BACKGROUND: Zika virus (ZIKV) disease is characterized by rash, joint pain, fever, and/or conjunctivitis. As of December 16, 2016, the Puerto Rico Department of Health (PRDH) reported 34,478 confirmed ZIKV infections among 65,778 individuals; however, additional ZIKV disease cases likely were not reported to PRDH due to ill individuals not seeking care or clinicians not suspecting ZIKV infection. We used household cluster investigations to identify clinical and demographic factors associated with unreported ZIKV disease cases.
METHODS: We offered participation to residents of all households within 100 meters of households where patients with confirmed ZIKV disease lived. Participants provided serum and urine specimens and answered a questionnaire to capture demographic and household characteristics, recent illnesses, and healthcare-seeking behaviors. We defined ZIKV infection as a laboratory-positive result by RT-PCR or IgM ELISA, and Zika-like illness as self-reported rash, arthralgia, fever, or conjunctivitis. We queried the PRDH surveillance database for participants with ZIKV infection who reported seeking care, and presence in the database was regressed on covariates using generalized estimating equations to account for clustering. We removed index participants from the analysis.
RESULTS: We carried out 19 cluster investigations during September–October 2016. Approximately half (242; 54.2%) of 446 households offered enrollment accepted, and 379 (58.8%) of 645 residents in enrolled households participated. Among 367 non-index participants, 113 (30.8%) were laboratory-positive; 90 (79.6%) by IgM ELISA, and 23 (20.4%) by PCR. Fifty-seven (50.4%) of these 113 participants reported a Zika-like illness in the prior 6 months (versus 19.7% of uninfected participants; p<0.0001). Among 57 laboratory-positive participants who reported Zika-like symptoms, median duration of illness was seven days (range: 2–45), and 33 (57.9%) sought care, of whom seven (21.2%) were laboratory-positive in the PRDH database. Among 33 laboratory-positive participants who sought care, factors significantly associated with being laboratory-positive in the PRDH database included having conjunctivitis (odds ratio [OR]: 12.48; 95% confidence interval [CI]: 1.20–129.92]) or abdominal pain (OR: 7.99; CI: 1.01–63.18).
CONCLUSIONS: Preliminary results suggest that at most half of people we detected with ZIKV infections were symptomatic, of whom roughly half sought care. One-fifth of those were reported to PRDH. Although having conjunctivitis or abdominal pain was associated with being laboratory-positive in the PRDH database, these findings may have been affected by ascertainment bias resulting from how infected individuals were identified. Additional educational outreach to clinicians should focus on the spectrum of illness associated with ZIKV disease and the need to report all patients with suspected ZIKV disease to PRDH.