severity of Norovirus Mixed Infections Among Children 12–24 Months of Age Living in the Suburban Areas of Lima, Peru

Abstract

In an active diarrhea surveillance study of children aged 12–24 months in Lima, Peru, norovirus was the most common pathogen identified. The percentage of mixed (bacterial and noroviral) infections was significantly higher among norovirus-positive samples (53%) than among norovirus-negative samples (12%). The combination of norovirus with the most common bacterial pathogens was associated with increased clinical severity over that of either single-pathogen norovirus or single-pathogen bacterial infections.

Studies conducted in the United States after the introduction of rotavirus vaccination showed that norovirus is now the most common etiologic agent of medically attended acute gastroenteritis among pediatric patients [1, 2]. In low-resource countries, norovirus infections are highly prevalent among symptomatic and asymptomatic children younger than 5 years [3–5]. Infectious gastroenteritis in children living in areas with poor sanitation is usually associated with more than 1 pathogen [5, 6]. In this study, we examined the characteristics of community acquired norovirus-associated gastroenteritis among children younger than 24 months living in periurban areas of Lima, Peru, and we paid special attention to mixed infections.

MATERIALS AND METHODS

Population

Participants were recruited for a randomized controlled trial aimed at measuring the effect of lactoferrin supplementation in preventing diarrhea among young children [7]. The study was conducted between January 2008 and May 2011. Eligible subjects were healthy, previously weaned, 12- to 18-month-old children living in the District of Independencia, Lima. All families with a child of target age were offered to participate; 555 children were enrolled and followed for 6 months with daily home visits for data collection and lactoferrin administration. The supplementation with lactoferrin did not have any effects on the incidence of diarrhea, the relative prevalence of pathogens, or the severity of the diarrheal episodes.

The results presented here represent a secondary analysis of the trial data. Because no statistically significant differences were observed between the group of the children on lactoferrin supplementation and the one on placebo, we merged the data to create a single population for the purpose of this analysis.

Methods

Diarrhea was defined as the presence of ≥3 loose or watery stools in 24 hours or >1 loose stool with blood. A new episode was defined as onset of diarrhea preceded by at least 2 consecutive days without symptoms. We used a modified Ruuska-Vesikari score to determine diarrheal episode severity [8]. An episode was defined as mild if the total score was up to 8, moderate if the score was 9–14, and severe if the score was >15. Information regarding medical history and household characteristics was collected at enrollment. Community health workers performed home visits 5 days/week to record clinical data and collect stools for each diarrheal episode. Growth was assessed through monthly follow-up visits.

Bacterial enteropathogens were detected by using conventional cultures. Diarrheagenic Escherichia coli infections were diagnosed by using a multiplex real-time polymerase chain reaction (RT-PCR) [9], rotavirus and adenovirus were diagnosed by using immunochromatography (Operon, Huerva-Zaragoza, Spain), and norovirus was diagnosed by using one step RT-PCR to detect genogroups I and II (GI and GII, respectively) using previously described primers designed to detect the ORF1 and ORF2 junctions of each genogroup [10]. Parasites were detected by direct microscopy, special stains, and concentration methods.

Statistical Analysis

Each incidence was calculated as the number of norovirus infections per child-year of follow-up. Each prevalence was calculated as the number of stool samples with results that were positive for a specific pathogen over the total number of samples analyzed for all pathogens (788). Incidences and prevalences are reported here with 95% confidence intervals (CIs). Anthropometric measures, weight for age and length for age, are reported with Z scores per World Health Organization guidelines. The clinical characteristics of the episodes associated with single-pathogen norovirus infection are compared to those of the episodes associated with mixed (noroviral and bacterial) infections and single-pathogen bacterial infections and reported with medians or percentages. χ², Fisher exact, Wilcoxon, and Kruskal-Wallis tests were used as appropriate. Logistic regression was used to identify individual and household risk factors for single-pathogen and mixed infections. Statistical analysis was performed using Stata 12 (Stata Corp, College Station, Texas).

This study was approved by the institutional review boards of the University of Texas Health Science Center in Houston and Universidad Peruana Cayetano Heredia in Lima.

RESULTS

As previously described, we enrolled 555 children and conducted 6 months of follow-up for a total of 250.3 child-years of observation. The median age at enrollment was 16 months; the mean (±SD) weight-for-age Z score at enrollment was 0.29 (±0.93), and the mean (±SD) height-for-age Z score was –0.57 (±0.95)

We observed 1235 episodes of gastroenteritis with an incidence of all-cause diarrhea of 5 episodes per child-year. According to the modified Ruuska-Vesikari score, 84.5% of the episodes were classified as mild, 14.5% as moderate, and only 0.9% as severe. A total of 915 diarrheal stool samples were collected; 788 (86.1%) were available for norovirus testing, representing 67% of all diarrheal episodes (the remaining samples were of insufficient quantity). Norovirus was the most prevalent pathogen detected in 34.8% of the samples. The other most common infectious agents were diarrheagenic E coli (31.0%), Campylobacter spp. (10.5%), Giardia lamblia (6.5%), and Shigella spp. (6.2%). Rotavirus and adenovirus were found in 3.3% and 2.5% of the diarrheal samples, respectively. In the analysis of the 788 samples included, more than 1 pathogen was found in 206 (26.1%), a single pathogen was found in 356 (45.2%), and no pathogen was identified in 226 (28.7%).

Norovirus-Associated Gastroenteritis

Of 275 norovirus-positive samples, 45 (16.3%) were positive for genogroup I, 228 (82.9%) were positive for genogroup II, and 2 (0.7%) were positive for both genogroups. The average incidence of norovirus-associated diarrhea was 1.1 episode per child-year (Figure 1). The incidence of GII norovirus did not vary according to years of follow-up or season. The incidence of GI norovirus was slightly higher in the summer months (0.3 episode per child-year [95% CI, 0.2–0.5 episode per child-year]) than in the winter months (0.1 episode per child-year [95% CI, 0–0.2 episode per child-year]) (P = .02). Thirty-five percent (193 of 555) of the children tested positive for norovirus during at least 1 diarrheal episode, 5.4% tested positive in 2 episodes, 3.2% tested positive in 3 episodes, 0.72% tested positive in 4 episodes, and 0.18% tested positive in 5 episodes. In 19 (23%) of 82 repeated episodes of norovirus-associated diarrhea, the genotype was different than that in the previous episode and therefore represented true new infections. In 2 cases, repeated episodes were caused by GI norovirus; each episode occurred rather soon after the previous one (5 and 9 days, respectively), thus likely representing prolonged shedding of the virus. For the remaining 61 cases in which GII norovirus was detected in consecutive bouts of diarrhea, the median time between episodes was 29 days (interquartile range [IQR], 17–55 days). Because norovirus can be shed in the feces for several weeks after initial infection and we have not yet analyzed stool samples collected during asymptomatic periods or quantified the viral load in diarrhea samples, it is difficult to determine whether these findings represent repeated infections or just prolonged excretion.

Figure 1.

Incidences of norovirus-associated diarrheal episodes according to month, January 2008 to May 2011. Incidence is reported as the number of episodes per child-year. The dashed line indicates the average monthly incidence of diarrhea, the solid line indicates the monthly incidence of norovirus-positive episodes, and the dotted line indicates the percentage of norovirus-positive episodes of all diarrheal episodes. Norovirus was detected year-round without a clear seasonal pattern, although there was a small increase in incidence in the late-spring and summer months over that in the winter months (see text under Results section, paragraph “Norovirus-associated gastroenteritis”).

The prevalence of norovirus among symptomatic children decreased with age; it was 44% among children between 12 and 14 months of age, 39% for those between 15 and 17 months, 33% for those between 18 and 21 months, and 31% for those older than 22 months (P = .03, test for trend).

The percentages of mixed infections were significantly higher among the norovirus-positive samples than among the norovirus-negative specimens (53% [146 of 275] vs 11.6% [60 of 513]; P < .001). Among the norovirus-positive samples, diarrheagenic E coli (45%), Campylobacter spp. (12%), G lamblia (9%), and Shigella spp. (3%) were the most common coinfecting agents. Among the norovirus-negative samples, the most common combinations were diarrheagenic E coli with a Campylobacter sp. (22%), E coli with rotavirus (15%), E coli with G lamblia (10%), and E coli with a Shigella sp. (10%).

Diarrheal episodes associated with mixed (noroviral and bacterial) infections were generally more severe than those with single-pathogen norovirus infections with regard to duration (median, 5 [IQR, 3–8] days vs 3 [IQR, 3–6] days, respectively; P = .001), number of bowel movements per episode (15.5 [IQR, 10–27] vs 10 [IQR, 7–10], respectively; P = .0003), and number of moderate-to-severe episodes (22.6% vs 8.5%, respectively; P = .001). To understand if these differences were a result of the coinfecting agents or a true negative effect of the combination of norovirus plus another bacterial pathogen, we compared the clinical characteristics of episodes associated with single-pathogen norovirus infections with episodes associated with the most common single-pathogen bacterial infections (Campylobacter spp., E coli, and Shigella spp.) and episodes associated with mixed norovirus and Campylobacter, E coli, and Shigella infections. Despite the wide difference in group sizes, we observed that in case of Campylobacter sp. or E coli infections, the duration of diarrhea and the number of bowel movements per episode were consistently increased in mixed infections compared with those of single-pathogen norovirus infections or single-pathogen bacterial infections (Table 1).

Table 1.

Comparison of Clinical Characteristics of Diarrheal Episodes Associated With Noroviral, Bacterial, and Mixed (Noroviral and Bacterial) Infections

Characteristic	(A) Noroviral, Single Pathogen (n = 129)a	(B) Bacterial, Single Pathogen (all) (n = 165)b	(C) Bacterial/Norovirus (all) (n = 90)c	(D) Campylobacter, Single Pathogen (n = 32)d	(E) Campylobacter/Norovirus (n = 17)e	(F) Shigella, Single Pathogen (n = 22)f	(G) Shigella/Norovirus (n = 7)g	(H) E coli, Single Pathogen (n = 111)h	(I) E. coli/Norovirus (n = 66)i
Median (IQR) duration (days)	3 (3–6)	4 (3–7)j	5 (3–8)k	5.5 (3–7)	6 (3–8)m	5 (3–8)	6 (3–7)n	4 (3–7)	5 (3–8)o
Median no. (IQR) of bowel movements per episode	10 (7–18)	13(8–22)j	16 (10–26)k	13.5 (10–22)	19(12–28)m	16.5 (8–30)	18 (15–24)n	13 (8–21)	14.5 (8–25)o
Percentage of episodes with fever	10.1	15.1	17.8	18.8	6m	18.8	28.5n	13.5	19.7o
Percentage of episodes with bloody diarrhea	1.6	6.6j	4.4	3.3	0	31.8	14.3n	1.8	4.6o
Percentage of episodes with vomiting	17	15.1	22.2	18.8	29.4m	22.7	14.3n	12.6	22.2o
Percentage of moderate-to-severe episodes (Ruuska-Vesikari score, ≥9)	8.5	13.3	24.4k,l	15.6	29.4m	31.82	14.3n	9	24.4o

We compared the clinical characteristics of episodes associated with Campylobacter sp., Shigella sp., E coli, and single-pathogen norovirus infections with the episodes associated with mixed infection with norovirus plus one of the mentioned bacterial infections. These microorganisms were chosen because they represent the most common pathogens and combinations of pathogens in our population. Diarrheal episodes associated with mixed (bacterial and noroviral) infections tended to have increased durations of diarrhea, total numbers of bowel movements, and severity scores over those of episodes associated with either noroviral or bacterial single-pathogen infections (Campylobacter sp., Shigella sp., and E coli infections considered as a single group). When data were split according to bacterial pathogen, we observed that mixed Campylobacter/norovirus and E coli/norovirus infections were associated with episodes of increased severity compared with either single-pathogen norovirus or single-pathogen Campylobacter sp. and single-pathogen E coli infections, respectively. Abbreviation: IQR, interquartile range.

^aNoroviral, single pathogen: diarrheal episodes in which stool samples tested positive for norovirus only.

^bBacterial, single pathogen (all): all diarrheal episodes in which stool samples tested positive for 1 bacterial pathogen only (a Campylobacter sp., a Shigella sp., or E coli).

^cBacterial/norovirus: all diarrheal episodes in which stool samples tested positive for norovirus plus a single bacterium (a Campylobacter sp., a Shigella sp., or E coli).

^dCampylobacter, single pathogen: diarrheal episodes in which only a Campylobacter sp. was detected in the stool samples.

^eCampylobacter/norovirus: diarrheal episodes in which stool samples tested positive for a Campylobacter sp. and norovirus only.

^fShigella, single pathogen: diarrheal episodes in which only a Shigella sp. was detected in the stool samples.

^gShigella/norovirus: diarrheal episodes in which stool samples tested positive for a Shigella sp. and norovirus only.

^hE coli, single pathogen: diarrheal episodes in which only E coli was detected in the stool samples.

ⁱE coli/norovirus: diarrheal episodes in which stool samples tested positive for E coli and norovirus only.

^jP < .05 for the comparison between episodes associated with single-pathogen norovirus infections to single-pathogen bacterial infections (column A vs B).

^kP < .05 for the comparison between episodes associated with single-pathogen norovirus infections to mixed (bacterial and noroviral) infections (column A vs C).

^lP < .05 for the comparison between episodes associated with single-pathogen bacterial infections to mixed (bacterial and noroviral) infections (column B vs C).

^mP < .05 for the comparison between episodes associated with single-pathogen norovirus infections, single-pathogen Campylobacter sp. infections, and mixed Campylobacter/norovirus infections (columns A, D, and E).

ⁿP < .05 for the comparison between episodes associated with single-pathogen norovirus infections, single-pathogen Shigella sp. infections, and mixed Shigella/norovirus infections (columns A, F, and G).

^oP < .05 for the comparison between episodes associated with single-pathogen norovirus infections, single-pathogen E coli infections, and mixed E coli/norovirus infections (columns A, H, and I).

When analysis was restricted to single-pathogen norovirus infections, the median age at diagnosis was 17 months (IQR, 15–19.5 months) for GI norovirus infection and 19 months (IQR, 17–20 months) for GII norovirus infection (P = .05). No other significant differences were observed in terms of clinical characteristics (eg, duration of episodes, number of bowel movements per episode, and proportions of children with vomiting, fever, and moderate-to-severe episodes). We did not observe differences in clinical severity of the first versus repeated episodes of norovirus-associated diarrhea when it was analyzed according to each genogroup.

An analysis aimed at describing the clinical characteristics of diarrheal episodes associated with single-pathogen norovirus in relation to age at the onset of diarrhea failed to find any significant differences. However, it should be mentioned that our population was quite homogenous in terms of age, with 50% of our samples collected from children aged 18 to 21 months and 86% from children aged between 15 and 21 months, which is probably too narrow of an age window in which to appreciate differences in clinical severity.

No associations were noted between norovirus infections and sex, anthropometric measures at enrollment, number of people living in the household, presence of indoor running water or a toilet, water storage inside the house, or presence of a refrigerator.

DISCUSSION

In this study, conducted among suburban communities of Lima, norovirus was the most prevalent enteric pathogen (detected in 35% of diarrheal cases).

The causal relationship between fecal shedding of norovirus and gastroenteritis is relatively easy to establish among children living in industrialized countries, because stools usually test positive for only 1 pathogen. However, causality is more complicated to establish in low-resources settings, where the carriage of multiple enteropathogens is common during the first years of life and diarrhea presumably occurs when either a new infection is acquired or a definite quantitative threshold of the “usual” pathogens is exceeded [5, 6]. We found that 53% of norovirus-positive episodes involved mixed infections. In such instances, it is difficult to establish whether norovirus was the cause of the disease, a contributing factor, or just an “innocent bystander.” Nonetheless, our study confirms that toddlers living in limited-resource settings are repeatedly infected with norovirus, which raises questions regarding the possible long-term morbidity of such exposure. Norovirus is known to preferentially infect the small intestine and cause temporary reduction of the villous surface, infiltration of mononuclear cells, and increased mucosal permeability [11]. Similar results have been found in individuals with environmental enteropathy, a chronic intestinal disorder commonly found among people living in areas with a high incidence of enteric infections and associated with chronic malnutrition in children [12]. Recurring intestinal infections are likely the cause of such enteropathy, although the exact pathophysiologies and relative contributions of specific pathogens are unknown. A recent birth cohort study, also conducted in Peru, revealed a possible association between norovirus infection and growth among children up to 1 year of age [3]. We observed an increased frequency of mixed infections in norovirus-positive samples over that in norovirus-negative samples and a tendency toward longer duration of diarrhea and higher number of bowel movements in cases of mixed (noroviral and bacterial) infections than in cases of single-pathogen infections. Whether norovirus predisposes to additional enteric infections or these findings just reflect heavy exposure to fecal pathogens that results from inadequate sanitation needs to be established, but the possible contribution of norovirus to chronic enteropathy and malnutrition deserves further investigation.

We did not observe a difference in clinical severity of the first versus repeated GI or GII norovirus-associated diarrheal episodes. However, it should be noted that we enrolled children 12 months of age or older, a time of life at which they have been already repeatedly exposed to both genogroups [3] and likely have acquired some form of immunity; the episodes we observed were probably all reinfections and thus did not significantly differ in clinical appearance.

Our study has several limitations. We tested only 67% of all diarrhea episodes for norovirus and thus may be providing conservative estimates of incidence and prevalence. More important is that we have not yet tested the stool samples collected from healthy children living in the same community; therefore, we do not know the prevalence of asymptomatic norovirus infections. Such information will be crucial for establishing the significance of norovirus detection in children with diarrhea. Nevertheless, the results of this study confirm that norovirus is a common and persistent pathogen among young children living in low-resource countries and suggest possible interactions with other enteric microorganisms that may be worth exploring in the future.