Differences in the clinical presentations of ZIKV infection which are occurring in different parts of the world might be an obstacle to setting a case definition that is good for use in all areas.
Until the French Polynesia outbreak, the Zika virus (ZIKV) was a self-limited disease with few human cases documented and restricted distribution on the Africa and Asia continents. In 2015, ZIKV began spreading rapidly in the Americas, and those countries saw increasing numbers of infections associated with complications such as Guillain-Barré syndrome, and non-previously predicted severe presentation such as congenital Zika virus syndrome. This changed the history of the arbovirus infection.
In 2015 through 2016, the fast globalization of ZIKV urged the creation of an interim case definition for the virus. It was hoped that this would optimize the detection of suspected cases to help public health organizations improve surveillance and management of the cases to gain control of the epidemic. However, we now know that slight differences in the clinical presentations of ZIKV infection which are occurring in different parts of the world might be an obstacle to setting a case definition that is good for use in all areas, particularly those with active co-transmission of other arboviruses. In this article, I discuss the points that need to be considered when developing a case definition for ZIKV infection.
Much like other arboviruses, such as Dengue (DENV) and Chikungunya (CHIKV) virus, those individuals infected with ZIKV may not have any symptoms or they may present with clinical manifestations that vary from mild to severe. In addition, because most arboviruses do not produce a singular and/or characteristic symptom, similar clinical features can be induced by distinct viruses.
In terms of coinfections of ZIKV and another arbovirus, the problem is not only that clinical manifestations do not distinguish a single arbovirus infection from another, but also that ZIKV co-infections cannot be “suspected” based on clinical presentation. Furthermore, most ZIKV cases are presumably asymptomatic, and make up those “suspected cases” that may be invisible to the surveillance system. As a result, the accuracy of case definitions can change, depending on the clinical features depicted to define the suspected cases in areas of arbovirus co-transmission.
What we do know is that suspected cases are considered probable when anti-ZIKV IgM reactivity is present. However, validated tests for the detection of anti-ZIKV IgM by serological methods are lacking, and those that are available have variable performance depending on the population tested. In areas of flavivirus co-transmission, such as areas with ZIKV and DENV, serum cross-reactivity occurs in moderate to high rates of flavivirus secondary infections. Therefore, ZIKV and DENV co-infection cannot be reliably diagnosed based on the serological reactivity.
Many times, testing biological samples with RNA-detection assays is key for the confirmation of ZIKV infection. However, although RNA shedding may remain for long periods in different biological fluids such as urine, blood, semen, vaginal fluid, and amniotic fluid, ZIKV RNA usually does not last longer than 5 to 13 days post-disease onset in serum and urine, respectively. In addition, in low viral load infections, the absence of RNA may not rule out infection in suspected cases.
Because laboratory tests show low sensitivity, ZIKV infection case definition accuracy also ends up affecting estimates of probable and confirmed cases which might compromise surveillance.
To this end, in a recent PLoS One article published on June 26, 2017, the authors propose a ZIKV case definition that performs well during simultaneous DEN, ZIKV, and CHIKV epidemics, making it more suitable for use in areas of arbovirus co-circulation. The group established a score prediction model (SPM) based on the clinical signs and symptoms “associated with ZIKV infection.” The study participants were a group of DENV, ZIKV, or CHIKV confirmed cases selected from a population of non-pregnant women and men presenting with acute febrile or exanthematous diseases. The cases were confirmed using reverse transcription polymerase chain reaction (RT-PCR). Those individuals with dual or triple arbovirus infections were excluded.
The data showed that the SPM had a sensitivity and specificity of 86.6% and 78.3%, respectively, suggesting moderate accuracy. In addition, “this Zika case definition also had the highest values for auROC (0.903) and R2 (0.417), and the lowest Brier score 0.096,” according to the study results. Another unique aspect of the PLoS One study is that the absence of fever and general clinical features actually empowers ZIKV infection case definition in areas of arbovirus co-transmission. These data also suggest that the same case definition is not applicable to a Brazilian co-transmission area.
To reinforce that differences of results exist in distinct areas, an additional study, this time of French Territories in America, found that maculopapular rash occurred in 84% of confirmed Zika cases in French overseas territories of America (French Guiana, Guadeloupe, Martinique, Saint Barthélemy and Saint Martin), as well as 93% of cases in French Polynesia. However, rash only occurred in 48.1% of cases in Rio de Janeiro, Brazil.
Based on these data, the robustness of case definition may only be achieved by tailoring the criteria to the transmission area studied.
Admittedly, using the SPM may lose almost 15% of cases and misdiagnose a little more than 20% of suspected cases. However, the accuracy of the results is still better than the ZIKV infection case definition used by the Pan American Health Organization (PAHO) 2016, the World Health Organization (WHO), the Centers for Disease Control and Prevention (CDC), the European Centre for Disease Prevention and Control (ECDC), and the Brazilian Ministry of Health (2016), the SPM showed the highest accuracy. This suggests that these global public health organizations may have worked with broader or universal criteria based on clinical evidence in other active transmission areas.
Marta G. Cavalcanti, MD, PhD, is a physician at Infectious Diseases Clinic, Hospital Universitario Clementino Fraga Filho at the Universidade Federal do Rio de Janeiro in Rio de Janeiro, Brazil, and an Editorial Advisory Board member for Contagion®.
José Mauro Peralta, PhD, is from the Instituto de Microbiologia Paulo de Góes at the Universidade Federal do Rio de Janeiro (IMPG/UFRJ) in Rio de Janeiro, Brazil.