Epidemiological Characteristics and Laboratory Findings of Zika Virus Cases in New York City,January 1,2016–June 30,2017

Background

Zika virus (ZIKV), first identified in 1947 in Uganda, is a Flavivirus that is closely related to dengue virus (DENV) and West Nile virus (Hayes 2009, Faye et al. 2014). Transmission primarily occurs through the bite of an infected mosquito, but also through pregnancy from mother to fetus (Pan American Health Organization 2015), sex with an infected individual (Petersen et al. 2016), or transfusion of infected blood products (Motta et al. 2016). Infection during pregnancy can result in severe congenital abnormalities (Mlakar et al. 2016). An outbreak of ZIKV infection first reported in Brazil in May 2015 rapidly spread throughout the Americas; in the United States, there has been local transmission in Florida and Texas (Hennessey et al. 2016). New York City (NYC) has a diverse population with ∼1.8 million residents who were born in these ZIKV-affected areas and who may travel to visit friends and family (U.S. Census Bureau 2016).

Diagnosis of ZIKV infection using current methodologies can be challenging. Commercially available tests include nucleic acid amplification-based tests (NAAT) and serology (ZIKV immunoglobulin M [IgM] assay). Detection of ZIKV RNA in serum or urine is considered the gold standard for diagnosis of acute infection. A cohort study of 150 patients in Puerto Rico found that ZIKV RNA persists on average for 14 days in serum and 8 days in urine following illness onset (Paz-Bailey et al. 2017), although reports have identified instances of prolonged viremia (Bingham et al. 2016, Campos De et al. 2016, Lustig et al. 2016, Meaney-Delman et al. 2016, Murray et al. 2017).

Methods available for serologic diagnosis during the period of this report relied on the presence of ZIKV IgM antibodies, thought to develop 4–7 days after infection (Lanciotti et al. 2008, Rabe et al. 2016, Oduyebo et al. 2017, Paz-Bailey et al. 2017). Data on IgM antibody persistence are limited, but recent unpublished data from a cohort of symptomatic patients in Puerto Rico showed a median time to negative IgM of 4 months (Centers for Disease Control and Prevention 2017a). Given the extensive crossreactivity between ZIKV and DENV antibodies and cocirculation of the two viruses in many parts of the world, a false ZIKV-positive result can occur in persons infected with DENV; thus the ZIKV IgM assay is considered a “screening assay.” Specimens with non-negative results for ZIKV IgM should be reflexed for testing using a plaque reduction neutralization test (PRNT) for both ZIKV and DENV antibodies (De Madrid and Porterfield 1974, Calisher et al. 1989, Oduyebo et al. 2017). PRNT measures virus-specific neutralizing antibodies, which persist for years after exposure. ZIKV-infected patients without prior DENV infection are expected to have neutralizing antibodies only to ZIKV, confirming diagnosis of a recent ZIKV infection (Rabe et al. 2016). However, those with a previous DENV infection tend to have neutralizing antibodies that react to antigens common to both viruses, making it impossible to confirm which virus caused the recent infection. The duration of IgM antibodies makes it important to consider the timing of a patient's exposure(s) when interpreting test results. IgM-positive pregnant patients with extensive travel to an endemic area may have been infected before and not during their current pregnancy (Rabe et al. 2016, Oduyebo et al. 2017).

The identification of the ZIKV outbreak in the Americas prompted an urgent need for the Centers for Disease Control and Prevention (CDC) to develop ZIKV testing guidelines. In early 2016 there were no commercial diagnostic options and limited information existed regarding optimal specimens or the duration of viral RNA or antibodies; therefore the initial CDC guidelines relied on knowledge of related viruses. Over time, new details about ZIKV and the changing epidemiology of the outbreak prompted subsequent updates to the CDC testing guidance.

The initial guidelines from early 2016 were based on available testing methodologies, aiming to optimize diagnosis of pregnant women (Rabe et al. 2016). Until July 24, 2017, NAAT on both serum and urine was recommended on specimens collected from patients meeting testing criteria ≤14 days following illness onset (for symptomatic patients) or last potential exposure through travel or unprotected sex with a traveler to an affected region for asymptomatic patients. ZIKV IgM testing was recommended when ZIKV RNA was not detected or for specimens collected within 2–12 weeks of the onset of illness or the last potential exposure (Rabe et al. 2016). This timeframe was established based on data from other flaviviruses, such as West Nile Virus and DENV and previous published data regarding ZIKV IgM antibody persistence.

In light of the waning ZIKV outbreak, in July 2017, the CDC published new recommendations, which eliminated antibody testing of potentially exposed pregnant women, and extended the timeframe for RNA testing. As of February 2018, there is no recommended test for persons whose potential exposure or illness onset occurred >12 weeks before specimen collection. Providers caring for pregnant women who have been exposed to ZIKV but are outside of the recommended testing window are advised to perform a standard evaluation at birth, unless abnormal findings are identified (Adebanjo et al. 2017).

In early 2016, CDC conducted ZIKV testing for Department of Health and Mental Hygiene (DOHMH). By late January all NAAT and ZIKV IgM testing was done by the NYC DOHMH Public Health Laboratory (PHL) and the New York State Department of Health (NYS DOH) Wadsworth Center (WC) Laboratory. PRNT is performed at WC. Both DOHMH and NYS DOH have been less restrictive since early 2016 than what was initially recommended by CDC and perform ZIKV IgM for any potentially exposed pregnant woman regardless of the timing after last potential exposure or illness onset. NAAT is performed at WC on serum and urine from pregnant women regardless of timing of specimen collection; at PHL, NAAT testing is performed on specimens collected ≤12 weeks after the date of last potential exposure or illness onset. For nonpregnant patients, testing is available for persons who developed one or more ZIKV symptoms (fever, maculopapular rash, conjunctivitis, or arthralgia) ≤4 weeks following travel to a ZIKV-affected area or an unprotected sexual exposure to a person who traveled to a ZIKV-affected area, or for persons who develop Guillain–Barré syndrome (GBS) ≤2 months following travel to a ZIKV-affected area.

Commercial laboratories began offering NAAT in May 2016 and IgM testing in September 2016; in addition to the Emergency Use Authorization requirements, NYS and NYC mandate through public health law that specimens positive by ZIKV serology and NAAT at commercial laboratories be forwarded to WC for repeat testing (New York State Department of Health and New York City Department of Health and Mental Hygiene 2016, U.S. Food and Drug Administration 2016). In November 2016, the DOHMH distributed a Health Advisory to providers encouraging ZIKV testing at commercial laboratories; PHL still offered testing for infants born to affected mothers, infants with ZIKV-related abnormalities regardless of the mother's test results, GBS patients, and persons lacking health insurance (New York City Department of Health 2016). PRNT's continue to be performed at WC. We summarized the demographic, clinical, and travel data along with the ZIKV laboratory test results of NYC residents with confirmed or probable ZIKV infection, highlighting a subset of patients with test results that demonstrate the advantages of the extended test period used by DOHMH and NYS DOH before CDC's new 2017 recommendations.

Methods

ZIKV testing for NYC residents was coordinated through DOHMH and performed at PHL and WC, with a small fraction of specimens sent to CDC early in 2016. The Zika Testing Call Center was established at DOHMH to screen NYC residents for eligibility for ZIKV testing, and, for those who met testing criteria, to prefill laboratory order forms and arrange transportation of specimens to PHL (Jacobs-Wingo et al. 2016, Lee et al. 2016). Pregnancy status, onset date, symptoms, and travel history were collected and imported into the surveillance database. All ZIKV test results from PHL, WC, and CDC, including negative test results, were electronically reported and imported into the surveillance database. As commercial testing became available, electronic reporting included ZIKV test results, but did not include patient illness onset date, pregnancy status, or dates of travel to ZIKV-affected areas.

This analysis included patients meeting the CDC/Council of State and Territorial Epidemiologists (CSTE) ZIKV national surveillance case definition (Centers for Disease Control and Prevention 2016b).

Confirmed cases had a specimen with either (1) detectable ZIKV RNA or (2) a ZIKV IgM-positive result with a positive ZIKV PRNT and negative DENV PRNT. Probable cases had a ZIKV IgM-positive result with both positive ZIKV and DENV PRNT. We define a ZIKV IgM-positive result as any of the following reported results: presumptive positive, possible positive (U.S. Food and Drug Administration 2017), positive, and equivocal (Centers for Disease Control and Prevention 2017b). This analysis focused on noncongenital cases. Each confirmed or probable ZIKV case patient was interviewed to obtain clinical and risk exposure information. If the patient was unavailable, we conducted a provider interview or medical record review. These data were entered into the surveillance database and analyzed using SAS version 9.2.

Pregnant cases were defined as having a potential exposure to ZIKV while pregnant or during the periconception period (beginning at 6 weeks before last menstrual period). For asymptomatic patients, the last possible exposure date was defined as the date of departure from a ZIKV-affected area. We used the last known date of unprotected sex as last potential exposure date if sexual transmission was identified as the route of infection; to be conservative, a patient was only categorized as a sexual transmission case if no reported travel occurred to a ZIKV-affect area. If the patient was symptomatic but symptom onset date was unknown, date of last potential exposure was used as a proxy for onset date. Those missing symptom information were counted as infections as per the CDC/CSTE case definition. For patients with more than one travel instance, but who were only tested once, the most recent trip before the collection date was used to calculate the interval from last possible exposure to collection.

Patients with unexpected laboratory test results based on timing of specimen collection and exposure or onset dates were manually reviewed and either reinterviewed to confirm exposure and onset dates, if possible, or excluded if we deemed the recorded dates were unreliable or additional laboratory tests suggested a false-positive result. For some patients with questionable test results, providers were asked to submit additional specimens; however, this was rarely done. Most specimens with detectable ZIKV RNA from a commercial laboratory were forwarded to WC for repeat testing. Specimens tested for ZIKV NAAT at PHL were confirmed with a second extraction. Asymptomatic, nonpregnant patients with only a single positive ZIKV NAAT laboratory result from a commercial laboratory were excluded from the analysis to reduce the possibility of counting false-positive results. For analysis of NAAT specimens, results on blood and plasma specimens were included with serum specimen analysis.

Results

Discussion

NYC had the largest number of travel-associated ZIKV cases in the United States (Centers for Disease Control and Prevention 2016a). The majority of cases (64%) were diagnosed by NAAT testing. More females (80%) than males were diagnosed with ZIKV, which is likely due to the screening and testing recommendations for pregnant women. Visiting friends and family was the most commonly reported reason for travel and the Dominican Republic was the most commonly reported country of travel. DOHMH targeted educational outreach efforts to residents and providers in areas with a larger population of persons born or with ancestry in the Dominican Republic and other ZIKV-affected areas (Lee et al. 2016).

We identified 108 NYC patients with ZIKV-positive specimens collected beyond the testing windows initially recommended by CDC (Petersen et al. 2016), 60 with a positive ZIKV IgM in a specimen collected >12 weeks, and 48 with detectable ZKV RNA in a specimen collected >14 days from onset or last exposure. While there are limited data on the expected duration of ZIKV IgM antibodies, the Zika Virus Persistence Study found that 87% of patients had detectable IgM antibodies >60 days after symptom onset, and a median of 122 days (range 8–210) to the first negative IgM result (Oduyebo et al. 2017, Paz-Bailey et al. 2017). The same study found the median time to the loss of detectable RNA was 14 days in serum and 8 days in urine (Paz-Bailey et al. 2017). Four asymptomatic pregnant women had ZIKV RNA detected in serum >60 days after last potential exposure (99, 85, 84, and 71 days). The longest detection of ZIKV RNA was at 97 days following illness onset in a symptomatic pregnant patient. Previously reported prolonged viremic periods for pregnant women include 107 days after symptom onset and 53 days after last potential exposure (Suy et al. 2016, St George et al. 2017). Prolonged viremia in pregnant women is hypothesized to be due to viral genetic material crossing the placenta following fetal infection and/or altered immune response during pregnancy (Kourtis et al. 2014, Meaney-Delman et al. 2016).

The expanded ZIKV testing algorithm offered by DOHMH, NYS DOH, and later commercially, identified several patients who otherwise would not have been identified following the initial 2016 CDC guidelines (Rabe et al. 2016). Among the 108 NYC patients who had positive ZIKV test results on a specimen collected outside the initial CDC recommended testing window, 19 had multiple specimens collected both within and outside of this testing window; however, the remaining 89 had specimens that were only collected outside this testing window. Of the 89, there were 57 pregnant women for whom expanded testing provided confirmation of infection status so that appropriate prenatal care and counseling could be provided (including referral to specialist care, if needed). Obtaining a correct diagnosis directs appropriate patient care and management. A positive test result in a specimen is valuable information, however, a negative result on a specimen collected outside the recommended testing window may provide a false reassurance that the patient was never infected in patients exposed early in pregnancy, but tested after the IgM response waned.

Study limitations include data accuracy regarding symptom history, onset date, travel history, and sexual exposures. Symptoms for ZIKV are nonspecific and onset dates can be difficult to recall, so time frames based on symptom onset date may be inaccurate. Patient interview was the primary source for these data, but when patients were unavailable, information was ascertained through the provider or by chart review, which might be less reliable. Data for each case with evidence of an extended ZIKV viremia, uremia, or antigenemia, were manually reviewed and in some cases we reinterviewed the patient or provider to verify dates of onset and risk exposure. However, it is possible patients failed to disclose accurate information regarding travel or sexual exposures. Current ZIKV diagnostics are imperfect and rely on appropriate screening and testing to get the most accurate results. Positive laboratory reports initiated most of our case investigations, however, given the limitations of diagnostic methodologies it is likely that not all patients who met the case definition were truly infected with ZIKV. We attempted to collect additional specimens for patients for repeat testing if there was suspicion of a false-positive result; however, this was not always possible.

In an outbreak setting of a rare infectious disease, CDC makes clinical recommendations based on the best evidence available and modifies those recommendations as our understanding of the pathogen improves through the course of the outbreak. This happened several times during the course of the 2016–2017 ZIKV outbreak. Our findings add to the growing body of evidence on ZIKV and support CDC's revised Interim Guidance, updated July 24, 2017, which recommends NAAT testing of symptomatic pregnant women up to 12 weeks following symptom onset (Oduyebo et al. 2017). Expanded testing may help confirm a diagnosis of ZIKV by NAAT rather than only being able to tell the patient of a “probable” diagnosis based on serologic testing only (positive ZIKV IgM, ZIKV PRNT positive, and DENV PRNT positive). This is increasingly important as more time elapses since the start of the outbreak to help discern an infection with ZIKV versus DENV. As ZIKV has been circulating in the Americas for over a year, many women will become pregnant after having been infected with ZIKV before pregnancy or the periconception period. A positive IgM in a currently pregnant woman who was never previously tested, but who had ongoing exposure to ZIKV both before and during pregnancy, offers multiple challenges. The timing of her infection cannot be defined (unless she has symptoms), complicating the clinical management of the pregnancy. For pregnant women who do not live in an area of active ZIKV transmission and have an exposure period restricted to their pregnancy or periconception period, IgM testing may be helpful. Given these diagnostic challenges, providers need to understand the limitations of diagnostic testing to appropriately counsel patients, especially pregnant women. Close coordination and consultation between providers and local and state health departments are important to help determine ZIKV infection status and guide clinical management, especially in pregnant women.