Drs. Petri, Kalunian, Ramsey-Goldman, and Strand have served as consultants to Genelabs Technologies, Inc.
Effects of prasterone on disease activity and symptoms in women with active systemic lupus erythematosus†
Results of a multicenter randomized, double-blind, placebo-controlled trial
Article first published online: 9 SEP 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 50, Issue 9, pages 2858–2868, September 2004
How to Cite
Petri, M. A., Mease, P. J., Merrill, J. T., Lahita, R. G., Iannini, M. J., Yocum, D. E., Ginzler, E. M., Katz, R. S., Gluck, O. S., Genovese, M. C., Van Vollenhoven, R., Kalunian, K. C., Manzi, S., Greenwald, M. W., Buyon, J. P., Olsen, N. J., Schiff, M. H., Kavanaugh, A. F., Caldwell, J. R., Ramsey-Goldman, R., St. Clair, E. W., Goldman, A. L., Egan, R. M., Polisson, R. P., Moder, K. G., Rothfield, N. F., Spencer, R. T., Hobbs, K., Fessler, B. J., Calabrese, L. H., Moreland, L. W., Cohen, S. B., Quarles, B. J., Strand, V., Gurwith, M. and Schwartz, K. E. (2004), Effects of prasterone on disease activity and symptoms in women with active systemic lupus erythematosus. Arthritis & Rheumatism, 50: 2858–2868. doi: 10.1002/art.20427
Presented in part at the 64th Annual Scientific Meeting of the American College of Rheumatology, Philadelphia, PA, November 2000, and at the Annual Meeting of the British Society for Rheumatology, Manchester, UK, April 2003.
- Issue published online: 9 SEP 2004
- Article first published online: 9 SEP 2004
- Manuscript Accepted: 29 APR 2004
- Manuscript Received: 15 AUG 2003
- Genelabs Technologies, Inc., Redwood City, CA
To determine whether prasterone administration results in improvement or stabilization of systemic lupus erythematosus (SLE) disease activity and its symptoms.
Women with active SLE were treated with prasterone 200 mg/day plus standard SLE treatments or with placebo plus standard SLE treatments for up to 12 months in this randomized, double-blind investigation conducted at 27 centers. Standard SLE treatments included prednisone (≤10 mg/day), antimalarials, and immunosuppressive agents; dosages were required to be stable for ≥6 weeks prior to enrollment and remain unchanged during protocol treatment. Responders were patients who experienced no clinical deterioration and had improvement or stabilization over the duration of the study in 2 disease activity measures (the SLE Disease Activity Index [SLEDAI] and the Systemic Lupus Activity Measure) and 2 quality of life measures (patient's global assessment and the Krupp Fatigue Severity Scale).
A total of 381 women with SLE were enrolled. Among patients with clinically active disease at baseline (SLEDAI score >2), 86 of 147 in the prasterone group (58.5%) demonstrated improvement or stabilization without clinical deterioration, as compared with 65 of 146 in the placebo group (44.5%) (P = 0.017). Acne and hirsutism were reported in 33% and 16%, respectively, of the prasterone group and in 14% and 2%, respectively, of the placebo group (P < 0.05 for both comparisons). However, most cases of acne and hirsutism were mild and did not require withdrawal from therapy. Myalgias and oral stomatitis were reported less frequently in the prasterone group (22% and 15%, respectively) than in the placebo group (36% and 23%, respectively) (P < 0.05 for both comparisons). Serum levels of high-density lipoprotein cholesterol, triglycerides, and C3 complement significantly decreased, while levels of testosterone and, to a lesser extent, estradiol increased in the prasterone group.
In adult women with active SLE, administration of prasterone at a dosage of 200 mg/day improved or stabilized signs and symptoms of disease and was generally well tolerated.
Systemic lupus erythematosus (SLE) is a chronic, potentially fatal autoimmune disease that occurs 9 times more frequently in women than in men (1). Although the multifactorial etiology of this disease is poorly understood, abnormalities of both estrogen and androgen metabolism in SLE patients have been reported (2, 3).
Dehydroepiandrosterone (DHEA) is a naturally occurring steroid produced by the adrenal glands. It is secreted primarily as its metabolite, DHEA sulfate (DHEAS), which is the most abundant circulating adrenal steroid in humans (4). Both DHEA and DHEAS are subsequently converted into androgenic and estrogenic steroids in peripheral tissues (5, 6). Decreases of ∼50% in circulating levels of DHEA and DHEAS have been observed in patients with SLE (7, 8). Previous studies in animal models of SLE have demonstrated improvement with androgen administration, including DHEA (9–14). In addition to serving as a precursor for other androgenic and estrogenic steroids (5), there is evidence that DHEA has an immunomodulatory role, including up-regulation of interleukin-2 (IL-2) and down-regulation of IL-6 expression (9, 15–17), both of which have been reported to be abnormal in SLE (18–20).
Prasterone is the United States Adopted Names generic designation for dehydroepiandrosterone. In open-label and placebo-controlled studies, Van Vollenhoven et al (21, 22) reported that SLE patients receiving oral prasterone 200 mg/day had improvement in a number of outcome variables, including reduction of steroid dosages, the number of disease flares, and global assessments of disease activity (21, 22). In an initial phase II/III trial comparing placebo with 100 and 200 mg/day of prasterone, it was demonstrated that 200 mg/day of prasterone allowed a sustained reduction in the glucocorticoid dosage (23). In addition, delay in time to SLE flare has been reported for women treated with prasterone (24).
The present study was conducted to determine whether prasterone administration results in improvement or stabilization in SLE disease activity and its symptoms.
PATIENTS AND METHODS
This prospective randomized, double-blind, placebo-controlled trial conducted at 27 study sites evaluated female patients with active SLE. Patients were required to meet the American College of Rheumatology 1982 criteria for a diagnosis of SLE (25) and have active disease, as determined by 2 disease activity indices. Initially, active disease was defined as a Systemic Lupus Activity Measure (SLAM) score ≥7 at baseline (26). While the study was ongoing and blinded, this was amended to also require a Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) score >2 at baseline (27). Eligibility was determined at screening and qualifying visits, which occurred no more than 10 days apart.
Patients in both treatment groups were allowed to continue taking standard SLE medications. Baseline medications that were allowed included oral glucocorticoids (≤10 mg/day of prednisone or equivalent), hydroxychloroquine, and/or immunosuppressive agents, including methotrexate, azathioprine, cyclophosphamide, and mycophenolate mofetil.
Patients were assigned by predetermined randomization codes to receive either prasterone 200 mg/day or placebo as once-daily morning doses for up to 52 weeks. Capsules containing placebo were identical to those containing prasterone. Physicians and patients were instructed to maintain the dosages of prednisone and other baseline SLE medications at the baseline dosages during protocol participation.
The primary end point was the proportion of patients who were “responders.” Responders were patients who showed improvement or stabilization in SLE disease activity and constitutional symptoms without clinical deterioration over the duration of the study.
Responder status was designed prospectively, in conjunction with lupus experts and with significant input by the US Food and Drug Administration, to be a single composite end point that integrated 3 domains of SLE: disease activity, organ damage, and health-related quality of life (28). Responders were those who demonstrated improvement or stabilization in mean on-treatment scores for 2 disease activity measures (the SLEDAI and the SLAM) and 2 health-related quality of life assessments (patient's global assessment and the Krupp Fatigue Severity Score [KFSS]) (29) without evidence of clinical deterioration (reflecting organ damage). A patient was deemed to have improved or stabilized in terms of each of the disease activity and health-related quality of life measures if the time-weighted mean of all on-treatment visit measurements for each of the instruments for that patient was less than the mean of 2 pretreatment values for each of the 4 parameters. “No change” was defined prospectively, while the study was ongoing and blinded. The definition of “no change” allowed for test–retest variability in these scoring instruments, which was defined as ±0.5 for the SLEDAI and KFSS, ±1.0 for the SLAM, and ±10 mm for the patient's global assessment (30, 31).
The clinical deterioration component of the responder end point was prospectively defined to include serious drug toxicity attributable to the study drug or other lupus therapy if it occurred during treatment with the study drug or within 6 weeks after discontinuation of the study drug, serious new or progressive lupus-related conditions, or requirement for increased dosage or institution of new therapy with immunosuppressive or cytotoxic agents for treatment of lupus. The occurrence of clinical deterioration and its onset date was determined by Genelabs' study monitors before the blinding was broken.
The following conditions qualified as serious drug toxicity: new-onset diabetes mellitus, defined as diabetes requiring drug therapy for ≥ 3 months; new gastric or duodenal ulcer not due to Helicobacter pylori and requiring hospitalization or transfusion; new-onset hypertension requiring drug therapy for ≥3 months; new myocardial infarction, as demonstrated by electrocardiographic or enzymatic criteria; new steroid myopathy; new elevation in serum transaminase levels (increases in aspartate aminotransferase or alanine aminotransferase levels to ≥8 times the upper limit of normal or a single measurement showing levels ≥3 times the upper limit of normal at multiple measurements over 3 months); or new fracture and/or vertebral collapse due to osteoporosis.
The following conditions qualified as major new or progressive organ disease. These conditions were assessed by the treating physician as being attributable to lupus or its treatment and which occurred during treatment with the study drug or within 6 weeks after discontinuation of the study drug. Central nervous system conditions were cerebrovascular accident transverse myelitis, retinal vascular occlusion, new onset of psychosis of >3 months' duration, or new onset of seizures that were refractory to therapy for at least 3 months. Renal conditions were new onset of end-stage renal disease or loss of renal function that required dialysis for at least 3 months. Pulmonary conditions were new or worsened pulmonary hypertension and/or interstitial lung disease with reduction in diffusion capacity, mean pulmonary artery pressure, and/or dyspnea at rest (New York Heart Association class IV). Cardiovascular conditions were pericarditis that was refractory to treatment for >3 months or that required pericardiectomy, cardiomyopathy that was refractory to therapy for >3 months with hemodynamic compromise (decreased cardiac index, left ventricular ejection fraction, and/or dyspnea at rest), or refractory arrhythmia. Gastrointestinal conditions were ischemic bowel disease that required bowel resection. Vasculitic conditions were vasculitis that resulted in infarction (excluding vasculitis described under any other organ systems). Hematologic conditions were thrombocytopenia that resulted in clinically significant hemorrhage with sequelae that did not resolve for at least 3 months or persistent leukopenia (white blood cell count <1,500/mm3) that resulted in recurrent infections without improvement in the incidence of recurrent infections for at least 3 months.
The following qualified as unacceptable increases in immunosuppressive or cytotoxic therapy for lupus: any SLE-related increase in dosages of concomitant methotrexate or azathioprine, or institution of new therapy with cytotoxic or immunosuppressive agents (methotrexate, azathioprine, cyclophosphamide, or cyclosporine), at any time during treatment with the study drug or within 6 weeks after discontinuation of study drug; and except for stress doses, prescribed prednisone dosage increase to >10 mg/day over the baseline dosage within the first 2 months of participation or, through the remainder of the study, prednisone dosage increase to >5 mg/day over the daily baseline dose for >2 consecutive months.
Prospectively defined secondary analyses included time to lupus flare and mean changes in individual scores on the SLEDAI, SLAM, KFSS, and patient's global assessment instruments. Time to lupus flare was not part of the initial protocol design. However, given the interest in lupus flares as a potential study outcome for future lupus protocols, it was proposed as a secondary outcome in this study.
Time to first lupus flare was analyzed from data derived from chart reviews of all enrolled patients and was determined while the study was ongoing and blinded. SLE flare was defined according to the following 5 criteria: 1) new or worse central nervous system lupus, vasculitis, or myositis requiring scoring on the SLEDAI and not present at a previous visit; 2) thrombocytopenia (<60,000 platelets/mm3), a hemoglobin value <7 gm/dl, or a decrease in the hemoglobin level of at least 3 gm/dl; 3) proteinuria with pyuria and/or hematuria treated with new use or increased dosage of glucocorticoids or immunosuppressive agents; 4) an increase in the glucocorticoid dose of ≥2.5 mg for at least 7 days for SLE-related reasons; or 5) new use or increase in dosage of immunosuppressive agents or antimalarials for at least 7 days for SLE-related reasons or hospitalization for new manifestation of SLE.
Scheduled evaluations at baseline and every 3 months included physical examinations, laboratory determinations, and scoring of the SLAM, the SLEDAI, patient's and physician's global assessments using 100-mm visual analog scales (VAS), and the KFSS.
Laboratory assessments were performed every 3 months. Blood samples were drawn after an 8-hour fast but were not timed to prasterone administration. Assessments included anti–double-stranded DNA (anti-dsDNA) antibodies, C3 and C4 levels, IgG and IgM anticardiolipin antibodies, serum lipid levels (total cholesterol, HDL-cholesterol, calculated LDL-cholesterol, and total triglycerides), serum chemistries, complete blood cell counts, urinalyses, and 24-hour urine collections for creatinine clearance and protein quantitations. Serum levels of 17β-estradiol, testosterone, and DHEAS were measured at baseline and the last visit. To avoid unblinding, these results were not reported to the investigators or study monitors until completion of the trial. All blood and urine assays were conducted at a central laboratory (Covance Laboratories, Indianapolis, IN), with the exception of DHEAS levels, which were performed by radioimmunoassay at Genelabs Technologies.
The protocol was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review board at each center. All patients gave written informed consent.
Given that this protocol utilized a responder end point that had never been used in a clinical trial, there could be no a priori estimation of responder rates. Hence, the sample size of 300 randomized patients was based on practical, rather than statistical, calculations.
The original protocol entry criterion required a SLAM score of ≥7 for the definition of active disease. There was no restriction on the SLEDAI score for patient entry. While the double-blind study was ongoing, the protocol was subsequently amended to incorporate a baseline SLEDAI score of >2 as an additional entry requirement. This requirement was based upon the outcome of an earlier Genelabs study, which revealed that SLE patients with little or no disease activity (SLEDAI ≤2) are likely to exhibit a high response regardless of treatment (23).
All randomized patients were included in the intent-to-treat analysis of safety (n = 381). All patients who met criteria for active disease (SLEDAI >2) at both the baseline and screening visits were included in the analysis of efficacy (n = 293). Patients without postbaseline assessments were designated, by default, as nonresponders.
The primary efficacy variable, proportion of responders, was analyzed using a logistic regression model using treatment as a factor. For secondary analyses, between-treatment comparisons of mean changes in disease activity indices (the SLEDAI and the SLAM), patient's global assessment, and the KFSS, and laboratory values were analyzed utilizing one-way analysis of variance, with treatment as a factor. Between-treatment comparisons for the number of patients with specific adverse events or clinically important treatment-associated changes in laboratory values were performed using chi-square test or Fisher's exact test. All statistical tests were 2-sided, and P values less than or equal to 0.05 were considered significant.
Characteristics of the study patients.
The trial was conducted at 27 office- or university-based rheumatology practices in the US, from February 1996 to June 1999. Three hundred eighty-one patients were randomized to receive treatment: 189 in the prasterone group and 198 in the placebo group. Treatment groups were well balanced at baseline with regard to age, menopause status, race, concomitant SLE medications, SLE scores, DHEAS levels, and other important laboratory values. Approximately 80% of patients in both groups were receiving standard SLE treatments at baseline, which included antimalarials, glucocorticoids, and/or other immunosuppressive agents. There were no statistically significant between-group differences in baseline characteristics for either the all-patient intent-to-treat analysis or the predefined patient group with active SLE (baseline SLEDAI >2) (Table 1).
|Intent-to-treat analysis group (all randomized patients)||Active SLE group (baseline SLEDAI >2)|
|Placebo (n = 192)||Prasterone (n = 189)||Placebo (n = 146)||Prasterone (n = 147)|
|Age, mean years||43.8||44.4||43.6||43.8|
|Prednisone dose, mean (median) mg/day||3.7 (2.5)||3.5 (3.8)||4.1 (5.0)||3.6 (5.0)|
|Medication use, %|
|Prednisone, immunosuppressives, or antimalarials||79.7||82.0||80.8||79.6|
|Composite responder index components, mean (median) score|
|SLEDAI||5.8 (5.0)||6.5 (6.0)||7.34 (6.0)||8.04 (8.0)|
|SLAM||12.0 (12.0)||12.2 (12.0)||12.46 (12.0)||12.69 (12.5)|
|Patient's global assessment||55.4 (57.0)||55.2 (57.0)||55.17 (56.7)||57.08 (58.5)|
|KFSS||5.6 (5.7)||5.5 (5.9)||5.56 (5.7)||5.61 (5.9)|
|Laboratory values, mean (median) [no. tested]|
|DHEAS, μg/dl||103 (50) [n = 163]||107 (61) [n = 165]||91 (47) [n = 121]||105 (61) [n = 127]|
|C3 complement, mg/dl||102.9 (102.0) [n = 192]||104.3 (100.0) [n = 187]||99.3 (97.0) [n = 146]||99.1 (97.0) [n = 146]|
|C4 complement, mg/dl||17.9 (16.0) [n = 192]||18.2 (17.0) [n = 187]||17.0 (15.0) [n = 146]||17.2 (15.0) [n = 146]|
|Anti-dsDNA antibody, IU/dl||23.45 (1.95) [n = 192]||36.08 (2.6) [n = 187]||29.4 (2.4) [n = 146]||43.5 (3.4) [n = 146]|
Seventy-four percent of the placebo group and 65.6% of the prasterone group completed 1 year of treatment (Table 2). There were no meaningful differences in withdrawals across the 2 groups, except for withdrawals due to adverse events, for which 5.7% in the placebo group and 14.3% in the prasterone group withdrew (P = 0.005). The differences in withdrawals primarily reflected androgenic adverse events in the prasterone group, the majority of which were mild, since there were no differences in the patterns of withdrawals for other types of adverse events.
|Placebo (n = 192)||Prasterone (n = 189)|
|Completed study drug||142 (74.0)||124 (65.6)|
|Discontinued study drug early||50 (26.0)||65 (34.4)|
|Lack of efficacy||9 (4.7)||11 (5.8)|
|Adverse event||11 (5.7)||27 (14.3)|
|Other||30 (15.6)||27 (14.3)|
Primary outcome measure, responder analysis.
The overall responder rates among the intent-to-treat group were 42.2% (81 of 192 patients) in the placebo group and 51.3% (97 of 189 patients) in the prasterone group (P = 0.074) (Table 3). In the population with active disease (SLEDAI >2), 44.5% (65 of 146 patients) in the placebo group and 58.5% (86 of 147 patients) in the prasterone group were responders (P = 0.017). In a post hoc analysis, significant differences between treatment groups persisted with increasing SLEDAI scores (Figure 1).
|Responders†||Patients with at least 1 SLE flare‡|
|Placebo||Prasterone||P||SLE flare||Time to first flare|
|Patients with active SLE||44.5 (65/146)||58.5 (86/147)||0.017||34.2 (50/146)||24.5 (36/147)||0.097||0.066|
|All patients||42.2 (81/192)||51.3 (97/189)||0.074||29.7 (57/192)||23.8 (45/189)||0.266||0.195|
Secondary outcome measures, time to SLE flare and mean changes in individual scoring instruments.
Among patients with active disease at baseline, fewer patients in the prasterone treatment group experienced a first flare during the study (24.5% taking prasterone versus 34.2% taking placebo, P = 0.066), and a trend for prolongation in time to flare was seen among patients who received prasterone (P = 0.097) (Table 3).
There were no statistically significant between-group differences in change from baseline in any of the individual components of the responder index (data not shown), but statistically significant differences in individual components of the responder analysis resulted in patients being classified as nonresponders (Table 4), suggesting a worsening of individual components of the composite responder index in more of the placebo-treated patients than prasterone-treated patients. The greatest differences between the prasterone and placebo treatment groups were defined by the proportion of patients reporting a worsening in the patient's global assessment (10.9% of the prasterone group versus 22.6% of the placebo group, P = 0.007) and in the SLEDAI scores (only 9.5% of the prasterone group versus 17.8% of the placebo group, P = 0.039 (Table 4).
|Criterion||No. (%) of placebo group (n = 146)||No. (%) of prasterone group (n = 147)||P|
|Clinical deterioration||13 (8.9)||15 (10.2)||0.705|
|Worsening of composite responder index component scores|
|SLEDAI score||26 (17.8)||14 (9.5)||0.039|
|SLAM score||15 (10.3)||10 (6.8)||0.288|
|Patient's global assessment score||33 (22.6)||16 (10.9)||0.007|
|KFSS score||21 (14.4)||16 (10.9)||0.367|
Although not defined in the specified analysis plan, it was of interest to assess responders among the patients who were not receiving antimalarials, glucocorticoids, or other immunosuppressive agents at baseline (∼20% of patients), given the high rate of background medication use in both treatment groups. Among patients not receiving these medications at baseline and who had active disease at study entry (i.e., baseline SLEDAI >2), responder rates were 42.9% (12 of 28 patients) in the placebo group and 70.0% (21 of 30 patients) in the prasterone group (P = 0.037).
Study drug administration was discontinued early in 65 (34%) of the 189 prasterone-treated patients and 50 (26%) of the 192 placebo-treated patients (P = 0.076). Early discontinuations were similar between treatment groups, with the exception of an increased number in the prasterone group withdrawing due to reported androgenic adverse events. Eleven patients (5.8%) with acne and/or hirsutism indicated these events as reasons for treatment discontinuation in the prasterone treatment group.
Serious adverse events were reported in 14% (27 of 189 patients) in the prasterone group and 17% (33 of 192 patients) in the patients in placebo treatment groups, respectively. Fourteen of these events in the prasterone group and 16 in the placebo group resulted in treatment discontinuation. While there were no deaths in the prasterone treatment group, there were 5 deaths during or shortly after completion of treatment in the placebo group, including 2 suicides, 1 death due to pulmonary hypertension, 1 sudden death, and 1 death from non-Hodgkin's lymphoma 6 weeks following protocol completion. Three patients were diagnosed as having cancer during the study, all of whom were in the placebo group: the patient with non-Hodgkin's lymphoma noted previously, 1 patient with carcinoma of the breast, and 1 patient with carcinoma of the lung.
Adverse events reported in 10% or more of the active and control treatment populations are shown in Table 5. Androgenic adverse events, including acne and hirsutism, were more commonly reported in patients receiving prasterone (42%) than in patients receiving placebo (18%) (P < 0.05). Most androgenic complaints were characterized as mild or moderate; none was severe.
|Adverse event||No. (%) of placebo group (n = 192)||No. (%) of prasterone group (n = 189)|
|Rash||62 (32.3)||75 (39.7)|
|Arthralgia||71 (37.0)||68 (36.0)|
|Acne||27 (14.1)||63 (33.3)†|
|Asthenia||51 (26.6)||45 (23.8)|
|Arthritis||42 (21.9)||45 (23.8)|
|Headache||56 (29.2)||42 (22.2)|
|Myalgia||69 (35.9)||42 (22.2)†|
|Flu syndrome||42 (21.9)||39 (20.6)|
|Hirsutism||3 (1.6)||31 (16.4)†|
|Stomatitis (mucosal ulcers)||44 (22.9)||28 (14.8)†|
|Depression||30 (15.6)||28 (14.8)|
|Alopecia||39 (20.3)||28 (14.8)|
|Abdominal pain||30 (15.6)||27 (14.3)|
|Fever||28 (14.6)||22 (11.6)|
|Peripheral vascular disease||20 (10.4)||19 (10.1)|
|Sinusitis||20 (10.4)||17 (9.0)|
|Chest pain||20 (10.4)||14 (7.4)|
Adverse events such as myalgias, stomatitis (oral ulcers), alopecia, and fever were reported less frequently in patients receiving prasterone in comparison to placebo (Table 5). These differences were statistically significant for myalgias and oral ulcers, suggesting potential beneficial effects of prasterone treatment on some of the typical signs and symptoms of SLE.
Treatment-associated changes in laboratory values.
Total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, and total triglyceride levels decreased in the prasterone treatment group, although decreases in LDL cholesterol were minimal (Figure 2). Treatment-associated decreases in total cholesterol and HDL cholesterol were significantly greater in the prasterone group compared with the placebo group. Serum triglycerides were also statistically significantly decreased in the prasterone group compared with the placebo group, in which they were increased. Reductions in serum HDL cholesterol, total cholesterol, and triglyceride levels were evident in the prasterone treatment group by month 3, and the reductions were maintained with subsequent study drug administration (data not shown). HDL cholesterol levels decreased from normal levels at baseline in 133 and 148 patients in the prasterone and placebo groups, respectively, to below 40 mg/dl (the currently recommended lower limit for HDL cholesterol ) at the end of treatment in 38 patients receiving prasterone (26.6%) compared with 15 patients receiving placebo (10.1%) (P = 0.001 by chi-square test).
Mean serum levels of C3 complement declined in the prasterone treatment group: –7% from baseline values compared with –2% in the placebo group (P = 0.015). This decline was not associated with renal deterioration or with institution of new immunosuppressive treatment. There were minor decreases in C4 complement levels in both treatment groups, but the differences between groups were not significant. Mean ± SD changes in anti-dsDNA levels from the baseline visit to the last visit were 5.8 ± 145.6 IU/ml (median 0.0) in the placebo group and 20.0 ± 130.0 IU/ml (median 0.0) in the prasterone group; the differences between treatment groups were not statistically significant.
Treatment-associated changes in sex hormone levels.
As expected, serum DHEAS values increased to pharmacologic levels in the prasterone treatment group. The mean DHEAS concentrations at the last visit were 811 μg/dl (median 607) in the prasterone treatment group. These levels were unchanged in the placebo treatment group 120 μg/dl (median 50).
Serum testosterone levels increased in prasterone-treated patients, especially among those who were postmenopausal (Table 6). In contrast, treatment-associated changes in estrogenic hormones were less consistent. In premenopausal women, no meaningful changes in serum estradiol were evident in either the prasterone-treated patients or the placebo-treated patients (Table 6). Changes in estradiol levels in postmenopausal women were analyzed according to the presence or absence of hormone replacement therapy (HRT). Patients reported by investigators to be postmenopausal but with baseline estradiol levels ≥20 pg/ml were excluded from this analysis as being perimenopausal. Mean serum estradiol levels increased in postmenopausal patients receiving prasterone and HRT; however, the cotreatment with HRT makes the interpretation of these findings difficult. In postmenopausal patients who were not receiving HRT, prasterone led to increases in serum estradiol levels that were similar to the levels previously reported with low-dose HRT (33). There were no correlations between changes in sex hormone levels and responder outcomes (data not shown).
|Change from baseline||Treatment comparison|
|Testosterone, mean (median) ng/dl|
|Placebo (n = 63)||20.5 (16.0)||19.3 (16.0)||−1.2 (−2.0)||0.5541||0.0001|
|Prasterone (n = 71)||23.8 (19.0)||58.9 (56.0)||35.0 (32.0)||0.0001|
|Placebo (n = 68)||20.4 (17.0)||18.2 (12.5)||−2.1 (0.0)||0.1648||0.0001|
|Prasterone (n = 52)||17.9 (12.5)||74.8 (54.0)||57.0 (41.5)||0.0001|
|Estradiol, mean (median), pg/ml|
|Placebo (n = 63)||97.8 (74.2)||88.9 (62.8)||−8.9 (0.3)||0.5098||0.5884|
|Prasterone (n = 71)||85.6 (63.9)||86.5 (65.2)||1.0 (1.5)||0.9367|
|Postmenopausal patients (no HRT)†|
|Placebo (n = 14)||2.5 (1.4)||2.3 (1.4)||−0.1 (0.0)||0.8127||0.0003|
|Prasterone (n = 18)||3.7 (1.9)||24.8 (22.2)||21.1 (19.4)||0.0003|
|Postmenopausal patients taking HRT|
|Placebo (n = 35)||104.4 (64.7)||82.4 (66.5)||−22.0 (1.7)||0.2950||0.0210|
|Prasterone (n = 30)||85.6 (79.4)||128.1 (105.1)||42.4 (34.1)||0.0170|
In this double-blind, randomized, placebo-controlled trial of prasterone treatment at 200 mg/day for up to 52 weeks, we found significant improvements in the patients taking prasterone. Significantly more patients in the prasterone group than in the placebo group experienced either improvement or stabilization of disease activity (P = 0.017).
Given the complexity of SLE, it was important to assess the disease in its entirety, and this study was the first of its kind to utilize an innovative composite end point that was designed to integrate all 3 SLE domains—disease activity, organ damage, and health-related quality of life—into an overall “responder” end point. Patients categorized as “responders” had to exhibit simultaneous improvement or stabilization in each of 2 disease activity measures (SLEDAI and SLAM) and 2 quality of life measures (KFSS and patient's global assessment), without clinical deterioration.
To qualify for enrollment in this study, patients had to have stable disease activity at baseline, without recent changes in cotreatments, including glucocorticoids, antimalarials, and immunosuppressives. Cotreatments with these drugs were required to be held at a fixed dosage for the duration of the study The high response rate (45%) in the placebo treatment group should be interpreted in the context that this was not a true placebo-controlled trial, since most patients were cotreated with standard SLE therapies during the study. Thus, the statistically significant improvement in the prasterone group (59% taking prasterone versus 45% taking placebo were responders) is both statistically and clinically meaningful.
Deterioration in the mean postbaseline measures in any 1 of the scoring instruments relative to baseline or clinical deterioration caused a patient to be designated as a “nonresponder.” Significantly more patients in the placebo group as compared with the prasterone group failed to achieve responder status based on the SLEDAI or patient's global assessment values. The latter is particularly noteworthy, in that significant differences in patient's global assessments, in favor of prasterone, have also been reported in 2 other studies comparing prasterone with placebo (22, 24).
As noted above, the protocol was amended to incorporate baseline SLEDAI scores of >2 as an additional patient entry requirement while the double-blind study was ongoing. This requirement was based upon the outcome of an earlier Genelabs study, which revealed that patients with little or no disease activity (SLEDAI ≤2) are not the most appropriate candidates for clinical study, since patients with little or no disease activity are likely to exhibit high response rates (23). The importance of requiring minimum disease activity as a criterion for enrollment into SLE clinical trials is consistent with the approach taken in other clinical trials of other rheumatologic conditions.
Patients enrolled in this study represented a wide spectrum of SLE disease activity. The differences between treatment groups persisted with increasing baseline disease activity. This observation is similar to the findings in an earlier study assessing steroid-sparing properties of prasterone versus placebo, in which the greatest difference between placebo and prasterone treatment occurred in the groups with baseline SLEDAI scores >8 (23). The 2 instruments that provided the greatest sensitivity to responder analysis appeared to be the SLEDAI and the patient's global assessment.
There were no significant differences between treatment groups in the individual components of the responder analysis, which may reflect the fact that most of the patients were receiving cotreatment with standard SLE therapies. Furthermore, only patients who had had stable disease and no change in treatments for at least 6 weeks prior to entry into the study were eligible for enrollment, and as such, the population we studied primarily comprised those with active, yet stable disease. The responder analysis, however, required simultaneous stabilization or improvement across 4 variables (2 disease activity measurement instruments, 2 health-related quality of life measurements) and no clinical deterioration, while holding cotreatments constant. It is in this composite end point that significant differences occurred between treatment groups.
The responder end point was designed to assess simultaneous improvement or stabilization across all 3 domains of lupus: disease activity, organ damage, and health-related quality of life. As such, it was designed to assess a treatment effect on overall lupus disease. While it would be desirable to determine how many patients “improved” or how many “stabilized,” there is no established definition as to what constitutes improvement versus stabilization. Furthermore, due to the stringent responder criteria in this trial, each of the scoring instruments and clinical deterioration was given equal weight in the outcome. So, to improve in only some scoring instruments while worsening in others would deem a patient a “nonresponder.” Thus, we believe that overall stabilization, which is in fact reflective of a combination of improvement in some instruments and no worsening in others, is also a successful outcome for lupus patients.
Additionally, there were trends toward a lower number of patients with a first flare and a delay in time to disease flare among patients with active SLE who received treatment with prasterone. These findings are qualitatively similar to those reported in a population of Taiwanese women with SLE, almost all of whom were receiving cotreatment with glucocorticoids and/or immunosuppressives. In that study, there was a statistically significant delay in time to disease flare among patients treated with prasterone (24).
Administration of prasterone appeared to be well tolerated in this randomized clinical trial. Reported adverse events were predictably related to known pharmacologic effects of androgenic steroids and were generally mild and primarily confined to acne and, to a lesser extent, some hirsutism.
While prasterone is a precursor of both androgenic and estrogenic hormones, changes in estradiol levels in patients receiving prasterone were less consistent than changes in testosterone levels. In premenopausal patients, there were no differences in estradiol levels between treatment groups. In postmenopausal women receiving prasterone who were not also receiving HRT, there were modest increases in estradiol levels compared with placebo. During prasterone treatment, mean estradiol concentrations were similar to the mean serum estradiol concentrations observed during transdermal estradiol administration (33) and remained well under the pretreatment baseline estradiol levels observed in SLE patients who were receiving HRT at baseline. Reported adverse events related to estrogenic effects, such as menometrorrhagia, thrombotic events, or weight gain, were not increased in either premenopausal or postmenopausal women receiving prasterone, suggesting that the biologic effects of prasterone are primarily androgenic rather than estrogenic. This is in contrast to postmenopausal women, in whom unopposed estrogen therapy has been associated with a menorrhagia rate as high as 66% (34).
Clinical laboratory findings associated with prasterone treatment in this randomized controlled trial also reflected androgenic activity, including declines in HDL cholesterol and triglyceride levels. These findings have also been observed in 2 other controlled studies with prasterone (23, 24).
Reductions in triglyceride and HDL cholesterol levels with administration of androgenic hormones have been reported to be a manifestation of increased hepatic lipase activity, which results in enhanced clearance of HDL particles (35–37). Thus, the decreases in HDL cholesterol and triglyceride levels may represent increased reverse cholesterol transport (i.e., removal of cholesterol from peripheral tissues via enhanced HDL clearance) rather than decreased production of HDL (38). However, since lupus is associated with increased cardiovascular morbidity and mortality (39) and since long-term studies will be needed to further characterize these effects, it may be prudent to follow the National Cholesterol Education Program guidelines (32) while monitoring lipids in patients who are receiving prasterone.
Although the mechanism that results in decreased serum levels of complement C3 levels is incompletely defined, in vitro incubation of human peripheral blood mononuclear cells and bone marrow cells with DHEA has been demonstrated to reduce the production of IL-6 (16, 17). Since IL-6 levels are elevated in active SLE (20, 40) and can stimulate hepatic secretion of C3 as an acute-phase reactant (41), decreased levels of C3 during prasterone treatment may reflect either a direct or an indirect effect on hepatic C3 production. Decreased serum C3 complement levels were observed during administration of prasterone 200 mg/day to normal premenopausal women who were participating in a preclinical pharmacokinetic/pharmacodynamic study (Genelabs, Inc.: unpublished observations), which is consistent with a physiologic effect of prasterone on hepatic complement production. Serum complement levels also decline modestly without evidence of autoimmunity during testosterone replacement therapy in patients with Klinefelter's syndrome (42).
It is important to note this clinical trial enrolled only adult women with SLE. There are no data from randomized controlled studies of prasterone administration in men or in children with SLE. A previous open-label clinical trial in a small number of patients in which administration of a synthetic, more-potent androgenic steroid was examined suggested worsening of disease in men with SLE (43).
Finally, current therapeutic options are limited for patients with mild-to-moderate active SLE. The available drugs (e.g., nonsteroidal antiinflammatory drugs, antimalarials, glucocorticoids) can manage the disease temporarily in many patients; however, control of the underlying inflammatory disease is often incomplete, and many patients continue to have residual signs and symptoms, with fluctuations in disease activity and with disease flares. In these situations, it is often a major therapeutic step for the physician and patient to contemplate initiation of additional immunosuppressive agents or large doses of glucocorticoids, which also do not necessarily provide the hoped-for efficacy and are associated with significant side effects. Prasterone could bring benefits to patients who cannot or do not wish to take additional therapies of immunosuppressive agents or large doses of glucocorticoids. Prasterone is not intended, however, to be a replacement for glucocorticoids or immunosuppressive treatments that are needed during acute flares of lupus.
In summary, prasterone treatment improved or stabilized overall SLE disease activity in women with mild-to-moderate SLE. The most common adverse events associated with prasterone treatment were androgenic in nature and included acne and hirsutism, which were generally mild and treatable.
The authors greatly appreciate the advice of, and careful study monitoring by, Karen Colbert and Bettina Sporkenbach. ACRO, Inc. (Morris Plains, NJ) were the statistical consultants for the study.
- 1The clinical presentation of systemic lupus erythematosus. In: WallaceDJ, HahnBH, QuismorioFPJr, KlinenbergJR, editors. Dubois' lupus erythematosus. 5th ed. Baltimore: William & Wilkins; 1997. p. 627–33..
- 16Serum dehydroepiandrosterone (DHEA) and DHEA sulfate are negatively correlated with serum interleukin-6 (IL-6), and DHEA inhibits IL-6 secretion from mononuclear cells in man in vitro: possible link between endocrinosenescence and immunosenescence. J Clin Endocrinol Metab 1998; 83: 2012–7., , , , , , et al.
- 23GL701 Study Group. Effects of prasterone on corticosteroid requirements of women with systemic lupus erythematosus: a double-blind, randomized, placebo-controlled trial. Arthritis Rheum 2002; 46: 1820–9., , , , , , et al, for the
- 32Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001; 285: 2486–97.
- 40Altered function of the hypothalamic stress axes in patients with moderately active systemic lupus erythematosus. II. Dissociation between androstenedione, cortisol, or dehydroepiandrosterone and interleukin 6 or tumor necrosis factor. J Rheumatol 2000; 27: 911–8., , , , , .
- 41Hormonal regulation of complement biosynthesis in human cell lines. II. Upregulation of the biosynthesis of complement components C3, factor B and C1 inhibitor by interleukin-6 and interleukin-1 in human hepatoma cell line. Mol Immunol 1990; 27: 197–201., , , , , .