To evaluate whether chloroquine (CQ) is more effective than meloxicam for treating early musculoskeletal pain and arthritis following acute chikungunya (CHIK) virus infection.
To evaluate whether chloroquine (CQ) is more effective than meloxicam for treating early musculoskeletal pain and arthritis following acute chikungunya (CHIK) virus infection.
During the 2006 CHIK epidemic, 509 rural community cases of acute CHIK virus infection were identified in the district of Sholapur in India. Seventy consenting adult patients (seropositive for IgM/IgG anti-CHIK antibody) with early persistent musculoskeletal pain and arthritis were randomized into a 24-week, 2-arm, parallel efficacy trial of CQ (250 mg/day) and meloxicam (7.5 mg/day). Assessors completed a rheumatology evaluation in a blinded manner and collected blood samples in the patients' homes, as per protocol. Laboratory parameters included serum cytokine assay (interleukin-6 [IL-6], interferon-γ [IFNγ], tumor necrosis factor α, CXCL10/IFNγ-inducible protein 10, and IL-13). Twenty-two patients who failed to meet the eligibility criteria (low pain cohort) were also followed up with similar evaluations. An intent-to-treat analysis was completed. At baseline, the 2 groups (38 patients randomized to receive CQ and 32 patients randomized to receive meloxicam) were well matched.
There were no significant efficacy differences between the meloxicam group and the CQ group (mean changes in the visual analog scale score for pain −3.9 and −4.2, respectively). Patients improved significantly. Cytokine levels remained several-fold increased, were disproportionate to the clinical response, and were not different from those in the low pain cohort. Seven patients withdrew. Adverse events were mild and infrequent.
This exploratory community intervention trial failed to identify an advantage of CQ over meloxicam to treat early musculoskeletal pain and arthritis following acute CHIK virus infection, but therapeutic efficacy of CQ was not ruled out. The inflammatory cytokine response was intense and was not consistent with clinical status.
The 2006 chikungunya (CHIK) virus epidemic originated in East Africa and rapidly spread to the Indian subcontinent ([1-3]). Unlike previous epidemics, thousands of new cases and patients with persistent musculoskeletal pain and arthritis continue to be reported. We previously observed a wide spectrum of post-CHIK virus musculoskeletal pain and arthritis, but the etiology remains unknown (). Oral chloroquine (CQ) has been used extensively to treat acute and chronic musculoskeletal pain and arthritis following CHIK virus infection; however, clinical data remain sparse ().
Acute CHIK virus infection is a self-limiting, excruciatingly painful arboviral illness of short duration. In our experience, acute symptoms subsided within 10 days, and approximately two-thirds of patients recovered within 3 weeks (). Against this background, we carried out a controlled evaluation of CQ to treat patients with early persistent musculoskeletal pain and arthritis following CHIK virus infection. Selected cytokine assays were carried out to study the inflammatory nature of musculoskeletal pain and arthritis and the response of these symptoms to CQ.
The study was carried out in the village of Bavi (district of Sholapur in south central India) during the epidemic. The protocol was approved by the ethics committee of the Center for Rheumatic Diseases, Pune. All eligible patients who volunteered provided informed consent (in the local language) prior to enrollment.
This was an investigator-initiated 24-week, noncommercial, randomized, controlled, 2-arm, assessor-blinded, parallel-efficacy active comparator drug trial. The trial was essentially exploratory in nature and was not statistically designed. Patients were examined in their homes at 4-week intervals, according to the protocol. The trial was carried out by a rheumatology team from the Center for Rheumatic Diseases, Pune.
Bavi is a remote village (∼2,000 population and 200 km southeast of the current study center) with a predominantly farming community. The village is situated in a known drought-prone area with scarce water supply and electricity. Hygiene and sanitation are deplorable. The village does not have a doctor or any medical facility. An educated (school level) volunteer who resided in the village was trained to be a healthcare provider for this study.
Among 1,192 respondents to a previous community survey in Bavi (), we identified 509 cases of symptomatic CHIK virus infection. Seventy patients fulfilled the eligibility criteria and were enrolled in the current study (Figure 1). The principal inclusion criteria were persistent musculoskeletal pain and arthritis of more than 6 weeks' duration following the onset of CHIK virus infection, seropositivity for IgG and/or IgM anti-CHIK antibody, and a maximum visual analog scale (VAS) score for pain of ≥4 cm (0–10-cm scale). The principal exclusion criteria were a history of rheumatoid arthritis (RA) or any inflammatory arthritis disorder, drug hypersensitivity, recent severe illness (other than CHIK virus infection), unwillingness to participate, and any other condition (based on the investigator's discretion).
Twenty-two consenting patients with low-intensity pain who failed to meet the eligibility criteria (the “low pain cohort”) were followed up with a similar clinical evaluation schedule.
The primary efficacy criterion was the maximum pain severity in the musculoskeletal tissue and joints during the preceding 24 hours, recorded on a 100-cm VAS. We used several American College of Rheumatology core set criteria (). Other efficacy measures included a tender joint count in 68 diarthrodial joints; a swollen joint count in 66 joints (excluding hips); physician's global assessment of disease activity and patient's global assessment of disease activity on a scale of 1 to 5, where 1 = asymptomatic and 5 = very severe; duration of morning stiffness (in minutes) on the day before the visit; and functional assessment using a Center for Rheumatic Diseases–Pune version of a modified Stanford Health Assessment Questionnaire suitable and validated for Indian use (). Oral acetaminophen intake was carefully recorded.
The erythrocyte sedimentation rate (ESR) was measured using the Westergren method, and peripheral blood smears were performed in the village. Blood chemistries (including serum creatine kinase), urinalysis, and routine total and differential cell counts including platelets were performed; autoantibodies customarily assessed in rheumatic diseases (rheumatoid factor [RF], antinuclear antibody [ANA], and anti–cyclic citrullinated peptide [anti-CCP]) and serum cytokines were measured according to the protocol; and a latex agglutination test for C-reactive protein (CRP; cutoff value 5 mg/dl) was performed. Selected cytokines (interleukin-6 [IL-6], interferon-γ [IFNγ], tumor necrosis factor α [TNFα], CXCL10/IFNγ-inducible protein 10 [IP-10], and IL-13) were assayed using a DuoSet ELISA Development Kit (R&D Systems) at week 0, week 12, and week 24. Dengue fever and malaria were excluded using serum IgM/IgG assays and peripheral blood smears, respectively. Lateral flow chromatographic immunoassay (CTK Biotech) and an indirect immunofluorescence technique (biochip antigen substrate–coated slides; Euroimmun) were used to detect anti-CHIK virus IgM and IgG antibodies, respectively.
As controls, we used blood samples that had been donated by 80 consenting healthy villagers (without any symptomatic chronic illness or arthritis in particular) in 2005 (a year prior to the CHIK virus epidemic in India) during a community-oriented program for control of rheumatic diseases in the current rural region () and suitably stored at the Center for Rheumatic Diseases, Pune. Healthy control samples tested seronegative for dengue virus and CHIK virus antibodies.
Patients were randomized to receive CQ (250-mg tablet) or meloxicam (7.5-mg tablet) once daily after lunch (Figure 1). A predetermined fixed quantity of oral acetaminophen (500-mg tablet) was provided to each patient for use as a rescue analgesic on an as-needed basis. No other analgesic, nonsteroidal antiinflammatory drug (NSAID), or oral steroid was permitted. Stable doses of concomitant medications (e.g., for diabetes or hypertension) were continued. All patients, including those in the low pain cohort (receiving rescue acetaminophen only) were provided medication at no cost.
Designated healthcare workers distributed medication and monitored compliance. Another group of healthcare workers recorded patient-centric measures in a blinded manner. Doctors completed the joint and safety evaluation in a blinded manner. After checks for consistency and errors were performed, blinded data were locked by one of the authors (AC), and a copy was given to another author (MS) and a biostatistician for decoding and analysis.
At least 30 eligible patients were enrolled in each arm of the study. An intent-to-treat analysis (last observation carried forward) was performed. Two-tailed P values less than 0.05 were considered significant. Data were entered into MS Excel spreadsheets and imported into SPSS version 12 for analysis.
Patients in the CQ treatment arm (n = 38) and the meloxicam treatment arm (n = 32) were matched at baseline for sex (24 women and 31 women, respectively), mean age (50.2 years and 45.4 years, respectively), mean duration of illness (9.6 weeks and 8.8 weeks, respectively), and efficacy measures. Although there is no known sex predilection for CHIK virus infection, women outnumbered men in this study. Patients had moderately severe pain but very few swollen joints (Table 1). None of the patients were seropositive for RF, ANA, and anti-CCP, at either baseline or study completion.
|Variable||Meloxicam (n = 32)||Chloroquine (n = 38)||Difference mean change|
|Tender joint count (range 0–68)||22.4||−17.2 (−22.5, −11.9)||20.1||−16.8 (−21.2, −12.3)||−0.45 (−7.3, 6.4)|
|Swollen joint count (range 0–66)||0.8||−0.6 (−1.1, 0.04)||1.3||−1.1 (−1.8, −0.4)||0.57 (−0.3, 1.5)|
|VAS score for pain (range 0–10)||5.7||−3.9 (−4.8, −3.1)||6.1||−4.2 (−48, −3.6)||0.24 (−0.8, 1.3)|
|HAQ score (range 0–24)||4.1||−3.1 (−4.2, −1.9)||3.9||−2.8 (−4.1, −1.4)||−0.31 (−2.1, 1.4)|
|Physician's global assessment (range 1–5)||1.7||−1.1 (−1.4, −0.8)||1.7||−1 (−1.2, −0.7)||−0.10 (−0.5, 0.3)|
|Patient's global assessment (range 1–5)||2.1||−1.3 (−1.7, −1.8)||2.3||−1.5 (−1.8, −1.1)||0.14 (−0.4, 0.7)|
|VAS score for general health (range 0–100)||50.9||−29.4 (−36.8, −22)||52.1||−29.2 (−35.9, −22.4)||−0.26 (−10.1, 9.6)|
|ESR (Westergren method), mm/hour||29.96||−10.95 (−19.9, −1.9)||37.64||−23.86 (−36.5, −11.3)||12.92 (−2.17, 28)|
Patients improved significantly (Table 1), as demonstrated by reductions in the VAS score for pain and several efficacy variables at completion. However, the difference by treatment intervention was not significant, except for a greater reduction in the ESR in the CQ group (P = 0.04 by Mann-Whitney test). The VAS score for pain and the tender joint count improved rapidly within 6 weeks of study initiation, and improvement continued at a slower pace until study completion, with no difference between interventions (Figure 2). There was no significant difference in the mean number of acetaminophen tablets consumed per month per patient by treatment group in the first month (n = 16.43 tablets in the CQ group and n = 11.85 tablets in the meloxicam group; P = 0.07), the sixth month (n = 7.73 and n = 10.3, respectively; P = 0.29), or over the total study period (n = 64.67 and n = 66.13, respectively; P = 0.87). Correspondingly, the mean monthly intake of acetaminophen in the low pain cohort was 11.80 tablets in the first month and 8.40 tablets in the sixth month.
Patients in both intervention arms (5 in the meloxicam group and 7 in the CQ group) reported mild, brief adverse events that were mostly related to upper abdominal pain and discomfort (nausea, vomiting, diarrhea). Fewer than 3 patients each reported itching, headache, and fatigue; skin rash was not reported. Seven patients (10%) withdrew from the study, and none of the withdrawals was attributable to an adverse event.
None of the patients ever showed cytopenia or a blood hemoglobin level of <9 gm/dl (data not shown). The standard biochemical, hepatic, and renal parameters and creatine phosphokinase levels remained normal and did not change significantly (data not shown).
The serum levels of cytokines (except IFNγ) were increased several-fold at baseline and decreased by study completion; except for IL-6 in the CQ group, there were no significant differences according to treatment (Table 2). Figure 3 shows the changes in serum cytokine levels over time, including those in the low pain cohort.
|Healthy controls (n = 80)||Meloxicam (n = 32)||Chloroquine (n = 38)||Difference mean change (95% CI)a|
|Week 0||Week 24||Change||Week 0||Week 24||Change|
|CRP, mg/liter‡||≤5.l||3.47||2.93||1.27||3.09||2.90||0.98||1.29 (0.91, 1.84)|
|IFNγ, pg/ml||3.12||4.85||2.97||1.01||1.11||4.33||2.85||0.37 (0.01, 123.99)|
|CXCL10, pg/ml||3.16||172.28||40.54||6.37||148.03||23.85||9.42||0.68 (0.20, 2.24)|
|TNFα, pg/ml||32.29||243.03||47.09||3.46||214.46||69.34||4.63||0.75 (0.40, 1.37)|
|IL-6, pg/ml||51.12||361.44||274.32||2.93||339.34||352.38||2.72||1.01 (0.61, 1.92)|
|IL-13, pg/ml||181.44||919.75||653.34||2.89||826.75||636.57||2.87||1.00 (0.99, 1.01)|
The current study did not support the preferential use of oral CQ over meloxicam to treat early persistent musculoskeletal pain and arthritis following acute CHIK virus infection. At the time of enrollment, patients had moderately severe musculoskeletal pain and arthritis (duration at least 6 weeks), a high ESR, and a several-fold increase in cytokine levels (except IFNγ) and were treated for 24 weeks. Patients improved significantly. Although treatment with CQ resulted in better numerical improvement (tender joint count, oral acetaminophen intake, and ESR), there were no significant differences between the treatment groups. Drug-related adverse events were mild, and no safety issues were associated with CQ.
The CHIK virus epidemic returned to India after 3 decades ([1, 3]), and both the physicians and the community were caught unaware. The illness was managed primarily by general practitioners. Acute CHIK virus infection is not a rheumatologic disorder, but soon rheumatology outpatient clinics were inundated with patients with rheumatologic sequelae. We previously described a wide spectrum of rheumatologic disorders following acute CHIK virus infection that had not been reported earlier (). However, from a rheumatic disease perspective, we considered it necessary to study the acute illness and nature of early persistent musculoskeletal pain and arthritis. We added evaluation of the therapeutic role of CQ to our agenda.
We began with a community survey (at the current study site) and followed up the cases for more than 18 months (). The diagnosis of acute CHIK virus infection was based on clinical definitions () that were appropriate for an epidemic setting. However, as in the current study, serologic assessment (IgM and/or IgG antibodies) supported the diagnosis. Patients enrolled in the current study had experienced at least 6 weeks of persistent moderately severe musculoskeletal pain and arthritis, mostly a combination of polyarthralgias and soft tissue pain. Except for a disproportionately low CRP level, the laboratory results showed a severe inflammatory response. The proinflammatory cytokine phenotype was as intense as that reported for RA (). We previously observed persistent musculoskeletal pain and arthritis among patients in the community (16% at 4 months, 12% at 1 year, and 5% at 2 years). Unlike patients referred from our rheumatology center (), very few patients from the community who had chronic disease had significant synovitis or inflammatory polyarthritis (6% of the community cases and 72% of the referral cases) (). In the community, selection bias was greatly reduced, and patients with very early disease were selected.
We were surprised by the rapidly widespread use of CQ to treat CHIK virus infection during the epidemic, because the efficacy of this approach was unproven. We later learned that this treatment was also instigated by the demonstration of effectiveness of CQ to treat post-CHIK arthritis in an uncontrolled study with a small sample size from South Africa (). Predominantly, oral CQ was used in India. A potential for the use of CQ to treat viruses, including human immunodeficiency virus and CHIK virus, has been increasingly recognized ([10, 11]). A unifying hypothesis is the ability of CQ to alter the acidic interior milieu of cytoplasmic vesicles that is critical to virus survival (). Chloroquine can further inhibit the activation of macrophages and CD4+ T cells and the release of cytokines (). So far, however, the clinical evidence for its antiviral effect has been elusive.
A controlled drug trial of CQ in acute CHIK virus infection (CuraChik; ClinicalTrials.gov Identifier: NCT00391313) () was terminated prematurely, but the investigators concluded that CQ had no therapeutic efficacy (). An open-label short-term study demonstrated that hydroxychloroquine offered no additional advantage over a combination of NSAIDs and steroids in the treatment of extended symptoms of acute CHIK virus infection (). We do not recommend steroid treatment in patients with acute CHIK virus infection and early persistent musculoskeletal pain and arthritis. It is very reassuring to note that the large majority of patients in the current study recovered substantially within 8 weeks without steroids and probably are best managed with analgesics and NSAIDs. CQ has been used empirically to treat chronic cases of post-CHIK virus inflammatory arthritis in rheumatology practice ([4, 17]), but long-term controlled studies are required to support the effectiveness of this approach.
Our secondary objective was to use the serum cytokine response to study the inflammatory nature of musculoskeletal pain and arthritis and its recovery, and the biochemical effect of CQ. We selected several representative cytokines according to our discretion (TNFα, IL-6, IFNγ, IL-13) as well as from a literature search (CXCL10/IP-10). The healthy control subjects resided in the region, and their serum samples were obtained long before the epidemic. Although patients experienced dramatic relief of their symptoms by the 8-week study end point (as shown by the VAS score for pain and the tender joint count), the cytokine phenotype response remained up-regulated for a prolonged period of time. By the 24-week end point (study completion), the levels of TNFα and CXCL10 were significantly decreased and near normal. TNFα is a potent stimulant of CXCL10, which was reported to be up-regulated in acute CHIK virus illness (). IL-6 and IL-13 levels remained strongly up-regulated throughout the study period. We did not expect serum IFNγ levels to be high or to show any change, but the response was erratic and difficult to explain. There were no significant differences in cytokine assay results according to treatment allocation. Intriguingly, the cytokine assay and response (except that of IFNγ) were similar in the study intervention arms (moderately severe musculoskeletal pain and arthritis) and the low pain cohort (who received only rescue acetaminophen). Against this latter perspective alone, it was evident that CQ did not demonstrate any discernible effect on the cytokine assay in the current study.
In the current study, it was difficult to identify a signature cytokine phenotype for musculoskeletal pain and arthritis. We were unable to explain the prolonged cytokine response following clinical recovery, although some patients may have experienced subclinical viral persistence. During the acute CHIK virus illness, we carried out a more comprehensive cytokine assay (monocyte chemoattractant protein 1, IL-4, and IL-10 in addition to the cytokines assessed in the current study) and observed an expected intense Th1 cell response and a fairly early heightened Th2 cell response; however, none of the cytokines was significantly correlated with the duration and/or persistence of musculoskeletal pain and arthritis symptoms and/or early resolution (Chopra A: unpublished observations).
We previously reported an unexplained persistent disconnect between a high IL-6 level and a low CRP level during long-term followup of symptomatic patients ([3, 18]), which was also observed in the current study. IFNγ expression is an immediate response in acute viral illness () and is generally short lived. It is prudent to add that the viremia phase in CHIK virus infection does not extend beyond the initial week or so (). The persistent heightened IL-13 and IL-6 responses observed in the current study were intriguing () and seem to be a unique phenomenon in CHIK virus infection that has not been reported in the literature. Overall, this demonstrates a strong Th2 cytokine response. We speculate that this reflects a failure to curb macrophage activation in response to some degree of viral persistence. The pivotal role of macrophages in CHIK virus infection and its sequelae is well recognized in experimental studies ([20, 21]). Although several surrogate markers for acute and chronic arthralgias associated with CHIK virus have been considered, no evidence supports the value of these markers in the etiology and prognosis ().
After collating these data with those of our earlier studies, we believe that chronic musculoskeletal pain and arthritis following CHIK virus infection is probably attributable to some degree of virus persistence that otherwise has been well demonstrated in the case of chronic Ross River virus infection (); Ross River virus and CHIK virus are closely related arboviruses. It is prudent to add that despite a long-term intense Th2 cell response in our patients, we did not observe an increased incidence of autoantibodies (e.g., RF, ANA, anti-CCP). We further speculate that there is a complex interaction with genetic factors for enhanced susceptibility, severity, and chronicity of musculoskeletal pain and arthritis and other well-defined rheumatologic syndromes following CHIK virus infection (). A critical unanswered question is whether CHIK virus infection per se will lead to an increased incidence of chronic autoimmune arthritis or connective tissue disorders.
Clinical epidemiologic field studies are challenging and cumbersome in any setting and especially in a grossly underdeveloped rural setting. The current study was limited by a relatively small sample size at a single site. Also, we did not carry out any investigations to isolate (culture) the virus or to measure the viral load using reverse transcription–polymerase chain reaction. We shuttled between our urban rheumatology center and a poorly accessible village to examine patients in their homes. We encountered several logistic problems. We also exposed ourselves to an epidemic that swept across central and south India in fewer than 6 months ([1, 3]); one of the team physicians contracted acute severe CHIK virus infection during the study. We are not aware of another truly community-based drug trial of similar nature or design. To the best of our knowledge, this was the first controlled evaluation of CQ in CHIK virus–related arthritis and rheumatism.
In conclusion, our data do not support a meaningful therapeutic role of oral CQ in the management of early persistent musculoskeletal pain and arthritis following acute CHIK virus infection in the community. Treatment of the symptoms with acetaminophen and NSAIDs may suffice. Several inflammatory cytokines remain up-regulated and persist despite clinical recovery, and these need further investigation.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Chopra had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Chopra, Saluja, Venugopalan.
Acquisition of data. Chopra, Saluja, Venugopalan.
Analysis and interpretation of data. Chopra, Saluja, Venugopalan.
We thank Drs. Meenaz Modak and Harbeer Ahedi for assisting in examining the patients, the dedicated Center for Rheumatic Diseases staff (paramedics, nurses, and laboratory technicians), and Dr. Ravi Ghorpade for his work ensuring timely completion of the study. The statistical analysis was carried out under the supervision of Dr. S. Sarmukaddam. We thank the villagers of Bavi for their wholehearted support and generous cooperation in completing the project.