Journal of Medical Virology 86:2005–2010 (2014)
NS1 Antigen Testing for the Diagnosis of Dengue in Returned Israeli Travelers Inbal Fuchs,1* Hana Bin,2 Sara Schlezinger,2 and Eli Schwartz3 1
The Pediatric Infectious Disease Unit, Soroka University Medical Center, Ben-Gurion University of the Negev, Beer-Sheva, Israel 2 National Center for Zoonotic Viruses, Central Virology Laboratory, Ministry of Health, Public Health Services, Sheba Medical Center, Tel Hashomer, Israel 3 Center for Geographic Medicine and Tropical Diseases, Sheba Medical Center, Tel Hashomer, Israel
Antigen testing with NS1 provides rapid diagnosis of dengue in endemic regions during the febrile phase of illness before appearance of IgM in serum. This study aimed to determine the diagnostic accuracy of NS1 antigen testing in travelers presenting with febrile illness and serologically confirmed dengue infection, upon return from dengue endemic countries to Israel, a region endemic for West Nile virus (WNV). Cases were sera obtained from febrile returning travelers with positive dengue-IgM antibodies. Sera of non-travelers with confirmed WNV disease and sera of returning travelers with confirmed non-dengue febrile illnesses were used as controls. All sera were tested for NS1 antigen using the Panbio Dengue Early ELISA assay within 21 days of symptoms. Demographic data, travel destination, and interval between disease onset and testing were retrieved from patient files. Fifty-eight sera from 40 dengue-infected travelers, 26 sera from 26 WNV- infected patients, and 15 sera of returning travelers with non-dengue febrile illness were tested. Sensitivity of NS1 testing in dengue patients was 87% during the first 3 days of symptoms and declined to about 70% after 12 days. No cases tested positive for NS1 after day 12. Specificity was 92% for the entire testing period. The NS1 Panbio assay is sensitive for the detection of dengue viral infection in returning travelers during the febrile phase of illness, and is highly specific in a region where WNV co-circulates. J. Med. Virol. 86: 2005–2010, 2014. # 2014 Wiley Periodicals, Inc. KEY WORDS:
dengue; diagnosis; non-structural protein 1; traveler
INTRODUCTION Dengue fever (DF) has become one of the world’s most prevalent arthropod-borne viral illnesses. AccordC 2014 WILEY PERIODICALS, INC.
ing to the World Health Organization (WHO), 40% of the world’s population lives in a dengue virus (DENV-14) endemic area [WHO, 2009]. Every year, 50–100 million people are exposed to the virus, with approximately 25,000 fatalities [Guzman and Isturiz, 2010]. The disease is caused by the DENV-1-4, a single stranded ribonucleic acid (RNA) virus of the family Flaviviridae in the genus Flavivirus, the most important vectors being Aedes aegypti and Aedes albopictus. Dengue (DENV) is also one of the fastest spreading vector-borne diseases in the world due to population growth, uncontrolled urbanization, spread of mosquito vectors and rising international mobility [Kyle and Harris, 2008; WHO, 2009]. International travelers may acquire and spread DENV infection [WilderSmith and Schwartz, 2005]. Sero-survey studies have shown high sero-conversion rates after travel to endemic regions, demonstrating that DENV is one of the most common infections among travelers [Potasman et al., 1999; Cobelens et al., 2002]. Laboratory diagnosis methods for confirming DENV infection have traditionally involved detection of the virus, viral nucleic acid, or antibodies. These methods are not ideal for diagnosis of acute infection in returning travelers for several reasons. First, viral isolation, as well as polymerase chain reaction (PCR), are not routine methods in most medical facilities in non-endemic countries, as these are costly laborintensive procedures, requiring trained personnel for a low volume of tests. The short period of viremia in DENV infection presents another limitation for virus
The authors declare that they have no conflict of interest, financial or other, that might influence the contents of the manuscript. Correspondence to: Inbal Fuchs, MD, The Pediatric Infectious Disease Unit, Soroka University Medical Center, Beer Sheva, Israel. E-mail: [email protected] Accepted 12 December 2013 DOI 10.1002/jmv.23879 Published online 6 January 2014 in Wiley Online Library (wileyonlinelibrary.com).
2006
isolation or RNA detection in returning travelers. Culture and molecular diagnostics are most useful in detection of DENV in serum, circulating blood cells, and other tissues, when the tests are performed within 4–5 days of symptom onset. This window for virological diagnosis may be missed in returning travelers [Simmons et al., 2012]. Finally, two recent reports involving returning travelers with serologically confirmed DENV infection reported successful viral isolation in 69% of patients who presented during the first 5 days of disease, and in 86% in sera tested during the first 3 days of symptoms using viral culture and realtime PCR, respectively [Teichmann et al., 2004; Huhtamo et al., 2010]. Thus, molecular methods are not always confirmatory, even in early sera. As a result, diagnosis of DENV infection in travelers returning to non-endemic regions is usually based on clinical suspicion and confirmed with serological tests. However, clear-cut, bedside diagnosis of a febrile patient is limited by the delays inherent in seroconversion. In serum, IgM antibodies can usually be identified within 5–7 days of symptom-onset and IgG antibodies within 7–15 days following primary infections [Schwartz et al., 2000]. Serology-based diagnosis is also severely limited by cross-reactivity with other flavivirus infections or vaccination (e.g., yellow fever, Japanese encephalitis, tick-borne encephalitis [TBE]) [Schwartz et al., 2000; Lindegren et al., 2005; WHO, 2009]. This cross-reactivity can result in diagnostic uncertainties when a traveler returns from a DENV-endemic area and develops a febrile illness during transmission season of a co-circulating flavivirus endemic to the traveler’s home country. The development of a reliable test based on the detection of non-structural protein 1 (NS1) provided an important breakthrough in DENV diagnosis [Young et al., 2000; Alcon et al., 2002]. Antigen detection using NS1 has been used in DENV endemic areas in recent years as a less expensive and more easily accessible alternative to molecular testing. The secreted glycoprotein can be detected in high levels in the serum of DENV infected patients during the febrile phase of disease even when virus can no longer be isolated [Chaiyaratana et al., 2009; Duong et al., 2011]. Results of NS1 testing in endemic areas showed sensitivity in a range of 60–73% when sampling occurred within the first week following symptom onset [McBride, 2009; Ramirez et al., 2009]. Specificity data among returning travelers are scarce, and do not exist for travelers returning to countries with endemic flaviviruses. West Nile virus (WNV), the most important flavivirus in Israel, is endemic in parts of Africa, Europe, the Middle East and Asia. Outbreaks of disease also occur annually in North America [Sotelo et al., 2012]. Cross-reactivity between DENV-1-4 and WNV in serological tests has been described [Papa et al., 2011]. This study describes the performance of NS1 antigen testing in travelers returning to Israel, a country where WNV is endemic. J. Med. Virol. DOI 10.1002/jmv
Fuchs et al.
MATERIALS AND METHODS In Israel DENV serology is performed mainly at the Israeli National Center for Zoonotic Viruses (INCZV) at Sheba Medical Center, Tel Hashomer. The NS1 study was performed on sera of patients with acute DENV infection who presented during the years 2007–2011 (cases) and on sera of two control groups within 21 days of the onset of febrile illness. Intervals from symptom onset to testing were documented in the INCZV records for all serum samples obtained. This study was approved by the institutional ethical board of Sheba Medical Center with an exemption from informed consent. Case Definition Patients who presented with clinical illness compatible with DENV infection after visiting DENV endemic regions, with detectable DENV IgM in at least one serum sample, were defined as having acute DENV infection. Patients with positive DENV IgG antibodies alone were not included. Control Group 1 Definition Control group 1 consisted of serum samples obtained from Israeli patients with laboratory confirmed WNV infection, who had not traveled to countries endemic for DENV. A confirmed WNV case was defined in concordance with clinical and epidemiological data, after ruling out any other flavivirus infection or vaccination. Serological testing for WNV was carried out by the InBios West Nile DetectTM IgM Capture ELISA, and InBios West Nile DetectTM IgG ELISA (Seattle, WA); commercial kits simultaneously. A case was confirmed when one or more of the following occurred: (1) High levels of IgM in the cerebrospinal fluid, regardless of IgG result. (2) A fourfold increase in the kit’s units between paired samples collected at least 7 days apart. (3) Seroconversion of IgM and/or IgG from undetectable to a positive reading between paired samples collected at least 7 days apart. Data on co morbidities and medical treatment were available from patient files. Control Group 2 Definition Control group 2 consisted of serum samples obtained from returning travelers with a febrile illness other than DENV. All diagnoses were based on history of exposure, compatible clinical presentation, and confirmed by laboratory methods (malaria smear for malaria, blood culture for typhoid fever, microscopic agglutination test for leptospirosis, real-time PCR from a nasal swab for influenza, and IgM seroconversion for Epstein-Barr virus and cytomegalovirus). For case patients and control group 2 patients, individual files were retrieved from the Center of Travel and Tropical Medicine, at Sheba Medical Center. Data regarding for clinical presentation, travel destination, and interval between day of
NS1 Dengue Diagnosis in Travelers
2007
sampling and disease onset, were re-captured and confirmed. NS1 Antigen Testing and DENV Serology Testing Tests were performed according to manufacturer’s instructions using the second generation Panbio Dengue Early ELISA E-DENO2P (Inverness Medical Innovations Sinnamon Park, QLD, Australia). Dengue capture IgM, and dengue IgG indirect were performed according to manufacturer’s instructions (Panbio, Brisbane, Australia). RESULTS Fifty-eight sera from 40 patients who returned from DENV-endemic countries with positive DENVIgM serology during 2007–2011 were included in the study-group as cases. Eight patients (20%) had repeated serum testing which showed seroconversion. In the remainder of the returning travelers, a diagnosis of acute DENV infection was based on characteristic presentations of DF (clinical and laboratory findings) that appeared within a week of their return from a DENV-endemic region, and was further confirmed by positive DENV IgM. Twenty-seven case patients (68%) were males. Information regarding travel destination was available for 38 patients, of whom 35 acquired the infection in South-east Asia (Thailand, Laos Vietnam, and Cambodia) or India. Other destinations were: Costa Rica (1 patient) Bolivia (1 patient) and Papua New Guinea (1 patient). A positive NS1 result was observed in 34/58 (58%) of sera obtained from case patients during the 21-day interval. The test sensitivity was highest during the first 3 days, reaching 87.5%, declined to 70% during days 10–12, and became completely negative from day 13 onward (Fig. 1). Sera drawn from a patient on separate days from day 1 to day 8 of illness, showed a peak in the reading of NS1 panbio units on day 3 (Fig. 2).
Fig. 2. Consecutive testing of serum NS1 in a patient infected with DENV.
Two control groups were included in the study. In the first group, sera from 26 patients with confirmed WNV infection during 2009 and 2011 were tested for NS1. All sera were negative for NS1 except two samples (7.6%), which had a positive and a borderline NS1 test result. Results for the two samples in Panbio units were 19.45 and 9.19, respectively. In the assay used in this study, Panbio units are calculated as optical density (OD) (absorbance sample/OD cut off value) 10. Panbio units in the range of 9–11 are interpreted as equivocal. The WNV patient with a positive NS1 result had multiple sclerosis, for which she had received intravenous immunoglobulins 2 weeks prior to her admission. The WNV patient with a borderline NS1 result was a 65-year-old man who had not traveled outside of Israel. Repeat testing was not performed on the serum with a borderline result. The second control group included sera obtained from 15 returning travelers who were diagnosed with a febrile illness other than DENV during 2009–2012. The diagnoses were distributed as follows: leptospirosis (five patients), malaria (four patients), typhoid fever (three patients), mononucleosis due to EpsteinBarr virus, acute cytomegalovirus infection, and acute influenza virus infections (one patient each). Ten patients (66%) were male. Travel destinations were known for 12 of the 15 patients and were: South-east Asia/India (seven patients), Africa (four patients), Central America (one patient). These patients all tested negative for NS1. Sensitivity, specificity, as well as positive and negative predictive values of NS1 testing for the case and control patients are shown in Table I. DISCUSSION
Fig. 1. Sensitivity of NS1 testing as a function of days elapsed since symptom-onset in sera of returning travelers with symptomatic and serologically confirmed DENV infection.
This study aimed to measure the performance of a commercially available DENV NS1 diagnostic assay in travelers returning from a DENV-endemic region, to a country where other flaviviruses co-circulate. In Israel, WNV is the major endemic pathogenic flavivirus. Peak WNV transmission season in Israel corresponds to the peak DENV season in the preferred Israeli travel destinations, namely Thailand and India [Schwartz et al., 2008]. Distinguishing between J. Med. Virol. DOI 10.1002/jmv
2008
Fuchs et al.
TABLE I. Accuracy of NS1 Testing for Diagnosis of Dengue at Increasing Intervals From Symptom Onset in Patients With Acute Febrile Illness Interval since symptom onset to NS1 testing of serum (days) 0–6
0–9
0–12
All time intervals (0–21 days)
Positive/tested % Positive/tested % Positive/tested % Sensitivity in dengue patients Specificity in West Nile patients
17/23 0/8 0/8
Specificity in patients with febrile disease other than denguea Positive predictive value Negative predictive value
73.9 100
29/40 72.5 34/49 69.4 1/18 88.8 1/23 92.0 1 borderline 1 borderline 100 0/14 100 0/15 100
(94.4–100b) 79.3
(93.5–96.7b) 73.8
Positive/tested
%
34/58 1/26 1 borderline 0/15
58.6 92.3
94.4 70.5
100 94.4 61.9
a Leptospirosis-5 patients, malaria-4 patients, typhoid fever -3 patients, mononucleosis due to Epstein-Barr virus, cytomegalovirus, and acute influenza virus infection-one patient each. b Ranges with and without the WN patient with multiple sclerosis who received immunoglobulins 2 weeks before admission.
the two infections in a returning traveler has important implications both from clinical management and public health surveillance standpoints. In this study, the Panbio assay could detect NS1 in 72.5% (29/40) of the case patients’ sera when testing was performed within 9 days of fever-onset. This time interval encompasses the entire phase of clinical disease when the physician needs to make a clear diagnosis, and is the time interval recommended for testing by the test manufacturer. Compared to a 7-day average length of antigenemia reported from endemic settings, the test in the current study could detect antigen in sera drawn 12 days following onset of symptoms, although with sensitivity declining to about 50% (Fig. 1) [Alcon et al., 2002; Duong et al., 2011]. Several other studies have tested the accuracy of NS1 in returning travelers. One study in Polish servicemen did not find any NS1 positive cases, probably due to the fact that the patients were tested many days after the onset of symptoms [Goljan et al., 2010]. In Finnish returning travelers, sensitivity of the NS1 test using a different assay, the BIORAD Platelia Dengue NS1 AG EIA, ranged from 70% to 84% during the first week since onset of symptoms [Huhtamo et al., 2010]. However, while in the current study, the highest sensitivity—87.5%— was during days 1–3, the highest sensitivity—84%— was observed on days 6–7 in the study using the
BIORAD assay. In a study that tested sera of Japanese returning travelers using either the Panbio or the BIORAD kit, reported NS1 antigen positive rates ranged from 88% to 96% on days 1–5 [Moi et al., 2013]. In previous comparative studies from DENV-endemic areas (non-travelers population), the BIORAD assay showed higher sensitivity compared to Panbio during the first week of testing (Table II) [Dussart et al., 2008; McBride, 2009; Ramirez et al., 2009]. It should be mentioned that these comparative studies assessed sensitivity using the first generation Panbio NS1 assay, the Panbio Dengue Early Rapid, whereas this study used the second generation Panbio Dengue Early ELISA test (http://panbiodengue.com/page/ products). Two factors, beside technical test performance, may contribute to reduced sensitivity of NS1 detection observed in early sera in this study. The first is the potential for secondary DENV infection in “seasoned” returning travelers. Some experts consider sensitivity of NS1 to be reduced in secondary infection, although others have found no such association [Alcon et al., 2002; Duong et al., 2011; Moi et al., 2013]. Presumed secondary infections were seen by Huhtamo et al. [2010] who observed that certain sera from ill returning travelers had a negative NS1 test result but positive DENV RNA [Huhtamo et al., 2010]. In the
TABLE II. Comparison of Sensitivity and Specificity of PanBio and BIORAD NS1 Testing Kits Between Travelers and Non-Travelers in Selected Studies During the First Week of Symptoms Panbio Dengue Early ELISA Source a
Current study Huhtamo et al. [2010] Ramirez et al. [2009]b McBride [2009]c Chaiyaratana et al. [2009] a
Study population
Sensitivity %
Specificity %
Sensitivity %
Specificity %
T T E E E
72.5–87.5 N/A 50.4–70.5 51.3–73.0 N/A
88.8–100 N/A 94.4 N/A N/A
N/A 74.1–84.2 61.0–80.0 53.8–88.5 27.3–100
N/A 100 86.1 N/A N/A
Second generation test T, travelers; E, endemic or epidemic region; NA, not assessed. Specificity analysis included 36 sera from febrile patients negative for DEN. Included samples up to 8 days of illness.
b c
J. Med. Virol. DOI 10.1002/jmv
Bio-Rad Platelia assay
NS1 Dengue Diagnosis in Travelers
current study, 2 out of 6 cases of alleged secondary infections, defined by serum IgG >40 panbio units detected in sera sampled in the first week since symptom-onset, were negative for NS1 [Meltzer et al., 2012]. Theoretically, the cut-off of IgG >40 Panbio units, could give false positive results in an endemic country for WNV such as Israel, attributable to a previous infection with other flaviviruses. Using the IgM/IgG OD ratio
NS1 Antigen Testing for the Diagnosis of Dengue in Returned Israeli Travelers Inbal Fuchs,1* Hana Bin,2 Sara Schlezinger,2 and Eli Schwartz3 1
The Pediatric Infectious Disease Unit, Soroka University Medical Center, Ben-Gurion University of the Negev, Beer-Sheva, Israel 2 National Center for Zoonotic Viruses, Central Virology Laboratory, Ministry of Health, Public Health Services, Sheba Medical Center, Tel Hashomer, Israel 3 Center for Geographic Medicine and Tropical Diseases, Sheba Medical Center, Tel Hashomer, Israel
Antigen testing with NS1 provides rapid diagnosis of dengue in endemic regions during the febrile phase of illness before appearance of IgM in serum. This study aimed to determine the diagnostic accuracy of NS1 antigen testing in travelers presenting with febrile illness and serologically confirmed dengue infection, upon return from dengue endemic countries to Israel, a region endemic for West Nile virus (WNV). Cases were sera obtained from febrile returning travelers with positive dengue-IgM antibodies. Sera of non-travelers with confirmed WNV disease and sera of returning travelers with confirmed non-dengue febrile illnesses were used as controls. All sera were tested for NS1 antigen using the Panbio Dengue Early ELISA assay within 21 days of symptoms. Demographic data, travel destination, and interval between disease onset and testing were retrieved from patient files. Fifty-eight sera from 40 dengue-infected travelers, 26 sera from 26 WNV- infected patients, and 15 sera of returning travelers with non-dengue febrile illness were tested. Sensitivity of NS1 testing in dengue patients was 87% during the first 3 days of symptoms and declined to about 70% after 12 days. No cases tested positive for NS1 after day 12. Specificity was 92% for the entire testing period. The NS1 Panbio assay is sensitive for the detection of dengue viral infection in returning travelers during the febrile phase of illness, and is highly specific in a region where WNV co-circulates. J. Med. Virol. 86: 2005–2010, 2014. # 2014 Wiley Periodicals, Inc. KEY WORDS:
dengue; diagnosis; non-structural protein 1; traveler
INTRODUCTION Dengue fever (DF) has become one of the world’s most prevalent arthropod-borne viral illnesses. AccordC 2014 WILEY PERIODICALS, INC.
ing to the World Health Organization (WHO), 40% of the world’s population lives in a dengue virus (DENV-14) endemic area [WHO, 2009]. Every year, 50–100 million people are exposed to the virus, with approximately 25,000 fatalities [Guzman and Isturiz, 2010]. The disease is caused by the DENV-1-4, a single stranded ribonucleic acid (RNA) virus of the family Flaviviridae in the genus Flavivirus, the most important vectors being Aedes aegypti and Aedes albopictus. Dengue (DENV) is also one of the fastest spreading vector-borne diseases in the world due to population growth, uncontrolled urbanization, spread of mosquito vectors and rising international mobility [Kyle and Harris, 2008; WHO, 2009]. International travelers may acquire and spread DENV infection [WilderSmith and Schwartz, 2005]. Sero-survey studies have shown high sero-conversion rates after travel to endemic regions, demonstrating that DENV is one of the most common infections among travelers [Potasman et al., 1999; Cobelens et al., 2002]. Laboratory diagnosis methods for confirming DENV infection have traditionally involved detection of the virus, viral nucleic acid, or antibodies. These methods are not ideal for diagnosis of acute infection in returning travelers for several reasons. First, viral isolation, as well as polymerase chain reaction (PCR), are not routine methods in most medical facilities in non-endemic countries, as these are costly laborintensive procedures, requiring trained personnel for a low volume of tests. The short period of viremia in DENV infection presents another limitation for virus
The authors declare that they have no conflict of interest, financial or other, that might influence the contents of the manuscript. Correspondence to: Inbal Fuchs, MD, The Pediatric Infectious Disease Unit, Soroka University Medical Center, Beer Sheva, Israel. E-mail: [email protected] Accepted 12 December 2013 DOI 10.1002/jmv.23879 Published online 6 January 2014 in Wiley Online Library (wileyonlinelibrary.com).
2006
isolation or RNA detection in returning travelers. Culture and molecular diagnostics are most useful in detection of DENV in serum, circulating blood cells, and other tissues, when the tests are performed within 4–5 days of symptom onset. This window for virological diagnosis may be missed in returning travelers [Simmons et al., 2012]. Finally, two recent reports involving returning travelers with serologically confirmed DENV infection reported successful viral isolation in 69% of patients who presented during the first 5 days of disease, and in 86% in sera tested during the first 3 days of symptoms using viral culture and realtime PCR, respectively [Teichmann et al., 2004; Huhtamo et al., 2010]. Thus, molecular methods are not always confirmatory, even in early sera. As a result, diagnosis of DENV infection in travelers returning to non-endemic regions is usually based on clinical suspicion and confirmed with serological tests. However, clear-cut, bedside diagnosis of a febrile patient is limited by the delays inherent in seroconversion. In serum, IgM antibodies can usually be identified within 5–7 days of symptom-onset and IgG antibodies within 7–15 days following primary infections [Schwartz et al., 2000]. Serology-based diagnosis is also severely limited by cross-reactivity with other flavivirus infections or vaccination (e.g., yellow fever, Japanese encephalitis, tick-borne encephalitis [TBE]) [Schwartz et al., 2000; Lindegren et al., 2005; WHO, 2009]. This cross-reactivity can result in diagnostic uncertainties when a traveler returns from a DENV-endemic area and develops a febrile illness during transmission season of a co-circulating flavivirus endemic to the traveler’s home country. The development of a reliable test based on the detection of non-structural protein 1 (NS1) provided an important breakthrough in DENV diagnosis [Young et al., 2000; Alcon et al., 2002]. Antigen detection using NS1 has been used in DENV endemic areas in recent years as a less expensive and more easily accessible alternative to molecular testing. The secreted glycoprotein can be detected in high levels in the serum of DENV infected patients during the febrile phase of disease even when virus can no longer be isolated [Chaiyaratana et al., 2009; Duong et al., 2011]. Results of NS1 testing in endemic areas showed sensitivity in a range of 60–73% when sampling occurred within the first week following symptom onset [McBride, 2009; Ramirez et al., 2009]. Specificity data among returning travelers are scarce, and do not exist for travelers returning to countries with endemic flaviviruses. West Nile virus (WNV), the most important flavivirus in Israel, is endemic in parts of Africa, Europe, the Middle East and Asia. Outbreaks of disease also occur annually in North America [Sotelo et al., 2012]. Cross-reactivity between DENV-1-4 and WNV in serological tests has been described [Papa et al., 2011]. This study describes the performance of NS1 antigen testing in travelers returning to Israel, a country where WNV is endemic. J. Med. Virol. DOI 10.1002/jmv
Fuchs et al.
MATERIALS AND METHODS In Israel DENV serology is performed mainly at the Israeli National Center for Zoonotic Viruses (INCZV) at Sheba Medical Center, Tel Hashomer. The NS1 study was performed on sera of patients with acute DENV infection who presented during the years 2007–2011 (cases) and on sera of two control groups within 21 days of the onset of febrile illness. Intervals from symptom onset to testing were documented in the INCZV records for all serum samples obtained. This study was approved by the institutional ethical board of Sheba Medical Center with an exemption from informed consent. Case Definition Patients who presented with clinical illness compatible with DENV infection after visiting DENV endemic regions, with detectable DENV IgM in at least one serum sample, were defined as having acute DENV infection. Patients with positive DENV IgG antibodies alone were not included. Control Group 1 Definition Control group 1 consisted of serum samples obtained from Israeli patients with laboratory confirmed WNV infection, who had not traveled to countries endemic for DENV. A confirmed WNV case was defined in concordance with clinical and epidemiological data, after ruling out any other flavivirus infection or vaccination. Serological testing for WNV was carried out by the InBios West Nile DetectTM IgM Capture ELISA, and InBios West Nile DetectTM IgG ELISA (Seattle, WA); commercial kits simultaneously. A case was confirmed when one or more of the following occurred: (1) High levels of IgM in the cerebrospinal fluid, regardless of IgG result. (2) A fourfold increase in the kit’s units between paired samples collected at least 7 days apart. (3) Seroconversion of IgM and/or IgG from undetectable to a positive reading between paired samples collected at least 7 days apart. Data on co morbidities and medical treatment were available from patient files. Control Group 2 Definition Control group 2 consisted of serum samples obtained from returning travelers with a febrile illness other than DENV. All diagnoses were based on history of exposure, compatible clinical presentation, and confirmed by laboratory methods (malaria smear for malaria, blood culture for typhoid fever, microscopic agglutination test for leptospirosis, real-time PCR from a nasal swab for influenza, and IgM seroconversion for Epstein-Barr virus and cytomegalovirus). For case patients and control group 2 patients, individual files were retrieved from the Center of Travel and Tropical Medicine, at Sheba Medical Center. Data regarding for clinical presentation, travel destination, and interval between day of
NS1 Dengue Diagnosis in Travelers
2007
sampling and disease onset, were re-captured and confirmed. NS1 Antigen Testing and DENV Serology Testing Tests were performed according to manufacturer’s instructions using the second generation Panbio Dengue Early ELISA E-DENO2P (Inverness Medical Innovations Sinnamon Park, QLD, Australia). Dengue capture IgM, and dengue IgG indirect were performed according to manufacturer’s instructions (Panbio, Brisbane, Australia). RESULTS Fifty-eight sera from 40 patients who returned from DENV-endemic countries with positive DENVIgM serology during 2007–2011 were included in the study-group as cases. Eight patients (20%) had repeated serum testing which showed seroconversion. In the remainder of the returning travelers, a diagnosis of acute DENV infection was based on characteristic presentations of DF (clinical and laboratory findings) that appeared within a week of their return from a DENV-endemic region, and was further confirmed by positive DENV IgM. Twenty-seven case patients (68%) were males. Information regarding travel destination was available for 38 patients, of whom 35 acquired the infection in South-east Asia (Thailand, Laos Vietnam, and Cambodia) or India. Other destinations were: Costa Rica (1 patient) Bolivia (1 patient) and Papua New Guinea (1 patient). A positive NS1 result was observed in 34/58 (58%) of sera obtained from case patients during the 21-day interval. The test sensitivity was highest during the first 3 days, reaching 87.5%, declined to 70% during days 10–12, and became completely negative from day 13 onward (Fig. 1). Sera drawn from a patient on separate days from day 1 to day 8 of illness, showed a peak in the reading of NS1 panbio units on day 3 (Fig. 2).
Fig. 2. Consecutive testing of serum NS1 in a patient infected with DENV.
Two control groups were included in the study. In the first group, sera from 26 patients with confirmed WNV infection during 2009 and 2011 were tested for NS1. All sera were negative for NS1 except two samples (7.6%), which had a positive and a borderline NS1 test result. Results for the two samples in Panbio units were 19.45 and 9.19, respectively. In the assay used in this study, Panbio units are calculated as optical density (OD) (absorbance sample/OD cut off value) 10. Panbio units in the range of 9–11 are interpreted as equivocal. The WNV patient with a positive NS1 result had multiple sclerosis, for which she had received intravenous immunoglobulins 2 weeks prior to her admission. The WNV patient with a borderline NS1 result was a 65-year-old man who had not traveled outside of Israel. Repeat testing was not performed on the serum with a borderline result. The second control group included sera obtained from 15 returning travelers who were diagnosed with a febrile illness other than DENV during 2009–2012. The diagnoses were distributed as follows: leptospirosis (five patients), malaria (four patients), typhoid fever (three patients), mononucleosis due to EpsteinBarr virus, acute cytomegalovirus infection, and acute influenza virus infections (one patient each). Ten patients (66%) were male. Travel destinations were known for 12 of the 15 patients and were: South-east Asia/India (seven patients), Africa (four patients), Central America (one patient). These patients all tested negative for NS1. Sensitivity, specificity, as well as positive and negative predictive values of NS1 testing for the case and control patients are shown in Table I. DISCUSSION
Fig. 1. Sensitivity of NS1 testing as a function of days elapsed since symptom-onset in sera of returning travelers with symptomatic and serologically confirmed DENV infection.
This study aimed to measure the performance of a commercially available DENV NS1 diagnostic assay in travelers returning from a DENV-endemic region, to a country where other flaviviruses co-circulate. In Israel, WNV is the major endemic pathogenic flavivirus. Peak WNV transmission season in Israel corresponds to the peak DENV season in the preferred Israeli travel destinations, namely Thailand and India [Schwartz et al., 2008]. Distinguishing between J. Med. Virol. DOI 10.1002/jmv
2008
Fuchs et al.
TABLE I. Accuracy of NS1 Testing for Diagnosis of Dengue at Increasing Intervals From Symptom Onset in Patients With Acute Febrile Illness Interval since symptom onset to NS1 testing of serum (days) 0–6
0–9
0–12
All time intervals (0–21 days)
Positive/tested % Positive/tested % Positive/tested % Sensitivity in dengue patients Specificity in West Nile patients
17/23 0/8 0/8
Specificity in patients with febrile disease other than denguea Positive predictive value Negative predictive value
73.9 100
29/40 72.5 34/49 69.4 1/18 88.8 1/23 92.0 1 borderline 1 borderline 100 0/14 100 0/15 100
(94.4–100b) 79.3
(93.5–96.7b) 73.8
Positive/tested
%
34/58 1/26 1 borderline 0/15
58.6 92.3
94.4 70.5
100 94.4 61.9
a Leptospirosis-5 patients, malaria-4 patients, typhoid fever -3 patients, mononucleosis due to Epstein-Barr virus, cytomegalovirus, and acute influenza virus infection-one patient each. b Ranges with and without the WN patient with multiple sclerosis who received immunoglobulins 2 weeks before admission.
the two infections in a returning traveler has important implications both from clinical management and public health surveillance standpoints. In this study, the Panbio assay could detect NS1 in 72.5% (29/40) of the case patients’ sera when testing was performed within 9 days of fever-onset. This time interval encompasses the entire phase of clinical disease when the physician needs to make a clear diagnosis, and is the time interval recommended for testing by the test manufacturer. Compared to a 7-day average length of antigenemia reported from endemic settings, the test in the current study could detect antigen in sera drawn 12 days following onset of symptoms, although with sensitivity declining to about 50% (Fig. 1) [Alcon et al., 2002; Duong et al., 2011]. Several other studies have tested the accuracy of NS1 in returning travelers. One study in Polish servicemen did not find any NS1 positive cases, probably due to the fact that the patients were tested many days after the onset of symptoms [Goljan et al., 2010]. In Finnish returning travelers, sensitivity of the NS1 test using a different assay, the BIORAD Platelia Dengue NS1 AG EIA, ranged from 70% to 84% during the first week since onset of symptoms [Huhtamo et al., 2010]. However, while in the current study, the highest sensitivity—87.5%— was during days 1–3, the highest sensitivity—84%— was observed on days 6–7 in the study using the
BIORAD assay. In a study that tested sera of Japanese returning travelers using either the Panbio or the BIORAD kit, reported NS1 antigen positive rates ranged from 88% to 96% on days 1–5 [Moi et al., 2013]. In previous comparative studies from DENV-endemic areas (non-travelers population), the BIORAD assay showed higher sensitivity compared to Panbio during the first week of testing (Table II) [Dussart et al., 2008; McBride, 2009; Ramirez et al., 2009]. It should be mentioned that these comparative studies assessed sensitivity using the first generation Panbio NS1 assay, the Panbio Dengue Early Rapid, whereas this study used the second generation Panbio Dengue Early ELISA test (http://panbiodengue.com/page/ products). Two factors, beside technical test performance, may contribute to reduced sensitivity of NS1 detection observed in early sera in this study. The first is the potential for secondary DENV infection in “seasoned” returning travelers. Some experts consider sensitivity of NS1 to be reduced in secondary infection, although others have found no such association [Alcon et al., 2002; Duong et al., 2011; Moi et al., 2013]. Presumed secondary infections were seen by Huhtamo et al. [2010] who observed that certain sera from ill returning travelers had a negative NS1 test result but positive DENV RNA [Huhtamo et al., 2010]. In the
TABLE II. Comparison of Sensitivity and Specificity of PanBio and BIORAD NS1 Testing Kits Between Travelers and Non-Travelers in Selected Studies During the First Week of Symptoms Panbio Dengue Early ELISA Source a
Current study Huhtamo et al. [2010] Ramirez et al. [2009]b McBride [2009]c Chaiyaratana et al. [2009] a
Study population
Sensitivity %
Specificity %
Sensitivity %
Specificity %
T T E E E
72.5–87.5 N/A 50.4–70.5 51.3–73.0 N/A
88.8–100 N/A 94.4 N/A N/A
N/A 74.1–84.2 61.0–80.0 53.8–88.5 27.3–100
N/A 100 86.1 N/A N/A
Second generation test T, travelers; E, endemic or epidemic region; NA, not assessed. Specificity analysis included 36 sera from febrile patients negative for DEN. Included samples up to 8 days of illness.
b c
J. Med. Virol. DOI 10.1002/jmv
Bio-Rad Platelia assay
NS1 Dengue Diagnosis in Travelers
current study, 2 out of 6 cases of alleged secondary infections, defined by serum IgG >40 panbio units detected in sera sampled in the first week since symptom-onset, were negative for NS1 [Meltzer et al., 2012]. Theoretically, the cut-off of IgG >40 Panbio units, could give false positive results in an endemic country for WNV such as Israel, attributable to a previous infection with other flaviviruses. Using the IgM/IgG OD ratio
