We enrolled all patients over 15 years of age referred for febrile arthralgia at the ED between March 1^st 2006 and May 31^st 2006. A clinical examination, a blood sampling and a questionnaire were completed for all included subjects as the time of admission at the ED. This was the only time those patients were examined and interrogated for the purpose of this study.

The questionnaire was written after a 2-months observation period (January to February 2006, when the epidemic was at its peak, with an incidence of 45,000 cases per week). The questionnaire was created and validated after the medical file analysis of 1,030 consecutive patients referred to the ED with febrile arthralgia evocative of Chikungunya. The data collected included age, gender, medical history (co-morbidity, recent travels abroad), symptoms and date of their onset (1^st day), the need for hospitalization and biological parameters.

Oral consent was obtained from each patient or a first-degree relative, as the investigations were carried out under the standard care procedure for this poorly characterized disease, in accordance with the recommendations of the Committee for Clinical Research of the GHSR. In France, written consent is mandatory only if the medical treatments or the products used are not standard for the diagnosis, treatment, or monitoring (art. 88-II, law 2004–806, Journal Officiel, 08/11/2004; art. 31-I, law 2006–450, Journal Officiel, 04/19/2006). The information was given in French or in Creole with the help of a translator when necessary.

Chikungunya is known to evolve in successive phases: first, a viremic acute phase that lasts around 5 days, followed by a post-viremic acute phase defined by a negative CHIKV RT-PCR, presence of anti-CHIKV IgM antibodies but absence of IgG, that lasts about a week before the appearance of anti CHIKV IgG [5]. Please note that this chronology only applies to patients originating from area where CHIKV has not circulated previously, as it is the case in La Réunion. CHIKV-specific IgM and IgG antibodies were detected by capture enzyme-linked immuno-sorbent assay (ELISA) [6] and CHIKV-specific IgG antibodies by ELISA developed at the French National Reference Centre for Arbovirus (Institut Pasteur, Lyon, France) and automated with an EtiMax 3000 apparatus (DiaSorin®, Italy). The cut-off values for IgGs and IgMs were determined from a series of 30 negative sera collected in La Réunion before the epidemic. A one-step TaqMan® real time quantitative RT-PCR was performed in serum samples using the Light Cycler 2.0® (Roche Diagnostics). The assay sensitivity for serum specimen was 350 copies per mL [7]. The patients enrolled in the study were separated into two groups: Group A, patients with Chikungunya, Group B, patients without Chikungunya (negative RT-PCR and negative IgM anti CHIKV). In Group A the patients were in the acute infection phase (RT-PCR + or CHIKV Specific IgM + and IgG -). Within Group A, two subgroups were defined: Group A1, patients with viremia (RT-PCR+) and Group A2, patients without viremia. Whether the patients in the acute infection phase (Group A) had been hospitalized or not was used as a severity criterion.

Patients with positive anti-CHIKV IgG (i.e. that had been infected with CHIKV previously) were excluded from the study.

The study subjects were tested for other agents susceptible of causing fever and arthralgia. These include dengue and malaria (present in other islands of this area of the Indian Ocean) as well as leptospirosis, a prevalent infection in Reunion Island [8]. If patients had travelled during the previous months in areas endemic for malaria, a blood smear was performed. If patients had travelled during the previous ten days, anti-dengue IgM and dengue RT-PCR assays were performed. Leptospirosis diagnosis using microagglutination and PCR assays were performed in the presence of clinical signs evocative of leptospirosis, such as jaundice.

Groups were compared using Chi-square or Fisher's exact test, when appropriate. Clinical and biological continuous parameters were compared with the Mann-Whitney test. Averages and standard deviation were calculated. A P value <0.05 was considered statistically significant.

Among all patients admitted to the ED during the study, 2.21% were acutely infected by CHIKV and 1.86% were viremic. In acutely CHIKV-infected patients with fever and arthralgia (Group A), the positive predictive value of the presence of these two signs for CHIKV acute infection was 82.3%.

To our knowledge, this is the first prospective study on the clinical and biological aspects of acute Chikungunya infections in adults with differentiation between the viremic and non-viremic phases during an outbreak. Previously published clinical and biological data were obtained from retrospective analyses, usually in the absence of virological confirmation of acute infection. Several studies have used anti-CHIKV IgM as an acute phase criterion but they are still present in more than 50% of patients a year after the onset of the disease [5]. Simon et al. conducted a prospective observational clinical study during the early stages of Chikungunya infections (within 10 days of the disease onset) but did not differentiate between the viremic and non-viremic phases [11]. In addition, they included in their study data from a retrospective questionnaire completed by patients examined 10 days after the onset of symptoms. Recent retrospective studies reviewing cases that occurred in the Indian Ocean were published. Taubitz et al. described clinical and biological results for 20 travellers admitted in their department 2 to 73 days after the onset of symptoms [12]. Borgherini et al. published early biological and clinical features for 157 adult patients during the outbreak on Reunion Island [13]. Borgherini's study displays the biases inherent to retrospective surveys. In particular, the patient's selection and data collection processes did not include frequent clinical signs such as myalgia. This study also differs from previous works since its patients constitute an unbiased representative sample of the population enrolling at the ED over a continuous three-month period, and describes the clinical and biological features of the viremic and post-viremic phases of the acute CHIKV infection, as assessed by RT-PCR and IgM assays in 216 consecutive patients.

Moreover, most of these retrospective studies were carried out in countries where co-infections with arboviruses such as Dengue and Yellow-fever viruses are likely to occur [14]. We also searched for co-infections. No case of dengue fever was diagnosed in our study, and no cases of dengue fever or other arbovirosis was reported in 2006 in La Réunion, therefore reinforcing the specificity of our findings. Malaria and leptospirosis were not diagnosed in our patients infected with CHIKV. We found rare and benign hemorrhagic signs, in contrast to what previously reported in India and Southeast Asia —areas where Dengue is endemo-epidemic— where CHIKV was incriminated in haemorrhagic fever outbreaks [15]. Hemorrhagic signs in other studies among adults were probably attributable to co-infection, especially with dengue or yellow fever viruses, even if Asian CHIKV genotypes differ from the one responsible for the Indian Ocean outbreak, which is of East African origin [16]. During the outbreak in La Réunion, hemorrhagic signs were described in our prospective study of neonatal Chikungunya infection [17] and thrombocytopenia and haemostasis troubles were also common in the neonatal context.

Since 1953, the classical clinical features of acute Chikungunya associate the triad fever, arthralgia and inconstant skin rash [18]. We showed here that fever is concomitant of viremia, with the highest temperatures in the 2 days that follow the onset of the symptoms.

Arthralgia is key to the clinical diagnosis of acute CHIKV infection. However, forms with secondary arthralgia or without arthralgia have been described in Asian studies [19]. In a random sample of the population of La Réunion during the post-epidemic phase, 9.7% of proven Chikungunya cases (i.e. with a positive serology) reported no arthralgia, including the 5% asymptomatic forms [3]. In our study, the presence of arthralgia was mandatory for inclusion, and asymptomatic forms were obviously not referred to the ED. CHIKV-associated arthralgia have not been characterized in detail so far [20], [21]. While in most cases, bilateral and symmetrical polyarthralgia were reported (although sometimes more intense on one side), we observed two cases of monoarticular arthralgia of the ankle. We recorded rare locations of arthralgia, such as of the temporo-mandibular joint, and forms that had never been reported in previous descriptions such as chondrocostal arthralgia, affecting up to 20% of the patients, and hip arthralgia (17%). CHIKV also appeared to revive pain along former fractures lines and ligament injuries, and clinical examination revealed entesopathies and talalgia, which had never been described previously.

The frequency of skin rash ranged from 20% to 85% [18], [22]. In our study, about half of the patients presented a skin rash, generally arising during the first three days of infection. Our study matches former studies in the location and aspect of the skin lesions but differs by the existence of vesicular and bullous forms as well as erysipela-like lesions of the lower limbs. By mean of RT-PCR, we could document the specificity of these skin manifestations by amplifying the viral genome within this vesicular fluid (our unpublished data).

Our study illustrates the variety of the other clinical manifestations that can be associated with CHIKV infection. Yet, these symptoms are non-specific as they can be present in a number of other viral infections. Some of these clinical symptoms may also correspond to decompensation of underlying conditions and to iatrogenic effects (e.g. paracetamol and hepatitis, non-steroidal anti-inflammatory and kidney failure).

Severe clinical forms with neurological [23], [24] and cardiac involvements [25], [26] have also been reported in the literature. In 1973, Obeyesekere [25] described the first cases of acute pericarditis and myocardiopathy associated with CHIKV, based on serology. Thus, CHIKV may, as other viruses, trigger pericarditis and myocarditis. Lemant et al. [27], in a retrospective survey of CHIKV-infected patients admitted to intensive care, reported a case of fatal fulminant myocarditis for which analysis of post-mortem myocardial tissue samples pathological examination demonstrated the presence of cytoplasmic viral inclusions in myocytes. Cardiac complications and myocarditis have also been described in infected newborns on the Island of La Réunion [17], [28].

Neurological symptoms (vertigo or confusion) observed in our study may be attributable to fever and dehydration. However, a neurotropism of CHIKV has several times been suspected [14], [29], [30]. Mazaud [30] reported a case of acute encephalitic syndrome, and Deller [29] described the onset of vertigo and a vestibular syndrome. In India, Carey [14] observed important neurological after-effects. In La Réunion, severe cases of meningo-encephalopathy among adults [27] and newborns [28], and acute polyradiculoneuritis were also attributed to CHIKV [24]. Consistent with these clinical observations, we have shown, in an animal model for Chikungunya, that CHIKV may disseminate to the central nervous system, and infect the meninges and the ependymal tissue [31].

Mazaud [30] observed biological similarities between dengue and chikungunya. Indeed, it was reported that these two arbovirosis are associated with brief leukoneutropenia (3 to 4 days) and thrombocytopenia [30]. According to this author, leukoneutropenia and thrombocytopenia have great orientation value in favour of these two arboviroses. In other studies, no leukopenia was reported [19]. In contrast to these retrospective studies in which the differential diagnosis between dengue and chikungunya could not be precisely assessed, we found in our series that 38% of patients were leukopenic and that in Group A1, lymphopenia dominated and was very closely associated with viremia; for 85% of the patients lymphocyte count was lower or equal to 1,000/mm³ and for 50% of the patients the level was lower than 500/mm³. Thrombocytopenia was found in 42% of the patients, but to moderate levels, between 100,000 and 150,000/mm³.

Cases of severe acute hepatitis occurring during the CHIKV infection have already been described in La Réunion. While it was suggested that this virus may target the liver –a finding confirmed in our animal model in the first hours of infection–, acute hepatitis seems to be mainly promoted by chronic ethylism, denutrition and paracetamol toxicity [32]. The serum ASAT level was increased for 32% of the patients as opposed to only 7% for the ALAT. These results suggest that the increase in the serum level of ASAT was partially related to muscle injury, with 60% of the patients with increased ASAT also presented a rhabdomyolysis.

When using the two cardinal clinical signs in acute phase, fever and arthralgia, for the diagnosis, we found a specificity of 99,6% and a positive predictive value of 84.6%. Leucopenia and lymphopenia could be used to aid in the diagnosis of CHIKV infection in acute phase, but not for formal diagnosis. The positive predictive values for these parameters, found in our work, are inappropriate. Indeed, the high number of CHIKV-infected patients compared to controls tends to artificially raise the positive predictive values and decrease the negative predictive value. So, leucopenia/lymphopenia values are suggested as only one parameter that could be a part of a diagnosis algorithm. Other mosquito-transmitted viruses with overlapping geographic distributions, specifically Dengue Fever, can have similar hematologic abnormalities. However because of the risk of transmission of the Chikungunya fever in European countries free of Dengue Fever, these hematologic abnormalities are of interest and value.

Factors influencing disease severity are dominated by age and comorbidities. However, these two criteria are strongly linked. Among viremic patients, age under 65 was an excellent predictor of non-severity (severity negative predictive value 85.6%), with only functional impotence due to lower limbs pain and inability to walk leading to hospitalization in this group.

Little is known about the pathophysiology of CHIKV infection, in contrast to better characterized alphaviruses such as Sindbis and Ross River viruses. Sourisseau et al. [33] have reported that CHIKV replicates in various human cultured cells, such as epithelial and endothelial cells, primary fibroblasts and, to a lesser extent, macrophages. CHIKV replication induces a significant cytopathic effect, and may also lead to apoptosis. CHIKV replication is significantly inhibited by type I IFNs, suggesting a prominent role of the innate immune system in the control of the infection and the rapid decline of viremia before the appearance of neutralizing antibodies during the acute phase of infection. It has also been reported that CHIKV may also infect muscle satellite cells in humans [34]. In a mouse model of chikungunya, we have shown that fibroblasts of tissues in which seats symptoms in humans (muscles, joints, skin), are the target cells of CHIKV [31], these findings matching viral tropism in human as assessed by CHIKV immunolabeling of infected human tissues. The infection of muscles and joints, which contain nerve endings may account for the nociceptive pain observed in the course of human acute infection. The functional neurological manifestations complicating the course of acute CHIKV infection may reflect CHIKV dissemination to the central nervous system, consistent with the positive CHIKV RT-PCR in the cerebrospinal fluid of these patients and the observation that CHIKV infect the choroid plexuses, meninges and the ependyma in the mouse animal model [31].