Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Accuracy of Zika virus disease case definition during simultaneous Dengue and Chikungunya epidemics

  • José Ueleres Braga ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Escola Nacional de Saude Publica, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil, Instituto de Medicina Social, Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brazil

  • Clarisse Bressan ,

    Roles Conceptualization, Data curation, Investigation, Methodology, Validation, Visualization, Writing – original draft

    ‡ These authors also contributed equally to this work.

    Affiliation Instituto Nacional de Infectologia Evandro Chagas, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Ana Paula Razal Dalvi ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft

    ‡ These authors also contributed equally to this work.

    Affiliation Escola Nacional de Saude Publica, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Guilherme Amaral Calvet,

    Roles Conceptualization, Data curation, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft

    Affiliation Instituto Nacional de Infectologia Evandro Chagas, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Regina Paiva Daumas,

    Roles Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Escola Nacional de Saude Publica, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Nadia Rodrigues,

    Roles Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft

    Affiliation Escola Nacional de Saude Publica, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Mayumi Wakimoto,

    Roles Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft

    Affiliation Instituto Nacional de Infectologia Evandro Chagas, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Rita Maria Ribeiro Nogueira,

    Roles Conceptualization, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – original draft

    Affiliation Laboratorio de Referencia de Flavivirus, Instituto Oswaldo Cruz, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Karin Nielsen-Saines,

    Roles Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation David Geffen UCLA School of Medicine, Los Angeles, United States of America

  • Carlos Brito,

    Roles Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Departmento de Clinica Medica, Universidade Federal de Pernambuco, Recife, Brazil

  • Ana Maria Bispo de Filippis ,

    Contributed equally to this work with: Ana Maria Bispo de Filippis, Patrícia Brasil

    Roles Conceptualization, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – original draft

    Affiliation Laboratorio de Referencia de Flavivirus, Instituto Oswaldo Cruz, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil

  • Patrícia Brasil

    Contributed equally to this work with: Ana Maria Bispo de Filippis, Patrícia Brasil

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Instituto Nacional de Infectologia Evandro Chagas, Fundacao Oswaldo Cruz, Rio de Janeiro, Brazil



Zika is a new disease in the American continent and its surveillance is of utmost importance, especially because of its ability to cause neurological manifestations as Guillain-Barré syndrome and serious congenital malformations through vertical transmission. The detection of suspected cases by the surveillance system depends on the case definition adopted. As the laboratory diagnosis of Zika infection still relies on the use of expensive and complex molecular techniques with low sensitivity due to a narrow window of detection, most suspected cases are not confirmed by laboratory tests, mainly reserved for pregnant women and newborns. In this context, an accurate definition of a suspected Zika case is crucial in order for the surveillance system to gauge the magnitude of an epidemic.


We evaluated the accuracy of various Zika case definitions in a scenario where Dengue and Chikungunya viruses co-circulate. Signs and symptoms that best discriminated PCR confirmed Zika from other laboratory confirmed febrile or exanthematic diseases were identified to propose and test predictive models for Zika infection based on these clinical features.

Results and discussion

Our derived score prediction model had the best performance because it demonstrated the highest sensitivity and specificity, 86·6% and 78·3%, respectively. This Zika case definition also had the highest values for auROC (0·903) and R2 (0·417), and the lowest Brier score 0·096.


In areas where multiple arboviruses circulate, the presence of rash with pruritus or conjunctival hyperemia, without any other general clinical manifestations such as fever, petechia or anorexia is the best Zika case definition.


To the present date 75 countries and territories have reported evidence of mosquito-borne ZIKV transmission since 2007 [1]. ZIKV continues to disperse, reinforcing the need for surveillance in countries with vector circulation. Twenty-nine countries or territories have reported microcephaly and other CNS malformations suggestive of congenital Zika infection. In addition, an increased incidence of GBS was reported by 21 countries and territories. In non-endemic countries imported cases and sexual transmission of ZIKV were also identified by health authorities [1].

Zika infection has affected countries endemic for other arboviruses, particularly Dengue and Chikungunya. The oldest and most frequent of arboviruses, Dengue has caused epidemic outbreaks for decades, with WHO estimates of more than 100 million cases per year in 100 countries where the disease is endemic [2]. Dengue is associated with risks of severe disease and deaths, but is not frequently associated with neurological complications or fetal malformations. In Brazil in 2015 there were 1,500,578 cases, with 833 deaths [3].

In 2005/2006, large outbreaks of Chikungunya took place in many countries where Dengue is endemic, starting in the Indian Ocean and Indian Islands, and by 2013 reaching the Western Hemisphere of the Caribbean island of Saint Martin. By July 2014 Chikungunya had spread to another 20 countries in the Caribbean and Central and South America, with more than 500,000 cases reported in 2016 [4, 5]. Chikungunya has as main complications significant disease severity and deaths occurring in the acute phase of illness, with frequent neurological complications [6]. Unlike the other two arboviruses, Chikungunya often results in chronic forms of disease, with about 40% of acute phase patients progressing to chronic arthritis for months to years [7].

PAHO/WHO warned of the risk of simultaneous circulation and potential outbreaks with more than one arbovirus, with epidemiological surveillance being the key element for identification and monitoring of clinically compatible cases for each one of the three arboviruses [8].

The major challenge for an epidemiological surveillance system where the three viruses are circulating simultaneously is the initial case detection in an epidemic based on the identification of a clinical picture compatible with the definition of a suspected case. With the progression of concurrent epidemics due to different co-circulating arboviruses, the surveillance system should be able to differentiate between distinct illnesses in order to enable adequate case reporting so that the data reflects the real epidemiological situation. This will allow for adequate prediction of potential complications due to each arbovirus.

The laboratory confirmation of the diagnosis of ZIKV infection can be difficult especially in countries endemic for other arboviruses such as Brazil. Detection of the virus by reverse transcription polymerase chain reaction (RT–PCR) is not always available and reference laboratories may be overloaded during epidemics. Besides, virus load levels may be low with sensitivity of molecular tests also being low [9, 10], also due to a relatively short window of detection of virus in blood or urine. ZIKV cross-reacts immunologically with Dengue and other flaviviruses, wich frequently co-circulate, thus complicating the interpretation of serological results.

Zika virus infection is the most recent of the three epidemics, and its clinical pattern of disease has been newly described. Different case definitions have been published, often based on small case series. Consequently the sensitivity of some case definitions has been questioned by clinicians [11]. For the diagnosis of Zika to be made using clinical and epidemiological criteria in resource-limited areas, or during epidemic periods, it is necessary to establish a sensitive case definition with good specificity.

An accurate case definition would optimize the allocation of human and financial resources according to the distribution of notifications and permit the prioritization of higher risk groups, such as pregnant women, or the monitoring of complications. Furthermore, it would be useful to the surveillance of imported cases in non-endemic areas and the development of clinical management protocols. Case definitions which are non-specific but with a high sensitivity can overestimate the Zika virus infection burden in a population. In this scenario, the rate of potential complications of ZIKV infection such as GBS or congenital Zika virus syndrome will be underestimated, as many of the reported "cases" would not have been true Zika cases in the first place. For example, if 10,000 cases of ZIKV are reported because of an overly sensitive case-definition, but the true number of cases is 1000, with an estimated frequency of GBS of 1% and adverse pregnancy outcome frequency of 40%, the true number of GBS cases would be 10 cases and the number of cases with adverse pregnancy outcomes 400. However, because the number of reported cases of Zika is erroneously estimated as 10,000, it appears that the frequency of GBS is 0.1% and that of adverse pregnancy outcomes 4% due to the larger denominator. This in fact has generated large discrepancies across different studies evaluating the frequency of ZIKV-related complications.

Thus, the aim of this study was to discuss the accuracy of existing case definitions of ZIKV infection and propose, based on our experience, a novel and effective case definition to be used particularly in situations where triple epidemics overlap.

Materials and method

Study population and eligibility criteria

Patients were recruited from August 2014 to May 2016 in the outpatient clinic for Acute Febrile Diseases (AFD) of the Instituto Nacional de Infectologia Evandro Chagas of the Oswaldo Cruz Foundation. During this period, all patients with a febrile illness or rash without an evident focus of infection who sougth care within seven days of onset of symptoms were invited to participate in the study.

Considering the period of documented local transmission of DENV, ZIKV and CHIKV in the municipality of Rio de Janeiro, we selected for this analysis only patients who had RT-PCR results available for confirmation of diagnosis during the epidemic period. Thus, we included patients evaluated between Aug 1 2014 and Dec 31 2014 if they had RT-PCR results for DENV; patients examined from Jan 1 2015 to Oct 31 2015 with RT-PCR results for DENV and ZIKV; and patients seen between November 1 2015 and July 31 2016 if they had RT-PCR results for DENV, ZIKV and CHIKV. Refusal to participate, absence of recorded physical examination and absence of laboratory test results were exclusion criteria.

Study design

A cross-sectional study was conducted in our outpatient population to enable the development of a clinical prediction model of ZIKV infection based on the presence of signs and symptoms. Besides the development of the derivation model, this approach allowed us to evaluate the diagnostic performance of the model and compare results with case definitions of suspected Zika proposed for surveillance by different public health organizations.

Clinical signs and symptoms and laboratory confirmation of infection

Presence of the following clinical features were assessed: maculopapular rash, itching, edema, macular rash, enanthema, lymphadenopathy, arthralgia, conjunctival hyperemia, oropharyngeal pain, earache, nasal congestion, purpura, fever, vomiting, hepatomegaly, abdominal pain, nausea, anorexia, headache, taste alteration, bleeding, prostration, lightheadedness, chills, myalgia, dyspnea, low back pain, cough, coryza diarrhea, gingivorrhagia, sweating, petechiae, hoarseness, choluria, dysuria, photophobia, retro-orbital pain and epistaxis. These signs and symptoms were evaluated in the first medical visit by history and clinical examination recorded in structured case report forms. Data on signs and symptoms present on the first or second clinic visit within the first week of disease was collected.

Confirmatory diagnosis of infection for ZIKV, DENV and CHIKV was made by real time PCR test of blood or urine specimens obtained during the same time period [10]. Dual or triple co-infections were excluded from analysis. Other febrile illnesses (OFI) were diagnosed based on negatives PCR results for all circulating arbovirus. Patients included before 2015 and those who tested negative for ZIKV by RT-PCR were classified as Zika-negative.

Development of the clinical prediction model and its performance

Odds ratio (OR) was used as the association measure to identify the signs and symptoms related to ZIKV infection. Crude and adjusted estimates in single and multiple regression models were obtained using STATA 13·0 software. 95% confidence intervals (CI) were calculated using the Mantel-Hanszel method. Explanatory variables with a marginal association with the outcome (p≤0·20) in the bivariate analysis were included into the multiple regression models. A stepwise multiple logistic regression was conducted with backward selection to identify the prediction model, with a significance level of 5%.

Clinical prediction model and evaluation of suspected case definitions

We compared the predictive models that presented the best fit by the likelihood ratio test (LR test) and R2 (coefficient of determination). We studied six (6) suspected case definitions of Zika used in epidemiological surveillance by the following groups: the Pan American Health Organization (PAHO) 2015 [12], PAHO 2016 [13], the European Center for Disease Prevention and Control (ECDC) 2016 [14], the Centers for Disease Control and Prevention (CDC) 2016 [15], the World Health Organization (WHO) 2016 [16] and the case definition used by the Brazilian Ministry of Health in 2016 [17]. A score based rule derived from the selected prediction model was evaluated and compared with the Zika suspected case definitions. The scores were defined by the ratio between the odds ratio for each associated signal and the one of smaller magnitude of the derived model. The comparisons were done using measures of accuracy sensitivity, specificity, diagnostic odds ratio, area under receiver operating characteristic (auROC) curve and corresponding 95% confidence intervals. The best cut-off value for the score was determined by the Youden index that maximizes the sum of sensitivity and specificity.

Ethical issues

This study was approved by the Institutional Review Board of the Instituto Nacional de Infectologia Evandro Chagas of the Oswaldo Cruz Foundation.


Socio-demographic characteristics of the study population

Among 659 participating subjects, 138 were positive by RT-PCR for ZIKV, 113 for DENV, 45 for CHIKV and 363 were diagnosed with OFI. The laboratory confirmation of cases was based on RT-PCR tests, which were performed for all circulating arboviruses in each period result of all cases. No patients included in the study had dual or triple infections. The mean age of the study population was 37 years (range: 9 to 78 years for Zika; 36 years (range 10 to 66 years for dengue; 43 years (range 17 to 72 years for chikungunya and 35 years (range 4 to 80 years for OFI).

Among subjects positive for ZIKV, 90 (65·2%) were women, 88 (63·8%) self-reported as white and 46 (33·3%) had a university degree. For dengue and OFI, most cases were in males and the majority of the study population self-reported as white and about one third reported having a university degree (Table 1).

Table 1. Socio-demographic characteristics of the population studied according to the presence of Zika, dengue or chikungunya diagnosis.

Clinical signs and symptoms associated with ZIKV disease

The simple logistic regression results are presented in Table 2.

Table 2. Clinical signs and symptoms associated with ZIKV infection.

Clinical prediction models of ZIKV disease

Three nested models with four, five and six clinical predictors were compared in order to identify the one with the best goodness-of-fit. Model 1 was chosen because it presented the highest R2 value and was different from model 2 as indicated by the LR test. The selected model for case definition for ZIKV included six signs and symptoms and was converted in a score based on their coefficients (Fig 1). Considering the best cut-off value, a score equal or higher than 7·5 was predictive of Zika virus disease (Table 3).

Fig 1. Prediction models area under ROC based on signs and symptoms of ZIKV infection.

Table 3. Clinical definitions of a suspected Zika case for surveillance purposes.

The multiple logistic regression results are presented in Table 4.

Table 4. Derived clinical prediction models of ZIKV infection.

As researchers consider the auROC measure of accuracy as the best measure of performance for diagnostic tests, we compared the model derived from our study with the other definitions used for a suspected case of Zika. The graph (Fig 1) indicates the superiority of the screening tool derived in our study in relation to others. It is noteworthy that the definition currently used in Brazil has an area of a little more than 0·5, which represents a lower power to discriminate suspected cases.

The suspected Zika case definitions were evaluated by significant differences in accuracy measurements (Table 5).

Table 5. Accuracy of Zika case definitions in simultaneous epidemics of dengue and chikungunya.


In this study, we formulated and tested a clinical based score to discriminate Zika cases from other febrile or exhantematic illnesses in a scenario where dengue and chikungunya viruses also circulate. The final score presented a sensitivity of 86·6% and a specificity of 78·3% and performed better than all other existing zika suspected case definitions to which it was compared.

In the definition proposed following our analysis, the symptoms most strongly associated with zika vírus disease were selected. These included presence of rash, pruritus, conjunctival hyperemia and absence of fever (axillary temperature less than 37·5°C), no petechiae, no anorexia.

Axillary temperature less than 37·5°C, petechiae and anorexia, signs and symptoms that are not included in existing suspected ZIKA case definitions, are included in the derived score prediction model with the best performance. This definition had the highest sensitivity and specificity, and also the highest diagnostic OR (23·4) value and auROC (0·903), as well as the lowest Brier score of 0·096.The Brier score approaching zero indicates that the Zika case infection definition proposed in our model is the one that best discriminates true cases of Zika symptomatic infection in our population.

The concurrent circulation of several arboviruses in many countries with endemic disease makes it difficult to differentiate between dengue, zika and chikungunya, which requires the use and standardization of more sensitive and specific criteria for definitions of suspected cases, in order to avoid notification errors and distortions in the analysis of arbovirus behavior.

The criterion derived from our prediction model did not include arthralgia, a symptom that was unable to discriminate between Zika and OFI, being frequent in both groups, with a prevalence of 69·6% and 56·3%, respectively. In fact although arthralgia is characteristic of Chikungunya, mainly due to its intensity and its longer duration (parameters not evaluated in the present analysis), it has been described with high frequency in other arboviral infections, as well as influenza and OFI [18]. In dengue, arthralgia is a symptom included in the WHO and Brazil MOH case definitions [19, 20] with frequencies above 70% being described in some studies [18, 21].

Fever (axillary temperature ≥37·5°C), a criterion present in most definitions, was inversely associated with Zika, as well as the presence of anorexia and petechiae, the more frequent manifestations of dengue fever and OFI prevalent in the study population.

Rash is not a mandatory symptom in the PAHO and CDC Zika case definitions, although it is listed, along with elevated fever, arthralgia and conjunctivitis as a common clinical manifestation of Zika virus infection. In both definitions, sensitivity was high (> 80%) and specificity was very low (<11%).

The problem of a criterion with high sensitivity and low specificity in countries with a triple arboviral epidemic is that it will not be able to discern which infection is responsible and will inevitably lead to a higher notification of Zika false positives, distorting the actual epidemiological data. This can lead to potentially unnecessary investigation of specific high risk groups, such as pregnant women for example, who will often turn up negative. In the CDC-2016 case definition, the simple presence of fever (measured or referred) defines a suspected case of Zika with a specificity of 1·4%, as fever is a very common symptom in almost all cases of Dengue and Chikungunya [19, 21] and common in most OFI in endemic and epidemic countries [18]. In the PAHO-2015 case definition which uses a mixed model, there was a slight increase in specificity (10·9%) by including, in addition to the presence of fever or rash, another symptom. However headache and malaise are considered non-specific and are also frequently seen in OFI, making it difficult to discern actual cases of Zika.

The two case definitions with the lowest sensitivity for detecting Zika cases were the ones used by PAHO and the Brazil–MoH. These fail the model for safe surveillance, as they may lead to a lower suspicion of true Zika cases which enables the detection of early cases and triggers an epidemic alert. However, although not sensitive, these definitions will likely report the highest number of true cases in triple-epidemic regions, because of acceptable degrees of specificity.

Despite the simultaneous circulation of more than one arbovirus [3, 22, 23] epidemics tend to produce different incidence patterns each year, with one arbovirus predominating over the others. The low sensitivity of Zika case definitions in Brazil allows many cases of Zika to be erroneously attributed to dengue fever. When using clinical criteria for the definition of a suspected case of dengue, 48% of confirmed cases of Zika in this study would have been reported as Dengue to the surveillance system. The misreporting of arboviral cases has serious consequences not only for the surveillance system but also for patient management and understanding of the pathogenesis of disease.

Our study, carried out in non-pregnant women and men, proposes three nested models. Maculopapular rash, itching, petechiae and temperature lower than 37·5C are signs common to all models. Anorexia was included in the first two models and non-purulent conjunctivitis, figures only in the first model. Model 1 showed a greater capacity of predicting Zika disease. Although this model requires a higher number of signs to be investigated in the potential Zika patient, it does not pose significant difficulty for its application in epidemiological surveillance.

The proposed definition originated from prediction models generated from data collected in our Acute Febrile Illness outpatient clinic over three years. These data are extremely rich because they reflect the epidemiological complexity of Rio de Janeiro in this period, which makes the differential diagnosis of viral diseases more difficult. On the other hand, it is not possible to determine if the predictive validity of the definition generated from these data will be reproducible in other scenarios, especially where there is circulation of other diseases that share some of Zika's symptomatology.

Surveillance systems use clinical and epidemiological criteria; therefore, in addition to clinical information, the epidemiological data on seasonality, laboratory confirmation of the first suspected cases, contact history, origin and dispersion rates are part of a suspected case definition. Some cases can be incorrectly classified, but this may have little impact in terms of public health, unless the classification system has so many limitations that high case numbers following epidemics with high attack rates are not perceived. The main objective in defining suspected Zika cases during outbreaks is to differentiate this condition from other epidemic arboviruses in a triple-epidemic scenario. Furthermore, the need to correctly identify circulating Zika virus is extremely important given its high pathogenicity to the developing fetus. As an unknown proportion of Zika cases are in fact asymptomatic, efforts to identify symptomatic disease are important and assist with sentinel surveillance.


The analysis of clinical findings in a sample of laboratory confirmed cases of arboviral infections enabled us to generate a clinical case definition model for suspected Zika infections which has the best sensitivity and specificity. This model is useful for countries experiencing triple arboviral epidemics. Based on our findings, we suggest that the official definitions of suspected Zika cases be reviewed.


  1. 1. World Health Organization. Situation Report. Zika virus microcephaly, Guillain-barré syndrome, January 05, 2017. (accessed Jan 13, 2017).
  2. 2. World Health Organization. Global strategy for dengue prevention and control 2012–2020. (accessed Jan 13, 2017).
  3. 3. Brasil. Secretaria de Vigilância em Saúde − Ministério da Saúde. Boletim epidemiológico. Monitoramento dos casos de dengue, febre de chikungunya e febre pelo vírus Zika até a Semana Epidemiológica 37, 2016 (in portuguese). (accessed Jan 13, 2017).
  4. 4. Morrison TE. Reemergence of chikungunya virus. J Virol. 2014;88(20):11644–7. pmid:25078691
  5. 5. Pan American Health Organization / World Health Organization. Number of Reported Cases of Chikungunya Fever in the Americas, by Country or Territory 2016 (to week noted), Cumulative cases, Epidemiological Week / EW 2 (Updated as of 13 January 2017). (accessed Jan 13, 2017).
  6. 6. Renault P, Solet JL, Sissoko D, Balleydier E, Larrieu S, Filleul L, et al. A major epidemic of chikungunya virus infection on Reunion Island, France, 2005–2006. Am J Trop Med Hyg. 2007;77(4):727–31. pmid:17978079
  7. 7. Rodriguez-Morales AJ, Calvache-Benavides CE, Giraldo-Gomez J, Hurtado-Hurtado N, Yepes-Echeverri MC, Garcia-Loaiza CJ, et al. Post-chikungunya chronic arthralgia: Results from a retrospective follow-up study of 131 cases in Tolima, Colombia. Travel Med Infect Dis. 2016;14(1):58–9. pmid:26419952
  8. 8. Pan American Health Organization / World Health Organization (PAHO / WHO). Epidemiological Alert Chikungunya and dengue fever in the Americas, 29 August 2014. (accessed Jan 13, 2017).
  9. 9. Duffy MR, Chen TH, Hancock WT, Powers AM, Kool JL, Lanciotti RS, et al. Zika virus outbreak on Yap Island, Federated States of Micronesia. N Engl J Med. 2009;360(24):2536–43. pmid:19516034
  10. 10. Lanciotti RS, Kosoy OL, Laven JJ, Velez JO, Lambert AJ, Johnson AJ, et al. Genetic and serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007. Emerg Infect Dis. 2008;14(8):1232–9. pmid:18680646
  11. 11. Brito CA, Cordeiro MT. One year after the Zika virus outbreak in Brazil: from hypotheses to evidence. Rev Soc Bras Med Trop. 2016;49(5):537–43. pmid:27812646
  12. 12. Pan American Health Organization / World Health Organization. Epidemiological Update: Zika virus infection, 16 October 2015.〈=en. (Accessed Sep 23, 2016).
  13. 13. Pan American Health Organization / World Health Organization. Case Definitions.〈=en (accessed Sep 23, 2016).
  14. 14. European Centre for Disease Prevention and Control. Case Definition. (accessed Sep 23, 2016).
  15. 15. Centers for Disease Control and Prevention. Zika Virus Disease and Zika Virus Infection 2016 Case Definition, Approved June 2016 (accessed Sep 23, 2016).
  16. 16. World Health Organization. Zika virus disease: interim case definition,12 February 2016 (accessed Sep 26, 2016).
  17. 17. Brasil. Ministério da Saúde. Secretaria de Vigilância em Saúde. Procedimentos a serem adotados para a vigilância da Febre do vírus Zika no Brasil (in portuguese). (accessed Sep 23, 2016).
  18. 18. Lorenzi OD, Gregory CJ, Santiago LM, Acosta H, Galarza IE, Hunsperger E, et al. Acute febrile illness surveillance in a tertiary hospital emergency department: comparison of influenza and dengue virus infections. Am J Trop Med Hyg. 2013;88(3):472–80. pmid:23382160
  19. 19. World Health Organization. Geneva, Switzerland: WHO; 2009. Dengue: Guidelines for Diagnosis, Treatment, Prevention and Control. (accessed Jan 13, 2017).
  20. 20. Brasil. Ministério da Saúde, Secretaria de Vigilância em Saúde, Departamento de Vigilância das Doenças Transmissíveis. Dengue: diagnóstico e manejo clínico: adulto e criança [recurso eletrônico] /– 5. ed.–Brasília: Ministério da Saúde, 2016. (accessed Jan 13, 2017).
  21. 21. Rezza G, El-Sawaf G, Faggioni G, Vescio F, Al Ameri R, De Santis R, et al. Co-circulation of Dengue and Chikungunya Viruses, Al Hudaydah, Yemen, 2012. Emerg Infect Dis. 2014;20(8):1351–4. pmid:25061762
  22. 22. Cardoso CW, Paploski IA, Kikuti M, Rodrigues MS, Silva MM, Campos GS, et al. Outbreak of Exanthematous Illness Associated with Zika, Chikungunya, and Dengue Viruses, Salvador, Brazil. Emerg Infect Dis. 2015;21(12):2274–6. pmid:26584464
  23. 23. Pessoa R, Patriota JV, Lourdes de Souza M, Felix AC, Mamede N, Sanabani SS. Investigation Into an Outbreak of Dengue-like Illness in Pernambuco, Brazil, Revealed a Cocirculation of Zika, Chikungunya, and Dengue Virus Type 1. Medicine (Baltimore). 2016;95(12):e3201.