- Top of page
- Materials and methods
Objectives The aim of this study was to assess the applicability and benefits of the new WHO dengue fever guidelines in clinical practice, for returning travellers.
Methods We compared differences in specificity and sensitivity between the old and the new guidelines for diagnosing dengue and assessed the usefulness in predicting the clinical course of the disease. Also, we investigated whether hypertension, diabetes or allergies, ethnicity or high age influenced the course of disease.
Results In our setting, the old classification, compared with the new, had a marginally higher sensitivity for diagnosing dengue. The new classification had a slightly higher specificity and was less rigid. Patients with dengue who had warning signs as postulated in the new classification were admitted more often than those who had no warning signs (RR, 8.09 [1.80–35.48]). We did not find ethnicity, age, hypertension, diabetes mellitus or allergies to be predictive of the clinical course.
Conclusions In our cohort of returned travellers, the new classification system did not differ in sensitivity and specificity from the old system to a clinically relevant degree. The guidelines did not improve identification of severe disease.
Objectifs: Evaluer l’applicabilité et les avantages des nouvelles directives de l’OMS pour la fièvre dengue dans la pratique clinique, pour les voyageurs au retour.
Méthodes: Nous avons comparé les différences dans la spécificité et la sensibilité entre les anciennes et les nouvelles directives pour le diagnostic de la dengue, et avons évalué leur utilité dans la prédiction de l’évolution clinique de la maladie. Nous avons aussi examiné si l’hypertension, le diabète ou les allergies, l’origine ethnique ou l’âge élevé influençait le cours de la maladie.
Résultats: Dans notre contexte, l’ancienne classification, par rapport à la nouvelle, avait une sensibilité légèrement supérieure pour le diagnostic de la dengue. La nouvelle classification avait une spécificité légèrement plus élevée et était moins rigide. Les patients atteints de la dengue qui avaient des signes d’alerte, comme énoncés dans la nouvelle classification ont été admis plus souvent que ceux qui n’avaient pas de signes d’alerte (RR: 8,09 [1,80 à 35,48]). Nous n’avons pas trouvé l’origine ethnique, l’âge, l’hypertension, le diabète sucré ou les allergies comme facteurs prédictifs de l’évolution clinique.
Conclusions: Dans notre cohorte de voyageurs au retour, le nouveau système de classification ne diffère pas de l’ancien système en termes de sensibilité et de spécificitéà un degré cliniquement significatif. Les directives n’ont pas permis d’améliorer l’identification de maladie grave.
Objetivos: Evaluar la aplicabilidad y los beneficios de las nuevas guías clínicas de la OMS para la fiebre del dengue en viajeros.
Métodos: Hemos comparado las diferencias en especificidad y sensibilidad entre las antigua guía para el diagnóstico del dengue y la guía nueva, y hemos evaluado su utilidad para predecir el desarrollo clínico de la enfermedad. También hemos investigado si la hipertensión, la diabetes o las alergias, la etnicidad o la edad influencian el curso de la enfermedad.
Resultados: En nuestro emplazamiento, la clasificación antigua, comparada con la nueva, tenía una sensibilidad marginalmente más alta para el diagnóstico del dengue. La nueva clasificación tenía una especificidad ligeramente más alta y era menos rígida. Los pacientes con dengue que tenían signos de alarma tal y como se postula en la nueva clasificación, eran ingresados más a menudo que aquellos que no los tenían (RR 8.09 [1.80-35.48]). No encontramos que el grupo étnico, la edad, la hipertensión, la diabetes mellitus o las alergias fuesen vaticinadores del desarrollo clínico.
Conclusiones: En una cohorte de viajeros que regresan, el nuevo sistema de clasificación no difería en sensibilidad o especificidad del sistema antiguo a un nivel clínicamente relevante. Las guías no mejoraron la identificación de la enfermedad severa.
- Top of page
- Materials and methods
Dengue is one of the major challenges for public health in tropical and subtropical regions, with around 2.5 billion people at risk and an estimated 50–100 million cases per year. Dengue fever is caused by the dengue virus, a mosquito-borne flavivirus, which is transferred by mosquitoes of the Aedes genus (Halstead 1988; Kouri et al. 1989). Until World War II, outbreaks of dengue were reported on almost all continents every 10–30 years, mainly taking a benign clinical course with high fever and bone and back pain (Mairuhu et al. 2004).
After 1945, a more severe form of dengue spread dramatically across the world, and cases complicated by haemorrhage and shock were increasingly documented in Southeast Asia (Mairuhu et al. 2004).
Many dengue-endemic countries are popular tourist destinations (Jelinek 2000). The disease has been diagnosed in an increasing proportion of febrile travellers, ranging from 2% in the early 1990s to 16% more recently (MacLean et al. 1994, Doherty et al. 1995; O’Brien et al. 2001). The incidence in returning travellers is probably underestimated because in many countries it is not a notifiable condition and because of the non-specific early clinical features of the disease (Gibbons & Vaughn 2002).
In 1974, the WHO Technical Advisory Committee formulated a dengue classification system in Manila, the Philippines, which identified dengue fever (DF), dengue haemorrhagic fever (DHF) and dengue shock syndrome (DSS) (Figure 1) (WHO 1975). This system was meant to classify the severity of the disease in endemic areas for triage and medical resource allocation. Also, it was used as an epidemiological tool. The classification was mainly based on clinical data collected in the 1950s and 1960s at the Children’s Hospital, Bangkok, Thailand.
Figure 1. Dengue classification according to the World Health Organization guidelines, issued in 1975 and 1997. Dengue is classified as undifferentiated fever, dengue fever (DF) or dengue haemorrhagic fever (DHF). In addition to fever and at least two clinical findings, diagnosis of DF requires epidemiological or laboratory evidence supporting a dengue virus infection. To meet a case definition of DHF, all the four following criteria are required: (1) fever, (2) haemorrhagic manifestations, (3) thrombocytopenia (platelet count, ≤100000 platelets/mm3) and (4) evidence of plasma leakage. Diagnosis of DHF does not require laboratory support.
Download figure to PowerPoint
Recent research by Balmaseda et al. (2005) and Bandyopadhyay et al. (2006) suggests that the 1975 classification may not be universally applicable for clinical management, because clinical characteristics of dengue appear to differ in various geographical regions and among different age groups. Severe clinical features also occurred in patients who did not meet the criteria for DHF, for example patients with organ failure, but without evidence of plasma leakage (Phuong et al. 2004; Srikiatkhachorn et al. 2010). In response, Alexander et al. (2011) developed a new dengue classification system, which is incorporated in the revised WHO guidelines of 2009 (WHO 2009). In these guidelines, a classification according to severity is outlined: dengue fever with and without warning signs, and severe dengue. From clinical data collected in numerous endemic areas worldwide, several clinical symptoms (lethargy, abdominal tenderness, mucosal bleeding, liver enlargement, persistent vomiting, haemoconcentration increase and platelet decrease) were identified as possible predictors of a severe clinical course and defined as warning signs (Figure 2).
Figure 2. Dengue classification according to the World Health Organization guidelines, issued in 2009. Dengue is classified as dengue with or without warning signs and severe dengue. Diagnosis of dengue requires the presence of fever and at least two clinical findings or any warning signs. Epidemiological or laboratory evidence is required to make the diagnosis. Severe dengue is defined as dengue with any of the following: (1) severe plasma leakage leading to shock or respiratory distress, (2) severe haemorrhage or (3) any organ failure.
Download figure to PowerPoint
Evaluations carried out in dengue-endemic areas have provided promising results, but have also evoked criticism. First, the new system would be too unspecific in defining severe dengue that needs close monitoring, therefore imposing a heavy workload on the health services in resource-poor countries; second, there would be too much focus on unusual dengue manifestations (Barniol et al. 2011; Srikiatkhachorn et al. 2011).
To date, the new classification had not been widely tested in returning travellers. In this population, the situation is different from permanent inhabitants of endemic areas, where only severe dengue is notified (Phuong et al. 2006). By contrast, travellers are inclined to seek medical attention early in the course of their disease. As a consequence, data from travellers better reflect the distribution over the different categories of severity. DHF and DSS are not reported frequently in this population. Of 483 dengue cases in the European Network on Imported Diseases, 2.7% were reported as DHF, with a 4.3 times higher risk among immigrants and travellers visiting friends and family, compared with tourists (Jelinek et al. 2002). DSS and death from dengue are uncommon. Two deaths were reported in a series of more than 250 cases of dengue imported into the United States between 1993 and 2000 (MMWR reported weekly Imported dengue US 1993–2000).
We compared the old and the proposed new dengue classification system in returning travellers for whom dengue serology was requested. We used clinical and laboratory characteristics of these patients to assess whether the warning signs as postulated by Alexander et al. (2011) apply to returning travellers, whether sensitivity and specificity of both diagnosing dengue classification systems differed, and whether clinical data currently gathered at the Academic Medical Center (AMC) were sufficient to categorise patients according to the proposed guidelines.
In the past, factors such as age, ethnicity and several chronic medical conditions (diabetes, hypertension and allergies) were described as potentially influencing severity of the disease, owing to alterations in the inflammatory response of the endothelium (Kouri et al. 1987; Cunha et al. 1999; Halstead et al. 2001; Lee et al. 2008; Figueiredo et al. 2010). Therefore, we analysed whether these factors influenced severity of the course of the disease.
- Top of page
- Materials and methods
In total, 1124 tests were obtained. After exclusion of those treated outside the AMC (200), double tests (285), missing files (31), onset of symptoms >6 months prior to consultation (14) and not having travelled (13), 581 patients remained.
Of these patients, 132 (22.7%) were diagnosed with dengue, 277 (47.7%) were diagnosed with no dengue, and 172 (29.6%) had indeterminate results (Table 1). In 17 indeterminate cases, serology and antigen tests were negative, but the clinical and epidemiological presentation was highly suggestive for dengue. In 31 indeterminate cases, serology was positive, but there was no clinical and epidemiological presentation typical of dengue. In 124 ‘indeterminate’ cases, serology was indeterminate without clinical suspicion of dengue.
Table 1. Characteristics of study population
| ||Dengue||No Dengue||Indeterminate||Total|
|Male (%)|| 63 (47.7)||132 (47.6)|| 72 (41.9)||267 (46.0)|
|European ethnicity (%)|| 91 (68.9)||173 (62.5)||117 (68.0)||381 (65.6)|
| Median|| 36|| 36|| 33|| 36|
| Interquartile range|| 28.0–51.3|| 27.0–47.0|| 26.0–48.0|| 27.0–47.0|
|Hospital admissions (%)|| 10 (7.6)|| 36 (13.0)|| 25 (14.5)|| 71 (12.2)|
|Probable dengue 1975 scheme (%)||111 (84.1)||157 (56.7)|| 98 (57.0)||366 (63.0)|
|Sensitivity old scheme: 84.1%|
|Specificity old scheme: 43.3%|
|Probable dengue 2009 scheme (%)||107 (81.1)||149 (53.8)|| 90 (52.3)||346 (59.6)|
|Sensitivity new scheme: 81.1%|
|Specificity new scheme: 46.2%|
|Dengue with warning signs (%)|| 44 (33.3)|| 59 (21.3)|| 35 (20.3)||138 (23.8)|
In 449/581 (77.3%) cases, one serologic sample was available; for 132 (22.7%) cases, samples in the initial and convalescent phase were collected. In 61/132 (46.2%) confirmed cases, serologic diagnosis was based on positive IgM in one sample.
Clinical symptoms and travel destinations
Symptoms and travel destinations of the 132 confirmed dengue patients are recorded in Tables 2 and 3. A total of 111 of 132 patients fulfilled the criteria of probable dengue, and 120 of 277 patients not diagnosed with dengue were negative according to the old system (sensitivity, 84.1%; specificity, 43.3%). One hundred and seven of 132 patients fulfilled the criteria of probable dengue in the revised system, and 128 of 277 patients not diagnosed with dengue were negative according to the new system (sensitivity 81.1%, specificity 46.2%). Haemorrhagic manifestations were reported in 52 patients (8.9%) of the total population and in 22 (16.7%) of the dengue population. There were no patients who fulfilled the criteria of DHF. Dengue with warning signs was recorded in 44 patients, eight of whom were hospitalised.
Table 2. Clinical findings of 132 patients diagnosed with dengue on serological and clinicoepidemiological grounds
|1997 Classification||Positive (%)||Negative (%)||Missing (%)|
|Fever*||123 (93.2)||4 (3.0)||5 (3.8)|
|Headache||98 (74.2)||8 (6.1)||26 (19.7)|
|Retro-orbital pain||44 (33.3)||15 (11.4)||73 (55.3)|
|Rash||88 (66.7)||22 (16.7)||22 (16.7)|
|Myalgia||80 (60.6)||13 (9.8)||39 (29.5)|
|Arthralgia||52 (39.4)||19 (14.4)||61 (46.2)|
|Petechiae||16 (12.1)||65 (49.2)||51 (38.6)|
|Leucopenia||22 (16.6)||62 (47.0)||48 (36.4)|
|Haemorrhagic manifestations*||6 (4.5)||98 (74.2)||28 (21.2)|
|Thrombocytopenia ≤100 000 platelets/mm3*||57 (43.2)||63 (47.7)||12 (9.1)|
|Rise in haematocrit*||16 (12.1)||72 (54.5)||44 (33.3)|
|All 4 recorded in chart: 19 (14.4%)|
| Fever||123 (93.2)||4 (3.0)||5 (3.8)|
| Nausea, vomiting||42 (31.8)||32 (24.2)||58 (43.9)|
| Rash||88 (66.7)||22 (16.7)||22 (16.7)|
| Aches and pains||122 (92.4)||10 (7.6)||0 (0)|
| Tourniquet test positive||0 (0)||0 (0)|| 132 (100)|
| Leucopenia||57 (43.2)||63 (47.7)||12 (9.1)|
| Tender abdomen†||25 (18.9)||96 (72.7)||11 (8.3)|
| Persistent vomiting†||0 (0)||74 (56.0)||58 (43.9)|
| Clinical fluid collection†||0 (0)||94 (71.2)||38 (28.8)|
| Mucosal bleeding†||10 (7.6)||88 (66.7)||34 (25.8)|
| Fatigue/lethargy†||28 (21.2)/9 (6.8)||49 (37.1)||46 (34.8)|
| Enlarged Liver†||6 (4.5)||98 (74.2)||28 (21.3)|
| Rise in Haematocrit†||1 (0.8)||66 (50.0)||65 (49.2)|
|At least 1 warning sign recorded in chart: 100%|
Table 3. Travel destinations of 132 travellers diagnosed with dengue
| ||Destinations N = 132 (%)|
|Central America|| 7 (5.3)|
|South America||31 (23.5)|
|South Central Asia|| 8 (6.1)|
|Southeast Asia||48 (36.4)|
|Sub-Saharan Africa|| 4 (3.0)|
|Other*|| 3 (2.3)|
In the total population, 71/581 (12.2%) were admitted versus 10/132 (7.6%) of those diagnosed with dengue. Patients with warning signs were more often admitted than those without warning signs in the total population with suspected dengue (RR, 1.43 [0.90–2.28]; P = 0.126). In those diagnosed with dengue, differences were significant (RR, 8.09 [1.80-35.48]; P = 0.001). Reasons for hospital admission in this group were high fever and severe illness (6), decreased fluid intake (2), suspicion of abdominal typhus (1) and hypotension (1). All admitted patients made a full recovery.
Risk factors, warning signs and hospital admission
Twenty-eight patients with dengue (21.2%) had a medical history, of whom 10 reported hypertension, three diabetes mellitus and five allergies. Nine of 132 patients with dengue (6.8%) were aged over 60 and 91 patients (68.9%) were of European origin. With regard to hospital admission or warning signs, no significant differences were found in those with and without a medical history in general, hypertension, allergies or diabetes mellitus (data not shown). Neither ethnicity nor age >60 were risk factors for hospital admission (data not shown).
Missing reported symptoms
Missing reported symptoms for defining ‘probable dengue’ were comparable in both classification systems (Table 2). The tourniquet test was never performed. To diagnose DHF, four separate symptoms are mandatory. In our data, only 14.4% of cases exhibited all four symptoms. Dengue with warning signs is diagnosed if one warning sign is present. In all cases, at least one warning sign was recorded.
- Top of page
- Materials and methods
This study was carried out to assess whether the new classification system is applicable to returning travellers. The assessment was carried out by comparing the current and the proposed WHO classification scheme in their potential to identify dengue and severe cases of dengue. Also, we assessed specific risk factors for travellers heading to endemic areas.
We found that sensitivity for diagnosing probable dengue was slightly higher in the old system than in the new system, and specificity was higher in the new system than the old. The small differences we found are probably not clinically relevant.
In those diagnosed with dengue, we found that significantly more patients with warning signs were admitted than those without. There were no cases of dengue with severe haemorrhagic manifestations or shock. If we assume hospital admission to be a clear indicator of potential severe disease, we may conclude that in all returning travellers in whom there is a suspicion of dengue, the presence of warning signs should be taken into account. However, only a small percentage of those with warning signs needed hospital admission, and warning signs themselves did not seem to help in identifying those who needed intervention.
Several criteria for diagnosing DHF in the old classification system are not used in routine clinical practice in many institutions (as well as ours) as their diagnostic value is not considered as high as other clinical signs: for example, the tourniquet test, which is scarcely performed, and hypoproteinaemia, which is rarely quantified. The tourniquet test is also used in the new classification system; however, here it is one of many warning signs and not one of the required symptoms to diagnose severe disease.
We did not find medical history, age or ethnicity to be a predictor of either clinical severity or hospital admissions, contradictory to findings of previous studies. Possibly, our small group size plays a role in this outcome. Patients diagnosed with dengue had often travelled to Southeast Asia, the Caribbean and South America.
Limitations of the study
Theoretically, patients with complications could have been missed in our retrospective cohort. Some studies recommend daily assessment (Wilder-Smith & Schwarz 2005). However, the chance that we missed admissions in other hospitals is small, as patients are routinely told to contact us in case of physical deterioration.
We did not carry out repeated measurements of haematocrit, and measurements of plasma leakage (in the form of ultrasound and chest X-rays), if it was not clinically indicated. For this reason, possibly, a certain degree of plasma leakage could have been missed in our patients. Previous studies have noted plasma leakage in a significant amount of returning travellers (Meltzer et al. 2012). This probably will not have much impact on the number of cases classified as DHF, but possibly would have increased the rate of those with warning signs.
Our study population only included patients for whom serology was requested. However, sensitivity and specificity may be different if a total population of travellers returning with fever would be included. We reason, however, that the sensitivity and specificity of recognizing probable dengue in this group are relevant, as these are the patients in whom dengue is featuring in the differential diagnosis and therefore those where the dengue classification system might be applied.
As a result of the high rate of indeterminate cases, our results may be biased. Seventeen patients were diagnosed with probable dengue on clinical and epidemiological grounds, without serologic confirmation and therefore were in the group of ‘indeterminate’ dengue. Perhaps these patients had not yet seroconverted, and no sample was taken in the convalescent phase. The rapid immunochromatographic assay has been extensively evaluated in India, Southeast Asia and the Caribbean, showing sensitivity from 45% to 100% and specificity from 57% to 100% (Blacksell et al. 2006). Sensitivity was highest when samples were collected later in the acute phase of infection. Another reason could be that the antigen tests were false negative in these patients or that a dengue-like illness, such as chikungunya, caused a clinical picture with negative serology in this group. A total of 155 patients had positive or indefinite serology without a clinicoepidemiological picture of dengue. This might be due to cross-reactivity with other flaviviruses such as yellow fever or Japanese encephalitis (Teles et al. 2005).