Active household surveillance for AGE was conducted as a longitudinal cohort study with a nested case–control design from April 2015 to April 2019. Study methods have been described in detail previously.^8,10 Households were randomly selected for participation from a publicly available census in a geographic area of approximately 8 km² in San Jeronimo, Cusco Province, Peru. San Jeronimo is a periurban district with an average population of 47,635 during 2015–2019, and it has an arid climate at an elevation of 3,400 m.¹¹ Household enrollment into the cohort was conducted annually, during which households were reconsented and replaced if necessary. During the yearly enrollment process, sociodemographic characteristics were collected, and field staff provided training on recognition of AGE symptoms and stool collection to at least one designated adult household representative. Households could be enrolled if at least 60% of household members (i.e., individuals who had lived in the household for ≥2 weeks at the time of enrollment with no plans to change residence in the next 2 months) agreed to participate. Consent to participate in the study was obtained from persons 18 years of age or older. Parents and/or legal guardians provided consent for persons younger than 18 years of age; additionally, assent was obtained from children ages 10–17 years. Household members could join or leave the cohort at any time after initial enrollment and only contributed person-time to surveillance when enrolled. Although the surveillance period was intended to encompass 4 years (April 2015–March 2019), surveillance could not be conducted during October 2016 because of administrative issues; surveillance was thus extended for an additional month through April 2019.

An AGE event was defined as any of the following in a 24-hour period: 1) three or more loose or liquid stools, 2) two or more vomiting episodes, or 3) one or more vomiting episodes and one or more loose or liquid stools. Field staff visited households two to three times weekly to inquire whether any household members were currently experiencing AGE or had experienced symptoms since the previous visit. When an AGE event was reported by a household representative, data on illness onset, symptoms, and risk factors were collected from the case. Materials were provided to collect one stool specimen from the case within 10 days of symptom onset. Cases (i.e., individuals with an AGE event) were then followed up twice a week for 15 days after illness onset, and a follow-up questionnaire was administered at each visit to assess the clinical course of illness, including symptoms, duration, and health care-seeking behaviors.

Index cases were defined as the first household member to experience AGE within a 15-day period. All household members were asked to provide a stool specimen during days 7–17 from the index case’s symptom onset and were tested for norovirus. Those who developed AGE symptoms during the index case’s follow-up period were treated as AGE cases and asked to provide a stool specimen within the 10 days after their own symptom onset. To assess household transmission, we limited to norovirus-positive index cases. Secondary cases were defined as household members with norovirus-positive AGE events with onset during the index case’s follow-up period. Only household members who provided a stool specimen were included in analysis.

Upon identification of every fifth AGE case, a control household that was matched on household size and the number of children younger than 5 years old was randomly selected from the cohort. For selection as a control household, all household members were required to be free of AGE symptoms for 14 days before and 7 days after the matched household’s index case symptom onset. One stool specimen was collected from each member of the control household during the period from 7 to 17 days after the matched index case’s symptom onset. If any control household members became symptomatic during this period, they became an AGE case, and another control household was matched to the index AGE case if still within the follow-up period. Households could be included as matched controls multiple times throughout the study period.

Stool specimens were self-collected by participants and refrigerated or stored in a cool, dark place until collected by field staff the same day; then, they were stored at −80°C. Frozen specimens were shipped on dry ice to the U.S. Naval Medical Research Unit SOUTH in Lima, Peru and tested for GI and GII noroviruses by real-time reverse transcription polymerase chain reaction using a 7500 Fast real-time platform.⁸ Norovirus-positive specimens were genotyped at the Centers for Disease Control and Prevention in Atlanta, GA using methods described previously.¹² Specimens with a cycle threshold (Ct) value of <40 were considered to be norovirus positive.

During April 2015–April 2019, study staff enrolled an annual mean of 247 households (range: 235–273), which included an annual mean of 1,555 participants (range: 1,242–1,667), resulting in 4,176 PY of participation (Table 1). The median duration of household participation was 36 months (interquartile range [IQR]: 22–43 months). Among those enrolled, between 9% and 17% of participants each year were lost to follow-up because of voluntary withdrawal or non-AGE death. The median household size was six household members (range: 4–9). The median age of participants was 29 years (range: younger than 1 month to 99 years).

In total, 1,099 AGE events were reported from 235 households and 566 participants. Specimens were collected from 1,014 AGE events (92%), of which norovirus was detected in 186 specimens (18%). These 186 norovirus-positive AGE events occurred among 149 participants from 98 households. Children younger than 2 years of age experienced 181 AGE events with 55 norovirus-positive specimens (30%). The mean durations of AGE and norovirus AGE symptoms were 3.3 days (SD: 2.3 days) and 3.2 days (SD: 2.2 days), respectively. Among all AGE cases, 202 (18%) sought medical attention (physician/nurse visit or emergency room), and 9 (0.8%) were hospitalized; no AGE-associated deaths were reported. Among norovirus-positive AGE cases, 40 (22%) sought medical attention, and 3 (2%) were hospitalized.

Overall AGE incidence was 26.3 events/100 PY (95% CI: 24.8–27.9). AGE incidence across the four surveillance years (April 2015–April 2019) ranged from 24.6 events/100 PY during April 2016–March 2017 to 27.3 events/100 PY during April 2015–March 2016 (Table 2). AGE incidence was highest among participants younger than 2 years of age at 200.6 events/100 PY (95% CI: 173.4–232.0) and lowest among adults ages 18–64 years at 17.1/100 PY (95% CI: 15.5–18.8).

Norovirus-associated AGE incidence was 4.4 events/100 PY (95% CI: 3.9–5.1). Children younger than 2 years of age had the highest norovirus AGE incidence at 60.9 events/100 PY (95% CI: 46.8–79.4); adults ages 18–64 years had the lowest norovirus AGE incidence at 2.6 events/100 PY (95% CI: 2.0–3.3), despite having the highest absolute number of events (n 5 432 AGE events). By surveillance year, norovirus AGE incidence ranged from 3.5 events/100 PY during April 2015–March 2016 to 5.3 events/100 PY during April 2017–March 2018 (Table 2). No distinct seasonal pattern was observed for norovirus detections (Figure 1).

Among the 186 norovirus-positive AGE specimens, 155 specimens (83%) tested positive for norovirus GII, 27 specimens (15%) tested positive for norovirus GI, and 4 specimens (2%) tested positive for both norovirus GI and GII. We did not detect a significant difference in the distribution of GI and GII viruses by age group. Of 127 norovirus-positive AGE specimens available for typing, 96 (76%) could be dual typed successfully, resulting in 19 different genotypes. Four genotypes (GII.27[PNA9], GII.1[P16], GII.6[P7], and GII.4 Sydney[P31]) were detected in 55 (57%) of the typed AGE specimens (Figure 2).

Of 56 norovirus-positive asymptomatic control specimens, 11 (20%) tested positive for norovirus GI, and 45 (80%) tested positive for GII. Among 24 typed control specimens, 13 different dual types were identified, with GII.6[P7], GII.27[PNA9], GII.1[P16], GII.4 Sydney[P31], and G1.3[P3] totaling to 67% of typed specimens (Figure 2). Norovirus GII.4 viruses accounted for 20% (19/96) of norovirus AGE events and 21% (5/24) of norovirus detections among controls. GII.27 viruses were found in 20% (n = 19) of specimens from norovirus AGE events and in 13% (3/24) of norovirus-positive control specimens.

Multiple (two to four) norovirus AGE events were experienced by 29 cohort participants, accounting for 66 total events. The median time to recurrence was 7.7 months (IQR: 5.0–16.2 months). Eleven individuals (38%) had both GI and GII norovirus events, whereas the remainder had only GII norovirus detected. Children younger than 5 years of age experienced the most recurring events, with 42 norovirus AGE events among 18 individuals in this age group. Specimens were typed from nine participants who experienced two or more norovirus AGE events. Six of these nine experienced repeat norovirus GII infections with differing genotypes. Only one individual tested positive for the same genotype (GII.27[PNA9]) from more than one AGE event. This individual had two detections with symptom onset dates occurring 31 days apart; however, Giardia lamblia was also detected during the second event.

Among the 186 norovirus events, 166 met the index case definition. In total, 473 household contacts provided symptom data and a stool specimen during the index case’s follow-up period. The majority (68%; n = 323) of household contacts were negative for norovirus, including 12 AGE cases. Around one third (32%; n = 152) of household contacts tested positive for norovirus, including 20 (13%) with AGE symptoms (i.e., secondary cases) and 132 (87%) who were asymptomatic. The secondary attack rate for symptomatic norovirus AGE was 4%. Among norovirus AGE events, the median time from index case illness onset to secondary case onset was 2 days (IQR: 1–7 days; n = 20).

There were 67 household contacts with genotyping data for whom the index case also had available genotyping data, including six symptomatic secondary cases. Of these, 51 (76%) were infected with the same genotype as the index case, including all six secondary cases. The four genotypes most commonly identified among cases accounted for 42 of these 51 secondary infections (82%): GII.1[P16] (n = 12), GII.6[P7] (n = 11), GII.4 Sydney[P31] (n = 10), and GII.27[PNA9] (n = 9). Symptomatic secondary cases were caused by GII.27[PNA9] (n = 3), GII.6[P7] (n = 2), and GII.12[P33] (n = 1). We did not observe statistically significant differences in household transmission of norovirus by the age of the index cases or the age of the household contacts.

Stool specimens were provided by 516 AGE-free participants in 128 matched control households. Control households had a median of four (range: three to six) household members. The age distribution among controls was similar to that of the cohort overall. The shortest duration between AGE onset as prior norovirus case and specimen collection as a control was 144 days. Among 672 stool specimens provided by control participants, 8% (56 specimens) were norovirus positive. Norovirus positivity among controls was highest among children younger than 2 years old (5/18; 28%). Across all age groups, the odds of norovirus detection were 2.2 (95% CI: 1.6–3.1) times higher among AGE cases than among healthy controls. As a sensitivity analysis, we reclassified norovirus positives based on stricter Ct cutoffs (Ct ≤35 for GI norovirus and Ct ≤37 for GII norovirus) and found that the odds of norovirus detection were 2.7 (95% CI: 1.9–3.8) times higher among AGE cases than among healthy controls.