The protocol for this investigation was reviewed by human subjects’ research advisors at the Centers for Disease Control and Prevention (CDC) and was deemed to be public health practice, not research. As such, institutional review board review was not required.

Patients with suspected ZIKV disease were reported to PRDH [5]. In brief, patients for whom a clinician suspected ZIKV disease had case report forms collecting demographic and clinical information submitted to PRDH, along with a serum specimen for diagnostic testing by real-time reverse transcriptase polymerase chain reaction (rRT-PCR) and/or immunoglobulin M antibody capture enzyme-linked immunosorbent assay (MAC-ELISA) to detect evidence of infection with ZIKV, DENV, and chikungunya virus [36, 37]. Copies of medical records were requested for (1) patients who tested positive for ZIKV infection by rRT-PCR or MAC-ELISA; (2) patients who had a reported date of illness onset between January 1 and December 31, 2016; and (3) patients for whom the variable “thrombocytopenia” (unquantified) on the case report form was reported in the affirmative.

Medical records were reviewed to confirm thrombocytopenia, as evidenced by a platelet count <100 ×10⁹/L [28, 29]. Such cases were defined as having confirmed thrombocytopenia. Medical records of cases with confirmed thrombocytopenia were abstracted to collect information on demographics, history of illnesses, clinical interventions, and outcomes. Severe thrombocytopenia was defined by (1) platelet count <20 ×10⁹/L or (2) platelet count <50 ×10⁹/L, along with a clinical diagnosis of and treatment for ITP (ie, administration of intravenous immunoglobulin [IVIG] or steroids). Cases with confirmed thrombocytopenia that did not meet the case definition of severe thrombocytopenia were defined as having nonsevere thrombocytopenia.

Cases with a platelet count <100 ×10⁹/L and an alternative etiology of thrombocytopenia were excluded from further analysis. Patients taking medications that could affect platelet count (eg, aspirin, antiplatelet agents, ranitidine, antidepressants) were only excluded if there was clinical documentation of thrombocytopenia as a result of the medication. Cases without another etiology of thrombocytopenia identified were defined as having ZIKV-associated severe or nonsevere thrombocytopenia.

Anatomical sites and severity of bleeding were abstracted. Bleeding manifestations were graded from 0 to 5 based on severity using the ITP-specific bleeding assessment tool (ITP-BAT) [35]. In summary, grade 2 was clinically significant bleeding requiring at least outpatient care. Grade 3 required hospital admission or surgical intervention in response to bleeding. Grade 4 required red blood cell transfusion or a decrease in hemoglobin of >2 g/dL. Grade 5 was fatal bleeding. A patient with evidence of bleeding (eg, decrease in hemoglobin >2 g/dL) but no objective evidence of bleeding was defined as having an occult hemorrhage. Available information from blood smear reports was reviewed for platelet number and size, fragmented red blood cells including schistocytes and other abnormalities to identify malignancy, DIC, pseudothrombocytopenia, or other etiologies of thrombocytopenia.

Cases were classified as confirmed or possible ITP. Confirmed ITP was defined by a clinical discharge diagnosis or end-of-visit diagnosis of ITP. Possible ITP was defined by (1) the absence of other causes or disorders that may be associated with thrombocytopenia except ZIKV and (2) a clinical course consistent with ITP as determined by a hematologist (E.V.) after medical record review.

Cases of severe thrombocytopenia were evaluated for the primary end points of platelet count 96 hours after administration of IVIG or corticosteroids, platelet count immediately after transfusion, and clinical response. Clinical response was defined as a platelet count ≥30 ×10⁹/L and ≥2-fold greater than baseline and the absence of bleeding [29]. Ninety-six-hour platelet count after ITP treatment was selected because peak response to IVIG occurs 2–7 days postadministration, and initial response to corticosteroids occurs 2–14 days after administration [28, 29]. Grade of bleeding manifestations was a secondary outcome [29].

The 2016 US Census population estimate for Puerto Rico (3 411 307 residents) was used to calculate annual incidence [38]. Data were collected and managed using Research Electronic Data Capture (REDCap; Nashville, TN) and analyzed with SAS, version 9.4 (Cary, NC).

As of February 1, 2018, 13 cases of severe thrombocytopenia (platelet count <20 ×10⁹/L without other underlying etiology) or ITP associated with ZIKV infection had been reported [10–15, 17]. This report nearly doubles the number of documented cases of severe ZIKV-associated thrombocytopenia that could be due to ITP. Moreover, this was the first investigation to determine the population-based incidence of ZIKV-associated thrombocytopenia. With 1.4 cases per 100 000 population and among 0.1% of patients reported with ZIKV disease, thrombocytopenia appears to be a rare but potentially severe manifestation associated with ZIKV infection. This analysis provided strong evidence that at least 33 patients with ZIKV-associated thrombocytopenia were possible cases of ITP and 6 were confirmed cases of ITP. Among the 12 patients with severe thrombocytopenia identified, 11 received treatment with IVIG or corticosteroids, and 9 responded to clinical treatment.

Similar to the patients identified in this investigation, case series from other countries reported patients with ZIKV infection and thrombocytopenia having acute phase bleeding manifestations, including gum bleeding, skin hematomas, oral mucosal bleeding, and hematuria [10–12, 14, 23, 33]. Blood abnormalities related to ZIKV infection are only sporadically reported but include thrombocytopenia, mild leucopenia, and the presence of activated lymphocytes [4]. In past case series of thrombocytopenia associated with ZIKV infection, bone marrow biopsies were consistent with ITP and Evan’s syndrome (hypercellularity and without dysplasia) [10, 12]. Findings from blood smears showed macroplatelets and marked reductions in platelets. One blood smear in a patient with nonsevere thrombocytopenia was consistent with hemolytic anemia.

Previous estimates indicate that up to 1.4% of adults and 0.4% of children with ITP experience intracranial hemorrhage, whereas 15% of patients with ITP experience other forms of severe bleeding [29, 39]. Predictors of severe bleeding include platelet count <20 ×10⁹/L, newly diagnosed ITP, and previous episodes of bleeding [39]. Among the 12 patients with severe thrombocytopenia reported here, death occurred in 1 patient. This patient was elderly, had intracranial hemorrhage, and detection of thrombocytopenia and treatment occurred late in the clinical course. In elderly patients with ITP, although platelet counts are similar to those of younger age groups, prognosis can be poor [40]. Therefore, due to the risk of fatal outcome among elderly patients with ZIKV-associated thrombocytopenia, heightened vigilance and rapid treatment are warranted to prevent severe bleeding and minimize mortality.

Among the 12 patients with severe thrombocytopenia, response to platelet transfusion seemed minimal. According to the evidence-based practice guidelines, in newly diagnosed pediatric patients with ITP, initial management recommendations for patients with no or mild bleeding should include observation alone regardless of the platelet count [28]. Initial pharmacologic treatment recommendations when indicated in pediatric patients include IVIG or a short course of corticosteroids [28]. In adults, treatment is recommended for patients with a newly diagnosed platelet count <30 ×10⁹/L, and longer-course corticosteroids are the preferred firstline treatment with IVIG used in conjunction when a more rapid platelet increase is required [28]. IVIG and Rho(D) immune globulin (anti-D) is the firstline treatment when corticosteroids are contraindicated [28]. Five out of 6 (83%) patients diagnosed with ITP who received IVIG and/or corticosteroids responded to treatment. Moreover, these responses were consistent with initial and peak responses to IVIG of 1–3 and 2–7 days, respectively, dexamethasone of 2–14 and 4–28 days, and prednisone of 4–14 and 7–28 days [28, 29]. Platelet transfusions may be indicated in severe and life-threatening bleeding when a rapid rise in platelet count is needed to achieve adequate hemostasis.

This investigation was subject to limitations. First, because not all medical records were available for review and not all patients with ZIKV-associated severe thrombocytopenia were likely reported, the identified incidence of thrombocytopenia is likely an underestimate of the true incidence. Second, because diagnostic testing for patients with thrombocytopenia is not standardized across hospitals in Puerto Rico, etiologies of thrombocytopenia other than ZIKV may have been responsible for thrombocytopenia in some patients. Similarly, half of the cases with thrombocytopenia only had serologic evidence of ZIKV infection. Because the duration of anti-ZIKV IgM antibody has not been defined but is likely to be at least several months, some individuals with residual IgM antibody may have been misclassified; however, 90% of patients had signs and symptoms consistent with ZIKV infection within 1 week of thrombocytopenia. Last, although anti-ZIKV IgM antibody may cross-react with DENV antigen and vice versa [36], DENV transmission was at a historic low in Puerto Rico in 2016 [6]. Hence, all patients in this report for which anti-ZIKV IgM antibody was detected were likely to have been the result of infection with ZIKV and not DENV.

Though rare, ZIKV-associated severe thrombocytopenia can be fatal. Treatment of ZIKV-associated thrombocytopenia might be indicated in some cases based on clinical expertise and recommendations. As demonstrated here, platelet transfusion alone does not appear to be sufficient to elicit an appropriate clinical response, whereas administration of corticosteroids, IVIG, or other ITP treatment could benefit some patients. Further investigation is needed to identify the mechanism of pathogenesis in patients with ZIKV-associated severe thrombocytopenia. This aspect is also relevant in the design of a vaccine to prevent ZIKV infection, as individual viral epitopes may be responsible for antibodies that cross-react with platelets and lead to severe thrombocytopenia [41]. Regardless of the mechanism, the findings from this investigation demonstrate that timely diagnosis of ITP among patients with ZIKV-associated severe thrombocytopenia is crucial to initiating life-saving interventions.

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.