Zika virus: causality, open science and risk of emerging infectious diseases
The Zika virus (ZIKV) outbreak in the Americas between 2015 and 2017 took the world by surprise. Within two years, over 1.5 million suspected or confirmed cases were reported. However, the true incidence is likely much higher, due to under-reporting and asymptomatic infections that are undetected. As of July 2019, 87 countries had reported ongoing or past circulation of ZIKV. ZIKV infection results generally in mild and transient symptoms. The disease caused by ZIKV is often asymptomatic or mild. However, infection during pregnancy can result in severe adverse congenital outcomes with microcephaly as most prominent. This was first noted in clusters of infants born with disabilities linked to ZIKV infection in Brazil in 2015, making ZIKV a disease with a serious public health impact. In this thesis, I explore different aspects of the ZIKV epidemic. I use different epidemiological methods to provide insight in the Zika virus as a cause of adverse outcomes, ZIKV as a sexually transmitted disease and the risk of future ZIKV outbreaks.
In Chapter 1, I provide an introduction to the history of emerging infections and the emergence of ZIKV specifically. I describe the investigation of causality, the use and accumulation of evidence during disease outbreaks, and how disease transmission can be investigated using mathematical models.
In Chapter 2, I provide insight in how evidence accumulates during an outbreak and more in general during new causal questions. Case reports and case series were the first studies to appear, followed by basic research (in vivo and in vitro studies). It took more than a year after the onset of the ZIKV outbreak for robust epidemiological studies to be published. Establishing early public health guidance thus requires a broad approach taking into account all evidence available. We have to make do with the low quality evidence. To minimize further delays, evidence should be accessible as soon as it becomes available through rapid and open access dissemination.
In Chapter 3, I extend a systematic review that was conducted earlier, and turn it into a living systematic review. I introduce the concept and implementation of living systematic reviews in the context of an emerging disease. I assess the evidence on the causal relation between ZIKV infection and adverse congenital and auto-immune neurological outcomes, published between May 30, 2016 and January 18, 2017, using a framework based on the causality dimensions of Bradford Hill. During this period, the evidence expanded that ZIKV was indeed a cause of congenital abnormalities and Guillain-BarrΓ© syndrome (GBS). I provide a proof of concept for the use of living systematic reviews to synthesize evidence about an emerging pathogen such as ZIKV.
In Chapter 4, I assess the evidence published between January 18, 2017 and July 1, 2019. I quantify the strength of association of the relation between maternal ZIKV infection and congenital adverse outcomes and between ZIKV infection and GBS. I found that the strength of association between ZIKV infection and adverse outcomes from case-control studies differs according to whether exposure to ZIKV is assessed in the mother (odds ratio (OR) 3.8, 95% CI: 1.7β8.7, I2=19.8%) or the foetus/infant (OR 37.4, 95% CI: 11.0β127.1, I2=0%). In cohort studies, the risk of congenital abnormalities was 3.5 times higher after ZIKV infection (95% CI: 0.9β13.5, I2=0%). The strength of association between ZIKV infection and GBS was higher in studies that enrolled controls from hospital (OR: 55.8, 95% CI: 17.2-181.7, I2=0%) than in studies that enrolled controls at random from the same community or household (OR: 2.0, 95% CI: 0.8β5.4, I2=74.6%). The heterogeneity between the studies could be partly explained by the heterogeneity in methods and sampled populations. Studies suffered from bias and uncontrolled residual confounding.
In Chapter 5, I present a framework to systematically assess the evidence for ZIKV as a sexually transmitted disease. I reviewed all available literature and concluded that the risk of sexual transmission of ZIKV is likely small, but relevant for certain risk groups. I found that in semen viral RNA could be detected for a median period of 34 days (95% CI: 28β41 days) and 35 days (no CI given) based on two cohort studies. Aggregated data about detection of ZIKV RNA from 37 case reports and case series indicate a median duration of 40 days (95% CI: 30β49 days) and a maximum duration of 370 days in semen. In human vaginal fluid, the median duration was 14 days (95% CI: 7β20 days) and the maximum duration was 37 days. Infectious virus in human semen was detected for a median duration of 12 days (95% CI: 1β21 days) and a maximum of 69 days. I highlight the poor quality of the evidence and the need for systematic observational studies that evaluate the risk of sexual transmission of ZIKV.
In Chapter 6, I present predictions on the future risk of ZIKV, based on data from Managua, Nicaragua, using mathematical modelling. The risk of a new outbreak in the next decades is low due to herd immunity. However, a next outbreak will disproportionally hit people in the young reproductive age hardest (age 15β29 years). Vaccination could curb this risk: Early introduction of vaccination in 15-year-old girls has the capacity to extend the herd immunity and be of benefit to the whole population. Introduction of a vaccine needs to happen within a decade after the 2016 outbreak to achieve this protection. The duration of immunity following ZIKV infection has impact on the speed at which outbreaks will reoccur.
In Chapter 7, I present an overview of the main findings and I discuss the interpretation and implications of these results. I discuss the strengths and limitations of the work, and outline follow-up questions emerging from the work.
In this thesis, I establish and use different frameworks and methods that help to make sense of the limited evidence that is available during disease outbreaks. ZIKV has been introduced on the American continent, and it is likely there to stay, thus we have to accept that ZIKV will continue to re-emerge. At the same time, due to the climate change, the European temperate region also becomes more suitable for vector-borne disease such as ZIKV. With the ZIKV epidemic on the wane, we now have time to consolidate findings and implement the lessons learnt. We need to be prepared for the re-emergence of ZIKV but also for the emergence of new diseases. The tools and methods I present in this thesis, will help us to be more prepared for a next outbreak.