The 2014 chikungunya virus outbreak in the U.S. Virgin Islands: epidemiology, long-term health outcomes, and economic burden

Thesis (Ph.D.)--University of Washington, 2016-06

Background: Chikungunya virus (CHIKV), an emerging and acutely debilitating alphavirus transmitted by the Aedes aegypti and Aedes albopictus mosquito, was introduced into the Americas in December of 2013. As of April 2016, almost 2 million suspected or confirmed cases have been reported in 45 different countries in the Americas. Acute symptoms of the virus include high fever, severe polyarthralgia (incapacitating joint pain in two or more joints), headache and myalgia. Symptoms often resolve within 7-10 days. However, up to 79% of cases in previous outbreaks have reported persistent arthralgia, defined as joint pain lasting more than two weeks, resulting in decreased quality of life for up to 36 months following initial illness. Currently no cure or vaccine exists for infection, there are no effective therapeutic treatments for chronic symptoms, and disease prevention measures have proven to be insufficient. This dissertation sought to estimate the following: demographic risk factors and clinical manifestations associated with symptomatic CHIKV infection (Aim 1), prevalence of persistent arthralgia among CHIKV cases compared to similar healthcare seekers 1-2, 6 and 12 months after illness onset (Aim 2), the direct and indirect costs associated with the 2014-2015 CHIKV outbreak on the U.S. Virgin Islands (USVI) (Aim 3), and household characteristics and individual behavioral practices of vector-control as potential risk factors for CHIKV disease (Aim 4). Methods: All four aims were addressed using CHIKV surveillance data from the USVI Department of Health. Aims 2-4 were also addressed using data from a year-long prospective cohort study of laboratory-positive CHIKV cases and similar healthcare seekers. For Aim 1, descriptive statistics were used to summarize and compare laboratory-positive and suspected laboratory-negative cases from the surveillance data. For Aim 2, three separate regression models were fitted for self-reported presence of persistent arthralgia 1-2, 6 and 12 months following illness onset, adjusting for age, gender and self-reported history of arthritis. Generalized linear models using the binomial family with robust variance estimators were constructed to estimate prevalence differences of persistent arthralgia among cases and the comparison group using the identity link. For Aim 3, direct medical costs were estimated by calculating the mean cost of inpatient and outpatient visits associated with a suspected CHIKV case and indirect costs were estimated by multiplying the mean number of work days missed by the average annual wage in the USVI. For Aim 4, generalized linear models using the binomial family with robust variance estimators were constructed to estimate prevalence differences of household characteristics and personal-protective measures among cases and the comparison group using the identity link. Results: CHIKV incidence was highest among individuals aged 55-64 years (13.06 per 1,000 cases) and lowest among individuals aged 0-14 years (1.77 per 1,000 cases). Incidence was higher among women compared to men (6.57 and 5.00 cases per 1,000, respectively). More than half of the reported laboratory-positive cases experienced fever lasting 2-7 days, chills/rigor, myalgia, anorexia, and headache. No clinical symptoms apart from the suspected case definition of fever >38 °C and arthralgia were significantly associated with being a reported laboratory-positive case. One to two months after disease onset, the difference in prevalence of persistent arthralgia between cases and the comparison group was 42% (95% CI: 32%-52%), after adjusting for age, sex and self-reported history of arthritis. The difference in prevalence of persistent arthralgia between cases and the comparison group at 6 months was 32% (95% confidence interval [CI]: 23-40%) after adjustment for potential confounders; at 12 months after onset, the difference in prevalence was 19% (95% CI: 11-28%). Twelve months after illness onset, cases were 1.81 (95% CI: 1.08-3.02) times more likely to have difficulty walking, 1.96 (95% CI: 1.24-3.12) times more likely to have difficulty climbing stairs, and 2.63 (95% CI: 1.31-5.29) times more likely to have difficulty getting in and out of a car compared to similar healthcare seekers. The total estimated cost associated with the 2014-2015 CHIKV outbreak in the USVI ranged from $36.9 to $37.1 million, of which 13% was direct medical costs and 87% was indirect costs due to absenteeism from work. Household characteristics and individual-level behavior practices of vector-control did not differ between laboratory-positive CHIKV cases and similar healthcare seekers during the 2014-2015 CHIKV outbreak in the USVI. Conclusions: These findings highlight the long-term impaired physical functionality of CHIKV cases, the need for therapeutic and vaccine research to manage and prevent acute illness and long-term morbidity, and the significant economic burden of the first outbreak in the USVI. These results will aid policy-makers in creating informed decisions about prevention and control measures for inevitable future CHIKV outbreaks. Larger-scale seroprevalence and long-term cohort studies will further aid in determining the acute and long-term burden, as well as the public health impact of CHIKV and other arboviral outbreaks in the Americas.