Effect of Nandrolone Decanoate Therapy on Weight and Lean Body Mass in HIV-Infected Women With Weight Loss: A Randomized, Double-blind, Placebo-Controlled, Multicenter Trial

Background  Weight loss is associated with accelerated mortality and disease progression in patients with human immunodeficiency virus (HIV) infection. Although studies have examined a variety of anabolic therapies in HIV-infected men, the safety and efficacy of such treatments in women have not been adequately studied.

Methods  In this randomized, double-blind, placebo-controlled, multicenter, phase I/II study, 38 HIV-infected women with documented weight loss of 5% or greater in the preceding year or a body mass index of less than 20 kg/m² were randomized to receive nandrolone decanoate (100 mg) or an equivalent volume of placebo every other week by intramuscular injection. Subjects received blinded treatment for 12 weeks, followed by open-label therapy for 12 weeks. Lean body mass and fat (bioelectrical impedance analysis) and weight were measured at baseline and at weeks 6, 12, 18, and 24. Biochemical assessments of safety (hematologic analyses, liver function tests, and sex hormone measurements) were performed at these same time points. Clinical signs and symptoms were monitored biweekly.

Results  Subjects randomized to receive nandrolone had significant increases in weight and lean body mass during blinded treatment (4.6 kg [9.0%] and 3.5 kg [8.6%], respectively; P<.001 vs baseline and placebo in each case). Fat mass did not change statistically significantly in either group. Although there were no statistically significant differences between groups in biochemical measures, the number of grade 3 or greater toxicities, or reports of virilizing effects, a full assessment of safety cannot be made in a trial of this size.

Conclusion  Nandrolone decanoate therapy may prove to be generally safe and beneficial in reversing weight loss and lean tissue loss in women with HIV infection and other chronic catabolic diseases.

W asting has been associated with increased risk of mortality and morbidity in patients with human immunodeficiency virus (HIV) infection, even in the current treatment era.^1-4 Although the incidence of HIV-associated wasting has declined in populations with widespread access to highly active antiretroviral therapy (HAART),^5-7 wasting persists in some patients undergoing HAART.⁸ Moreover, weight and lean body mass (LBM) are not consistently restored after the initiation of HAART.^9,10 Although studies have examined the safety and efficacy of a variety of anabolic therapies in men, treatment options have not been adequately studied in women.

Nandrolone decanoate is an injectable derivative of 19-nortestosterone that is approved by the Food and Drug Administration as a treatment for anemia in men and women with chronic renal failure. Treatment with nandrolone has been associated with increases in LBM in men and women with chronic renal failure¹¹ and in postmenopausal women.^12,13 Despite considerable off-label use of nandrolone in patients with HIV infection, limited data are available regarding its safety and efficacy in HIV-positive men, and none are available in HIV-positive women. In open-label studies^14-17 in HIV-infected men, nandrolone use has been associated with increases in weight and LBM. We report herein the first study of nandrolone dec-anoate therapy in HIV-infected women with weight loss.

Thirty-eight HIV-infected women with documented involuntary weight loss of 5% or more during the preceding 12 months or a body mass index (calculated as weight in kilograms divided by the square of height in meters) of less than 20 were recruited at 9 clinical sites between January 1997 and October 1999. Subjects who used antiretroviral drugs were required to be following stable, optimized regimens as determined by their primary care providers for 30 days or more preceding enrollment and were asked to plan to remain doing so during the study. Subjects who had not received antiretroviral drugs during the 30 days preceding enrollment and who had no plans to initiate therapy during the study were also eligible. All subjects of childbearing potential were required to have a negative pregnancy test (serum β-human chorionic gonadotropin) result at screening, to have used adequate contraception for at least 30 days before study entry, and to avoid pregnancy while participating in the study and for 3 months after the completion of treatment. Other eligibility requirements included a Karnofsky score of 60 or greater, a life expectancy of 6 months or longer, and a self-reported daily energy intake of 80% of estimated requirements or greater, as determined by diet history.

The exclusion criteria consisted of an opportunistic infection or other major systemic illness in the preceding 30 days; a platelet count less than 50×10³ μL; aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase levels more than 5 times the upper limit of normal (ULN); a direct bilirubin level greater than 1.5 times ULN; a prothrombin time/partial thromboplastin time greater than 1.5 times ULN; receipt of immune modulators, megestrol acetate, anabolic steroids, systemic glucocorticoids, growth hormone, anticoagulants, or hypoglycemic agents within 30 days of entry; a history of hypersensitivity to nandrolone decanoate; breastfeeding; the presence of a persistent, palpable, dominant breast mass of undetermined etiology; any history of invasive cervical cancer; high-grade squamous intraepithelial lesions (SIL) or cervical intraepithelial neoplasia grade II or worse; diabetes mellitus; any condition that would require anticoagulation therapy during the study; a history of cardiomyopathy or congestive heart failure; impaired oral intake; grade 2 or greater intractable nausea or vomiting; persistent diarrhea (4 or more stools daily); use of total or partial parenteral nutrition or tube feeding; weight gain of 3% or more within 4 weeks of entry; and current systemic chemotherapy or radiotherapy for B-cell lymphoma, malignancies, or Kaposi sarcoma. Subjects who used oral nutritional supplements, dronabinol, cyproheptadine hydrochloride, or pentoxifylline could enroll. The study was approved by the institutional review board at each clinical site, and signed informed consent was obtained from all subjects before screening.

Subjects were randomly assigned to receive 100 mg of nandrolone decanoate (Deca-Durabolin) or an equivalent volume of placebo (both donated by Organon USA Inc, West Orange, NJ) by deep intramuscular injection every 2 weeks for 12 weeks. This dose of nandrolone is half the maximum dose approved by the Food and Drug Administration for women with anemia associated with chronic renal failure. Nandrolone and placebo were provided in sterile sesame oil solutions that contained 10% benzyl alcohol as a preservative. All subjects, study personnel, and investigators were blinded to group assignments. Injections were administered by study personnel. Subjects who completed the 12-week placebo-controlled phase were eligible to receive open-label nandrolone under the same dosing regimen for the subsequent 12 weeks. Treatment was discontinued after 24 weeks.

The protocol stipulated that use of the study medication would be discontinued if a subject developed any new malignancy other than cutaneous Kaposi sarcoma, hypertension, or peliosis hepatis; had a Papanicolaou smear or a cervical biopsy specimen with high-grade SIL or cervical intraepithelial neoplasia grade II or worse results; became pregnant; or developed any grade 4 toxicity (defined by the Division of Acquired Immunodeficiency Syndrome Table for Grading Severity of Adult Adverse Experiences) related to study medication. Subjects with progressive weight loss during the study (≥5% from the baseline value) could be withdrawn and referred for treatment. The amount of study medication used could be reduced by half at the discretion of the subject or the investigator if the subjects experienced symptoms of virilization, mood changes, changes in libido, menstrual irregularities, or clitoral enlargement.

Efficacy and safety were assessed at baseline and at weeks 6, 12, 18, and 24. Subjects were asked to fast for 8 hours or more before these study visits. Height was measured before randomization, following a standardized procedure. Weight was measured on the same calibrated scales at each visit, with subjects wearing a hospital gown, underwear, and socks. Body composition was measured by using single-frequency bioelectrical impedance analysis (RJL Quantum, Clinton Township, Mich). Electrode placement was standardized during central training of study coordinators. Fat mass and LBM were calculated using published equations that have been validated in HIV-negative and HIV-positive individuals.¹⁸ Fasting blood samples were collected for measurement of chemistry values, liver function tests, hematologic analyses, and CD4 lymphocyte counts. Serum samples were stored for subsequent batch analysis of sex hormones. Three-day written food intake diaries were reviewed by a dietitian, and energy and macronutrient intake were calculated. Subjects also completed written quality-of-life and physical activity questionnaires, with the latter used to provide semiquantitative information on activity patterns. Clinical signs and symptoms were evaluated biweekly, when subjects received their injections of study medication. Pregnancy testing was performed at 6-week intervals. Follow-up cervical and breast examinations were performed at weeks 12 and 24.

Routine laboratory chemistries, hematologic analyses, and CD4 lymphocyte counts were measured in clinical laboratories for each site. Estradiol levels were measured by radioimmunoassay, and luteinizing hormone and follicle-stimulating hormone levels were measured by immunoradiometric assays (Coat-a-Count; DPC Inc, Los Angeles) on stored, batched serum samples in the Core Hormone Laboratory of the General Clinical Research Center at San Francisco General Hospital.

The sample size was determined by using change in LBM as the primary outcome. The protocol team considered an increase in LBM of 2.5 kg to be the minimum physiologically significant improvement that would offset the potential toxicity of nandrolone. Using an SD of 2.5 kg, derived from studies of nandrolone therapy in postmenopausal women¹² and growth hormone use in HIV-infected individuals,¹⁹ it was determined that 16 subjects per group would provide 80% power to detect a significant effect, using a 2-sided P = .05. With an anticipated dropout rate of 20%, the sample size was set at 19 subjects per group.

The primary analysis compared changes during blinded treatment (baseline to week 12) in the 2 arms in an intent-to-treat manner. Wilcoxon rank sum tests were used to assess between-group differences in continuous variables; differences between groups in categorical variables were assessed using exact testing procedures. Within-group changes were assessed using Wilcoxon signed rank tests. In the case of missing data, the last available value was not carried forward. Values for reported changes are in subjects for whom paired data are available.

Nineteen women were randomized to each treatment group (Figure 1). Subjects in both groups were predominantly nonwhite (Table 1). There were no significant differences between groups in the number of women undergoing antiretroviral therapy at baseline, and approximately half of the women in each group were following protease inhibitor–containing regimens. Seven women (3 in the nandrolone group and 4 in the placebo group) were receiving hormone replacement therapy. Only 1 woman reported active intravenous drug use at baseline, but most women in both groups (14 in the nandrolone group and 18 in the placebo group) had never used intravenous drugs. Mild cervical dysplasia was noted in 4 women at baseline (3 in the nandrolone group and 1 in the placebo group). The CD4 lymphocyte count was less than 200 cells/μL in 13 women (6 in the nandrolone group and 7 in the placebo group). There were no statistically significant differences between groups in these or any body composition or nutritional variables at baseline (Table 2). A total of 33 women (87%) completed the blinded treatment phase (16 in the nandrolone group and 17 in the placebo group); 31 (82%) completed the open-label portion of the study (Figure 1).

During blinded treatment, subjects randomized to receive nandrolone had significant increases in weight and LBM (4.6 kg [9.0%] and 3.5 kg [8.6%] respectively; P < .001 vs baseline and placebo in each case) (Figure 2). In contrast, weight and LBM did not change statistically significantly in the placebo group. Fat content did not change statistically significantly in either group. During open-label therapy, subjects originally randomized to receive placebo gained LBM (1.5 kg; P = .007 vs week 12) (Figure 2). Weight gain in this group approached, but did not achieve, a level of statistical significance (2.2 kg; P = .10 vs week 12). The women who were originally randomized to receive nandrolone tended to sustain the gains in weight and LBM that were achieved during the blinded treatment phase, but they did not experience statistically significant further gains with extended treatment. These increases in weight and LBM were not associated with any statistically significant differences in self-reported energy or macronutrient intake, quality of life, or activity levels at baseline or changes during blinded treatment (data not shown).

Hemoglobin and hematocrit values tended to increase with nandrolone therapy (Table 3), with the increase in hematocrit level approaching statistical significance (P = .054 vs placebo). During open-label therapy, hematocrit levels increased in both groups. Although there was a wide range of CD4 lymphocyte counts at baseline, there were no statistically significant changes in either group during the study (Table 3).

There were no statistically significant changes or trends in alanine aminotransferase, aspartate aminotransferase, and alkaline phosphatase levels during blinded treatment or open-label therapy, apart from a decrease in alkaline phosphatase levels during open-label therapy in subjects originally randomized to the placebo group (Table 4). Estradiol levels tended to decrease progressively in subjects randomized to receive nandrolone during blinded treatment and subsequent open-label therapy, and luteinizing hormone levels tended to decrease. Fasting glucose, triglyceride, and total cholesterol levels were unaffected by nandrolone therapy.

More than half of the women in each treatment group experienced at least one grade 3 or 4 toxicity, but there were no statistically significant differences between groups in overall or specific toxicity rates, including sex-specific toxic effects (Table 5). There were no reports of liver toxicity, consistent with the absence of changes in liver function test results reported in Table 3. There were no statistically significant differences between groups in the number of reports of total or specific sex-related toxicities of any grade (Table 6), although there is a noticeable elevation in the number of such events in the nandrolone-treated women during open-label therapy (10 vs 3 in those originally assigned to receive placebo). However, many of these events occurred in the same women.

Study medication use was discontinued during blinded treatment in 5 women (3 in the nandrolone group and 2 in the placebo group). The reasons for treatment discontinuation were noncompliance with study visits (2 in the nandrolone group and 1 in the placebo group) or the discovery of cervical intraepithelial neoplasia grade II or greater (1 in each group). There was 1 death during open-label therapy in a woman originally randomized to receive placebo. The cause of death was unrelated to use of the study medication.

Of 36 subjects with evaluable baseline Papanicolaou smears, 28 had findings within the reference range or benign cellular changes. The remaining results included low-grade SIL (n = 4), atypical squamous cells of undetermined significance (n = 2), and inflammatory smears (n = 2). Three women randomized to receive nandrolone had postentry evaluations that revealed high-grade SIL. One of these subjects did not have a baseline evaluation; her first Papanicolaou smear at week 12 revealed high-grade SIL, and a second examination performed 34 weeks after enrollment (off treatment) showed low-grade SIL. The second patient had normal smear results at entry, atypical squamous cells of undetermined significance at week 12, and high-grade SIL 36 weeks after enrollment. The third woman had low-grade SIL at baseline and week 12 and subsequently progressed to high-grade SIL in an examination performed 31 weeks after enrollment (off treatment). Three women randomized to receive nandrolone therapy who had abnormal smear findings at baseline had subsequent findings that were within the reference range or benign cellular changes. One woman randomized to the nandrolone group had abnormal follow-up breast examination findings at week 24. Three women randomized to the placebo group who had findings within the reference range or benign cellular changes at baseline progressed to low-grade SIL (n = 2) or atypical squamous cells of undetermined significance (n = 1) at week 12. Overall, there was evidence of progression and reversion in women randomized to receive nandrolone and of progression in women randomized to receive placebo, with no statistically significant differences between groups in progression rates.

To our knowledge, this is the first randomized, double-blinded, placebo-controlled study of a therapy that statistically significantly increases weight and LBM in HIV-infected women with a history of weight loss. Treatment with nandrolone produced increases of 9.0% and 8.6% in weight and LBM, respectively, in 12 weeks. The increases in weight and LBM with nandrolone use occurred with no statistically significant change in fat mass. Studies in both men²⁰ and women²¹ have shown that wasting in HIV infection is associated with the depletion of both fat and lean stores. Thus, by producing a robust increase in LBM coupled with no further loss of fat, nandrolone could be a particularly beneficial therapy for HIV-associated wasting. Although drug costs vary worldwide, a further advantage of nandrolone is that it is relatively inexpensive and, thus, may be more financially accessible than other treatments for wasting.

A variety of pharmacologic therapies for HIV-associated wasting have been investigated in randomized, double-blind, placebo-controlled studies.^19,22-27 Although many of these studies did not specifically exclude women, enrollment of women was generally minimal and precluded subgroup analyses. The present study, which recruited women exclusively, demonstrated that women with HIV-associated wasting have considerable anabolic potential. Indeed, the magnitude of the increases in weight and LBM in this trial exceeds gains seen in trials of other anabolic therapies. Although the HIV-infected women studied were primarily nonwhite, we have no reason to expect that this would limit the generalizability of our results.

It should be emphasized that we studied nandrolone as a pharmacologic intervention for wasting in women. Some HIV-infected women have been shown to have low testosterone levels,^21,28,29 and the effects of physiologic testosterone replacement therapy have been investigated in these women.^30,31 In 1 such study,³¹ hypogonadal women randomized to receive testosterone replacement therapy experienced significant increases in 4 of 8 tests of muscle strength. Muscle mass, estimated from urinary creatinine excretion, tended to increase (P = .08), but there was no significant increase in weight. Overall, testosterone replacement in hypogonadal women warrants further investigation.

Our sample size does not allow us to perform a full assessment of the safety of nandrolone. However, given the potential for abuse of nandrolone and other anabolic/androgenic agents, we present in this article detailed information on the effects of nandrolone on biochemical measures and the results of cervical and breast examinations and sex-specific adverse effects. There were no grade 3 or 4 elevations in transaminase or bilirubin levels or statistically significant changes in the median values for these variables in either group. Overall, injectable anabolic steroids have been associated with less hepatotoxicity than oral formulations such as oxymetholone and oxandrolone.^25,32 Although there was no change in total cholesterol levels in the present study, nandrolone and other synthetic anabolic steroids have been consistently shown to reduce high-density lipoprotein cholesterol levels.^11-13,32-34 Although our results show no evidence of a deleterious effect of treatment on CD4 lymphocyte counts, the effects of nandrolone therapy on HIV viral load are not known.

There tended to be a higher rate of sex-related side effects in women originally randomized to receive nandrolone, who were exposed to active treatment for up to 24 weeks. Although this trend did not achieve statistical significance with this relatively small sample size, women treated with nandrolone should be monitored for signs of virilizing effects. They should undergo cervical screening examinations with appropriate follow-up as recommended for all HIV-infected women by other published guidelines.³⁵

A limitation of the present study is that bioelectrical impedance analysis was used to measure changes in total LBM and fat mass. However, we are confident that our data were obtained under optimized conditions for that technique. Height and weight, which can contribute to the variability in LBM and fat calculated by bioelectrical impedance analysis, were measured on calibrated devices, using standardized procedures. Study coordinators underwent supervised central training for measuring height and weight and for standardized electrode placement. All measurements were made under fasting conditions to minimize the potential effects of food and fluid intake. Finally, LBM and fat mass were calculated using published equations that had been developed and validated in a diverse population that included patients with HIV infection.¹⁸

In summary, the results of the present study demonstrate that treatment with nandrolone can produce robust increases in weight and LBM, with no decrease in fat mass, in women with HIV-associated weight loss. Nandrolone therapy was not associated with statistically significant toxicities, although the study was too small for a full assessment of its safety. The long-term impact of reversing weight loss on morbidity and mortality in patients with HIV infection is not known. These caveats notwithstanding, nandrolone therapy may prove to be a beneficial means of increasing weight and LBM in women with HIV infection and other chronic catabolic diseases.

Correspondence: Kathleen Mulligan, PhD, Division of Endocrinology, San Francisco General Hospital, 1001 Potrero Ave, Bldg 30, Room 3501K, San Francisco, CA 94110 (kmulligan@sfghgcrc.ucsf.edu)

Financial Disclosure: None.

Funding/Support: This study was supported in part by grants AI27663, AI38855, and M01-430 from the National Institutes of Health. Hormone analyses were supported by grant 5-M01-RR-00083 and were performed in the Core Hormone Laboratory of the General Clinical Research Center at San Francisco General Hospital.

Acknowledgment: We acknowledge the contributions of the other members of the AIDS Clinical Trials Group 329 protocol team (listed alphabetically): Michael T. Basar; Stacy A. Bradley; Margaret Chesney, PhD; Margaret Nettles, PhD; Sharon Shriver, RPh; and Stacey Welch, MA. We thank the following AIDS Clinical Trials Group sites for enrolling subjects in this study: Medical Center of Louisiana at New Orleans; City Hospital, San Juan, Puerto Rico; Los Angeles County/University of Southern California Medical Center; University of Puerto Rico, San Juan; University of California, San Francisco–San Francisco General Hospital; Duke University, Durham, NC; Northwestern University, Chicago, Ill; Howard University, Washington, DC; and University of Hawaii at Manoa, Leahi Hospital, Honolulu. We also thank Barbara Chang of the General Clinical Research Center at San Francisco General Hospital for performing the hormone assays.

Acknowledgment: We grieve for our friend and colleague, Robert Zackin, ScD, whose contributions to all aspects of this study were vital to its success.