Human immunodeficiency virus–associated polymyositis: A longitudinal study of outcome




To determine the clinical course and optimum treatment of human immunodeficiency virus (HIV)–associated myositis.


Sixty-four patients attending a county outpatient HIV/acquired immunodeficiency syndrome facility were referred for the presence of elevated creatine kinase (CK) levels or muscle weakness. Patients underwent neurologic and rheumatologic evaluation, electromyography, and muscle biopsy after exclusion for recreational drug or alcohol use, metabolic/endocrine disorders, zidovudine therapy, and other infections.


Thirteen patients (20%) had biopsy-proven myositis. The median duration of HIV infection prior to diagnosis of myositis was 4.3 years (range 0–11 years). Six patients had concomitant diffuse infiltrative lymphocytosis syndrome. There was no correlation of severity of weakness, stage of HIV infection, or retroviral treatment with the CK level at diagnosis. Eight patients received prednisone (60 mg/day) with 5 attaining complete resolution of myositis. The remaining 3 patients received immunosuppressive therapy (azathioprine or methotrexate and intravenous immunoglobulin) and had normalization of strength and CK. Four patients had spontaneous resolution of their myositis without treatment.


HIV-associated myositis occurs at any stage of HIV infection, has a relatively good prognosis, responds well to immunosuppressive therapy, and has little evidence of adverse outcome on the HIV infection.


Myalgias and weakness are common in human immunodeficiency virus (HIV)–infected patients. These symptoms are caused by a heterogeneous group of disorders (1). Whereas uncomplicated myalgias associated with viremia (i.e., HIV or hepatitis C) and fibromyalgia typically require only symptomatic treatment, myopathies associated with HIV or its treatment deserve special consideration because these disorders are potentially disabling and life threatening.

Thus far, the myopathies associated with HIV include a necrotizing noninflammatory myopathy (2, 3); rhabdomyolysis due to a variety of infectious, toxic, or other causes (4); nemaline rod HIV myopathy (5, 6); pyomyositis (7, 8); zidovudine (ZDV)–associated myopathy (9–11); dermatomyositis (12); inclusion body myositis (13); and an HIV-associated polymyositis (14, 15). HIV-associated polymyositis is clinically and histologically indistinguishable from an autoimmune polymyositis (14). In addition to this, HIV-infected individuals have been diagnosed with infectious myopathies (excluding pyomyositis) due to microsporidiosis (16), cryptococcus neoformans (17), and toxoplasmosis (18). Although many cases of polymyositis attributed to HIV have been described, the prevalence, pathogenesis, clinical course, and appropriate therapy remain unclear. Confounding these considerations are the descriptions of both asymptomatic HIV-associated polymyositis and a ZDV-associated inflammatory myopathy, thus potentially underestimating the number of cases of HIV-associated polymyositis.

The present report describes and analyzes 13 cases of HIV-associated polymyositis that were evaluated prospectively and followed longitudinally in an outpatient HIV county clinic setting. An additional 2 patients, evaluated by at least 1 of the coauthors independent of this study, are also reported. The natural history, electrophysiologic manifestations, pathologic features, and results of treatment of this disorder are described.


From July 1993 to the present we have prospectively studied 13 HIV-1 infected patients who have been diagnosed with an HIV-associated inflammatory myopathy. The presence of HIV infection was detected by enzyme-linked immunosorbent assay and confirmed by Western blot. The patients' lymphocyte subsets, acquired immunodeficiency syndrome (AIDS) defining illnesses, and current medication regimens were recorded. Eleven patients in whom a myopathy was suspected (subjective or objective complaints of weakness with an elevated creatine kinase [CK] level) were initially evaluated and followed for their HIV infection at the Thomas Street Clinic of the Harris County Hospital District, Houston, Texas, which provides ambulatory care to HIV-infected patients. Two additional patients enrolled in this prospective study were initially identified at the University of Texas Rheumatology Clinics and The Diagnostic Clinic, both located in Houston, Texas. Informed consent was obtained from all patients, and this study was reviewed and approved by the Committee for the Protection of Human Subjects at the University of Texas-Houston Health Science Center.

During the course of initial evaluation for a potential inflammatory myopathy, patients found to have an elevated CK level (≥2 times the upper limit of normal [≥448 IU/liter]) on 3 consecutive occasions were referred to the University Clinical Research Center (UCRC) at Hermann Hospital in Houston, Texas. In patients concurrently receiving ZDV, the medication was withdrawn for a period of 6 weeks. If the patient's symptoms and CK did not improve with the discontinuation of ZDV, the medication was not considered the attributing cause of the myopathy and the patient was allowed to resume the therapy. Exclusions for participation in this study included positive urine toxicology screens for amphetamines, cocaine, narcotics, or barbiturates; and previous diagnosis of muscular dystrophy, thyroid disease, persistent electrolyte disorders, polymyositis, or other systemic connective tissue disease.

Those qualifying patients thus identified underwent a formal neurologic evaluation and the following serum testing: ultrasensitive thyroid stimulating hormone, antinuclear antibody (ANA) by immunofluorescence, human T cell lymphotrophic virus type I (HTLV-I), Toxoplasma gondii antibody titers by indirect immunofluorescence, and antibodies to echovirus and coxsackievirus A/B by complement fixation. Electromyography (EMG) and a muscle biopsy were performed on those patients with no identifiable alternative cause for their weakness and/or elevated CK. Needle EMG was performed with concentric, disposable needle electrodes. The duration of 20 motor units sampled from 1 muscle was obtained and compared with normative values. A myopathic EMG was defined as the following in 2 or more muscle groups: 1) the presence of small, brief, and polyphasic motor unit potentials; 2) fibrillation, positive sharp waves, and increased insertional irritability; and 3) complex repetitive discharges. The muscle biopsy was performed on the vastus lateralis contralateral to the side upon which the EMG was performed. Biopsies were examined by light and electron microscopy using standard techniques.

Inflammatory myositis was defined by the presence of a mononuclear inflammatory infiltrate and atrophy, degeneration, necrosis of muscle fibers especially in a perifascicular distribution, and regenerating fibers. Additionally, the presence of mitochondrial changes or inclusion bodies (suggesting ZDV-associated myositis) were recorded and individuals with either finding were individually evaluated for the potential exclusion.

All patients were prospectively followed in a clinical setting. After full disclosure of potential consequences, symptomatic patients electing treatment for their inflammatory myositis were initiated on prednisone at 60 mg per day. Nonsteroidal antiinflammatory drugs (NSAIDs) were administered for symptomatic relief of myalgias or arthralgias. Three patients who had only partial response to the prednisone therapy were given a trial of intravenous immunoglobulin (IVIG) at 2 gm/kg. Response to treatment was monitored in monthly clinical evaluations. Muscle strength testing, serum CK, and the occurrence of opportunistic infections, malignancies, or other complications were evaluated and noted. Immunosuppressive therapy was tapered as tolerated based on the patient's clinical improvement and decrease in CK levels.

Statistical analyses.

Comparisons among continuous variables were performed using an analysis of variance using the EPI-INFO statistical program (Centers for Disease Control and Prevention, Atlanta, GA). Continuous laboratory variables, such as the CD4+ counts or CKs, are presented as mean ± standard deviation.



From July 1993 to July 2001, 64 HIV-1 infected patients had been referred to the rheumatology clinic at the Thomas Street Clinic specifically for myalgias, muscle weakness, or an elevated CK level. Of those patients, 11 (17.2%) were subsequently identified to have biopsy-proven inflammatory myopathy that was not attributable to the use of ZDV. During this same period of time, there were 4,988 unreduplicated new patients seen at this county HIV outpatient facility, giving a prevalence of 11/4,988 or 0.22%. Of all patients evaluated at the UCRC (11 from the county clinic and 2 from the faculty clinics), 13 demonstrated an HIV-associated polymyositis (Table 1). Twelve patients were men; the mode of HIV transmission included homosexual activity (7 patients), intravenous drug use (3 patients), and heterosexual activity (3 patients). There were 8 African Americans, 3 Hispanics, and 2 whites. The patients' ages ranged from 30 to 67 years (mean, 42 years) and the mean duration of HIV infection prior to diagnosis of myositis was 4.3 years (±3.4 years). At the time of presentation, 6 (46%) patients had been diagnosed with AIDS characterized by opportunistic infections or depressed CD4+ T cell count (CD4+ < 200/mm3) (19). Three patients had been receiving ZDV at the time of suspected diagnosis; however, as the protocol specified, their symptoms and CK levels persisted 6 weeks after discontinuation of therapy. Three patients were receiving no antiretroviral therapy and the rest of the patients were receiving a combination of protease inhibitors and nucleoside reverse transcriptase inhibitors.

Table 1. Demographics of 13 patients with HIV-associated polymyositis*
PatientSexDOBEthnicityMOTYr of HIV DxAIDSYr of myositis Dx
  • *

    HIV = human immunodeficiency virus; DOB = date of birth; MOT = mode of HIV transmission; Yr of HIV Dx = year of HIV diagnosis; AIDS = acquired immunodeficiency syndrome; Yr of myositis Dx = year of myositis diagnosis; AA = African American; HMS = homosexual; H = Hispanic; HTS = heterosexual; IVDU = intravenous drug use; W = white.

2M3/41HHTS1996 1997
3M2/54AAHMS1989 1994
6M7/65AAHMS1990 1997
7M6/59AAIVDU1991 1994
8M11/54AAIVDU1989 1994
12M1/58AAIVDU1984 1995
13M10/32WHMS1998 1999

Clinical features.

Most patients (70%) presented with insidiously progressive muscle weakness primarily affecting the lower extremities (Table 2). Although activities of daily living were limited in those with muscle weakness, only 1 patient was debilitated due to myopathy, requiring use of a wheelchair for ambulation (a detailed evaluation of function using standardized validated instruments such as the Health Assessment Questionnaire or Short Form 36 was not carried out here). No patient had any gross evidence of diaphragmatic involvement clinically (i.e., requiring mechanical ventilation), but more detailed testing of diaphragmatic function was not performed. Only 1 patient complained of mild dysphagia, which could not solely be explained by muscular involvement. Four patients with and an additional 2 without muscle weakness (46% in total) experienced significant myalgias. Two patients presented with only an elevated CK level and sicca symptoms with associated parotid enlargement. Six patients (46%) had a concomitant diagnosis of diffuse infiltrative lymphocytosis syndrome (DILS) as diagnosed by the presence of sicca symptoms, parotid enlargement, and a positive minor salivary gland biopsy and/or positive gallium citrate Ga67 scintinography (20). These patients did not exhibit any other extraglandular manifestations associated with DILS. Weight loss of greater than 10% total body weight was demonstrated in 2 patients. Additionally, 2 patients complained of a glove distribution sensory loss and paresthesia, which was attributed to carpal tunnel syndrome. No other systemic manifestations seen in patients with idiopathic polymyositis, such as fibrosing alveolitis, interstitial lung disease, polyarthritis, Raynaud's phenomenon, or sclerodactyly, were encountered in these patients. Moreover, none of these 13 patients at the time of the evaluation had cancer, were taking other drugs implicated in myopathy (such as statins, colchicine, antimalarials, clofibrate, or corticosteroids), or were knowingly exposed to any toxins or occupational exposures (all were currently unemployed) that could have otherwise explained their muscle disease. All but 1 of the patients were men, and we did not specifically inquire about the presence of silicone implants of any sort (none were obvious on physical exam).

Table 2. Clinical features of 13 patients with HIV-associated polymyositis*
PatientSymptomsDILSCK (IU/liter)CD4+CD8+EMGMuscle biopsyHIV medication
  • *

    HIV = human immunodeficiency virus; DILS = diffuse infiltrative lymphocytosis syndrome; CK = serum creatine kinase level, in IL/liter; EMG = electromyography; pmw = proximal muscle weakness; nm = necrotizing myopathy; NRTI = nucleoside reverse transcriptase inhibitor; PI = protease inhibitor; nl = normal; if = insertional fibrillation; ZDV = zidovudine. CD4+ and CD8+ lymphocyte counts at time of diagnosis of myositis, in cells/mm3. Muscle biopsy results presented as degree of inflammatory infiltrate.

1Pmw, myalgia+2,5129521,596nmModerateNRTI
4Pmw 2,354921,540nmModerateNRTI/PI
5Pmw, myalgia 2,211105715nlMildNRTI/PI
6Pmw 1,4104151,516nmModerateNRTI/PI
7Pmw, myalgia+7766091,740nlProminentZDV
9Pmw, myalgia 1,38553872nlProminentNone
10Pmw 2,810117890nmProminentNRTI/ZDV
12Myalgia 1,832552984nlMildNone
13Myalgia 7,490392133nmMildPI/ZDV

Laboratory findings.

CK levels were elevated over a wide range (536–7,490 IU/liter) with a median value of 2,484 IU/liter (±1,787 IU/liter) without a significant correlation with severity of weakness. In fact, the patient with the highest CK experienced only myalgias. There was no significant difference in CK levels between those patients with asymptomatic HIV versus AIDS or between different antiretroviral therapies. At the time of diagnosis of myositis, the median CD4+ count was 381/mm3 (±263/mm3) and the median CD8+ count was 1,329/mm3 (±791/mm3). It is noteworthy that the CD4+ counts ranged widely at diagnosis (53–952/mm3). The average CD4+/CD8+ ratio was 0.51. Test results for antibodies for HTLV-I, Toxoplasma gondii, echovirus, and coxsackievirus A/B were negative in all patients. Results for ANAs by indirect immunofluorescence were negative in all. No comprehensive evaluation was done for other autoantibodies associated with idiopathic polymyositis, such as anti-Ro, anti-La, anti-Sm, anti-RNP, or tRNA-synthetase antibody.

Electrodiagnostic findings.

Electromyography revealed necrotizing myopathy with polyphasic small motor unit potentials, fibrillations, and increased insertional irritability in 8 patients (61%). One patient's EMG displayed only insertional irritability and 4 patients (31%) had a normal EMG. As with the CK, there was no correlation between EMG findings and stage of HIV infection or antiretroviral therapy.

Pathologic findings.

Although the quadriceps, iliopsoas, and deltoid were weak in all patients, the iliopsoas is not easily accessible for biopsy and the deltoid muscle has normally an uneven mix of type 1 and type 2 fibers; thus we selected the vastus lateralis for biopsy. All showed the presence of mild to moderate mononuclear inflammatory endomysial, perimysial, or interstitial infiltrates. Three patients had a prominent inflammatory response, 4 patients had a moderate response, and 6 had mild inflammation. Seven patients had fiber atrophy and 4 had necrotic fibers. Although neither nemaline rods, abnormal mitochondrial/fiber inclusions, nor inclusions compatible with inclusion body myositis were seen, 1 patient did have ragged red fibers. Although this can be seen with the use of ZDV (8), review of the medical records revealed no evidence of its use in this patient. Of uncertain significance, 5 patients had a type 2 fiber predominance, beyond that which would be expected in the biopsied muscle group (21).

Clinical course.

With the presence of demonstrable muscle weakness, 9 patients (70%) received immunosuppressive therapy (Table 3). Prednisone was initially administered to all 9 of these patients at 60 mg/day and tapered as soon as the clinical response permitted. Five patients (55%) attained complete remission of their myositis (both strength and CK normalized) and were able to discontinue prednisone in a mean of 9 months (±3 months). An additional patient, who was responding well to the use of prednisone, died of severe Pneumocystis carinii pneumonia 4 months into his immunosuppressive treatment at which time he was on 20 mg of prednisone a day.

Table 3. Clinical course of 13 patients with HIV-associated polymyositis*
PatientMyositis treatmentMyositis resolutionSurvivalStatusLast f/u
  • *

    HIV = human immunodeficiency virus; Survival = Survival after myositis diagnosis; f/u = followup; CS = corticosteroids; MTX = low-dose oral methotrexate; AZA = azathioprine; IVIG = intravenous immunoglobulin.

  • Patient currently alive.

  • Patient died prior to complete resolution.

1CS/MTX/AZA/IVIGIncomplete42 monthsDeceased12/99
2CSResolved40 monthsDeceased9/99
3NoneResolved37 monthsDeceased3/97
4CSResolved50 monthsAsymptomatic4/01
5CS/IVIGResolved96 monthsMyalgias1/01
6CS/IVIGResolved48 monthsAsymptomatic5/01
7CSResolved58 monthsDeceased2/99
8NoneResolved84 monthsMyalgias5/01
9CSResolved36 monthsAsymptomatic3/01
10CSIncomplete4 monthsDeceased12/93
11CSResolved44 monthsDeceased10/98
12NoneResolved71 monthsWeight loss2/01
13NoneResolved22 monthsAsymptomatic5/01

Three of the patients initially treated with corticosteroids had a suboptimal response to this therapy with persistence of muscle weakness or elevated CK levels. All 3 received IVIG at approximate doses of 2 gm/kg in addition to continuation of the prednisone. At the initiation of the IVIG, 2 patients had adverse events (worsening of muscle weakness in 1 and severe headaches in another) of sufficient severity to preclude further dosing. One of these patients had previously taken (and proved intolerant to) methotrexate, 7.5 mg/week and subsequently azathioprine, 150 mg/day. The third patient had no adverse effects, but after 12 months of therapy he had no significant improvement and further treatments were withheld. Over a period of 2–4 years, these patients' CK levels normalized, as did their muscle strength (albeit, some persistent mild subjective generalized weakness was noted).

Four patients who had no demonstrable evidence of muscle weakness received no immunosuppressive therapy. For those with myalgias, NSAIDs were given with good response. The CK levels improved spontaneously over a period of 10 months to 6 years. None of these patients ever exhibited objective evidence of muscle weakness.

Since the inception of this study, 7 patients (54%) have died, with 5 deaths attributed to complications of their HIV infection. Although the treatment of myositis with prednisone may have contributed to an opportunistic infection ultimately leading to 1 patient's death, no death could be directly related to complications of myositis. Not surprisingly, 4 of the 7 patients who have thus far died had AIDS at the time of diagnosis of the myositis. The current mean survival after diagnosis was 48 months (±25 months).


In addition to the 13 patients described in this cohort, authors of this report (S.K. and J.D.R.) have clinically evaluated 2 other HIV-infected patients who were diagnosed with HIV-associated polymyositis (these patients were not included in the group above because they did not undergo the systemic prospective evaluation as outlined in the Patients and Methods section). One of these patients was a 40-year-old white homosexual man who had been diagnosed with HIV in 1988. In 1990, when he presented with proximal muscle weakness and an elevated CK level, he had developed AIDS with a CD4+ cell count of 78/mm3 and Pneumocystis carinii pneumonia. His CK level at the time of diagnosis was 341 IU/liter; however, he had both necrotizing myopathy on EMG and an inflammatory myopathy on muscle biopsy. No nemaline rods, ragged red fibers, or inclusion bodies were identified. Due to his debilitated state of health secondary to his HIV infection, he never received therapy for his myositis. He died of acute respiratory failure and sudden cardiac death 1 month after he had been diagnosed with HIV-associated polymyositis.

The second patient was a 55-year-old African American man who contracted HIV through intravenous drug use. He had been diagnosed with HIV infection in 1986 and in 1991 presented with proximal muscle weakness, myalgias, and a CK level of 2,450 IU/liter. His CD4+ and CD8+ lymphocyte counts at that time were 358/mm3 and 1,992/mm3, respectively. He too had necrotizing myopathy on EMG and inflammatory myopathy on muscle biopsy. He received prednisone at 60 mg/day with initial improvement in both his symptoms and CK level. However, he has subsequently had several recurrences of an elevated CK level and muscle weakness and thus has remained steroid dependent.


Although musculoskeletal complaints frequently occur in HIV-infected patients, the low prevalence shown in this study suggests that HIV-associated polymyositis is relatively rare. Many case reports have been documented and a few case series have been reported on this entity; however, this report describes the largest series of patients with biopsy-proven polymyositis prospectively followed in a longitudinal study that spans 8 years. We encountered no cases of inclusion body myositis or dermatomyositis in this cohort, not altogether surprising, given that these have been only the subject of a few case reports (12, 13).

This study evaluated only those patients with a muscle biopsy documenting an inflammatory infiltrate. As has been shown elsewhere and in this study, interstitial inflammatory infiltrates may often be mild and focal and thus a single muscle biopsy may miss a diagnostic lesion. Additionally, multiple cases of “regenerating myopathy” and atrophy exclusive of inflammatory infiltrate have been described that may represent postinflammatory polymyositis lesions (22). Furthermore, polymyositis in HIV/AIDS appears to be a more subtle disease from a symptomatic standpoint, and mild proximal muscle weakness may not be recognized in a population of people who often present with generalized weakness and a debilitating course. In fact, 4 of the 13 patients from this cohort would not have been diagnosed with an inflammatory myopathy had a CK level not been assessed by routine screening for ZDV toxicity.

As in autoimmune polymyositis in HIV-negative individuals, the pathogenesis of HIV-associated polymyositis is unclear. The direct role of HIV in this process remains controversial. Whereas 1 study reported HIV-1 nucleic acids in the myocyte nuclei (23), the majority of investigations have detected only HIV sequences and antigens in the surrounding lymphoid cells. Using in situ hybridization, polymerase chain reaction, immunocytochemistry, and electron microscopy, no study has thus far been able to demonstrate the virus itself within the muscle fibers (24).

In both HIV-positive and -negative polymyositis, the predominant cells invading or surrounding major histocompatibility complex I antigen expressing non-necrotic muscle fibers are cytotoxic CD8+ lymphocytes and macrophages (14). Cytokines or toxic lymphokines released by the endomyseal inflammatory cells may precipitate muscle antigen exposure whereby the muscle fibers then act as antigen-presenting cells for which there is no self tolerance, thus generating an autoimmune response (24). Consequently, this immune process may be triggered by a systemic HIV infection with no direct viral infiltration of the muscle. This may be further aggravated by the relative imbalance of CD4+ lymphocytes, but the significance of this is uncertain.

In this study, HIV-associated polymyositis appears to be a milder variant of idiopathic autoimmune polymyositis. Overall muscle strength was diminished but generally was mild to moderate, with most activities of daily living preserved. Development of polymyositis in HIV-positive patients appears to be unrelated to the degree of associated immunodeficiency. It occurred in all stages of HIV infection from asymptomatic individuals to those with advanced AIDS and with no correlation to CD4+ or CD8+ lymphocyte counts. Previously undescribed, nearly half of the patients in this cohort had been diagnosed with DILS either prior to the diagnosis of myositis or concomitantly. Similarly, as in the inflammatory endomyseal infiltrate, DILS is characterized by elevated levels of circulating and infiltrating CD8+ lymphocytes. It is unclear if the concurrent occurrence of these 2 disorders represents a spectrum of a single disease, a separate syndrome, or a coincidental phenomenon of the cytotoxic effects of CD8+ lymphocytes.

By the criteria established by Bohan and Peter (25), 6 of the 13 patients in this cohort met definitive criteria for polymyositis; 5 met probable criteria; and only 2 were limited to possible polymyositis. These 2 patients with possible polymyositis lacked objective muscle weakness and had normal EMGs. Normal EMGs were encountered in 30% of the patients. Despite this, all patients had inflammatory infiltrates on muscle biopsy. Thus a normal EMG should not preclude a muscle biopsy in an HIV-infected patient in whom there is a clinical suspicion of a myopathy. There was a wide spectrum of inflammatory infiltrates exhibited on muscle biopsies ranging from mild and focal to prominent and diffuse. Nonetheless, the severity of inflammatory infiltrate did not correlate with symptoms, stage of infection, nor response to corticosteroids. The finding of ragged red fibers in 1 patient was perplexing. Although this is often indicative of ZDV therapy in HIV-infected patients, this patient had not received ZDV in the last 14 months. It had not previously been recognized that he had muscle weakness or an elevated CK level while taking ZDV. Despite the association of ragged red fibers with ZDV-associated myopathy, Simpson et al (26) have reported ragged red fibers in ZDV-naïve patients and thus suggest the use of mitochondrial abnormalities is unreliable to distinguish myopathies related to HIV versus those related to ZDV.

The treatment of polymyositis in an HIV-positive patient is similar to that in HIV-negative individuals. Corticosteroids in doses of 0.75–1 mg/kg are effective for improving muscle strength and lowering CK levels. In our series, 6 of the 9 patients were given corticosteroids alone. There was no significant difference between responders versus nonresponders regarding stage of infection or degree of inflammatory infiltrate. Despite reluctance to use immunosuppressive agents in an immunodeficient patient, our study and others (1, 11, 15) would suggest that most patients have few adverse effects if used judiciously. Careful monitoring of the patients' clinical status, CD4+ lymphocyte counts, and HIV mRNA levels should be performed at least monthly, especially in those with advanced disease. Whereas resistant cases do not appear to respond to IVIG, other immunosuppressive agents such as methotrexate or azathioprine may prove to be useful (1).

The fact that most of the cases were diagnosed before 1998, when protease inhibitors and nonnucleoside reverse transcriptase inhibitors became widely available in our clinic, suggests that these agents may affect predisposition to HIV-associated myositis. An effect of antiretroviral treatment on a related HIV-associated rheumatic disease, DILS, has also been described (20). It is possible that antiretroviral treatment itself may affect the prevalence of this disease. However, most patients' symptoms resolved regardless of antiretroviral treatment.

Whether asymptomatic patients should be treated remains unclear; however, our study would suggest these patients might have a spontaneous resolution of their inflammatory myopathy. Overall, HIV-associated polymyositis responds well to corticosteroids, rarely causes disability, and appears to have no adverse outcome on the course of the HIV infection.