Causes of cachexia may include but are not limited to secondary infection [3], fever [3], malabsorption [4, 5], diarrhea [6], and disease-induced alterations in cytokine concentrations [7] and thyroid hormones [8] or disease-induced anorexia [9]. Given the many potential identifiable and unidentifiable causative factors, reversal of AIDS wasting is difficult. If the causes of wasting are identified and treated, body weight may stabilize and increase [10]. When weight loss is severe, however, depression, anorexia, and the accompanying loss of physical function can make it difficult to increase caloric intake without parenteral or enteral nutrition support or appetite stimulants even for the most motivated patients.

In high doses (480 to 1600 mg/d), the synthetic progesterone megestrol acetate (17 -acetyloxy-6-methylpregna-4,6-diene-3,20-dione) has been shown to stimulate appetite, improve the sense of well-being, and increase weight in most patients with breast cancer [11]. In patients with various malignancies, megestrol acetate treatment has been shown to improve appetite in most patients and to promote weight gain of 2.27 kg (5 pounds) or more in approximately 25% of patients [12-16]. Fourteen patients with HIV infection who had cachexia were treated with 320 mg of megestrol acetate per day by Von Roenn and colleagues [17], and they noted an average weight gain of 0.5 kg per week. Others have reported that most patients with HIV infection and cachexia gained weight after treatment with megestrol acetate [18]. Although megestrol acetate therapy can increase the body weight of patients with AIDS who have cachexia, the number of 40-mg tablets required to increase appetite may be intolerable. In this study, we tested the acceptability and effects of a 20-mL dose of a lemon-lime flavored suspension that contained 800 mg of megestrol acetate.

This multicenter, double-blind, placebo-controlled clinical trial compared the effects of high-dose megestrol acetate and placebo on the medical status of patients with AIDS who had cachexia and anorexia. Specifically, we compared the effects of megestrol acetate and placebo on nutritional status variables, such as caloric intake, body weight, body mass index and body composition, and the patient's subjective sense of well-being.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Eligible patients fulfilled the 1987 criteria from the Centers for Disease Control for AIDS, were 18 years of age or older, had documented severe weight loss, had a projected life expectancy of 20 weeks or more, had a Karnofsky performance status of 50 or more, and considered weight loss to be a detriment to their well-being. Severe weight loss was defined as follows: For patients whose premorbid body weight was similar to ideal body weight, a minimum weight loss to 10% below ideal body weight was required; for patients whose premorbid body weight was greater than ideal body weight, a documented decrease of 20% from usual body weight was required; and for patients whose premorbid weight was less than ideal body weight, a documented loss of 10% or more of usual body weight was required. Patients treated with zidovudine had to have been on a stable dose for 8 weeks before study entry. Patients were ineligible if they were pregnant or menstruating or had a history of poorly controlled hypertension, congestive heart failure, or deep venous thrombophlebitis. Other reasons for ineligibility included functional obstruction to food intake, apparent digestive or absorptive impairment, dementia or mental incompetence, uncontrolled diarrhea, ascites, pleural effusions, active opportunistic infection, exacerbation of established opportunistic infections, malignancy associated with weight loss, or a history of substance abuse with questionable future abstinence. These exclusion criteria were developed to purposefully eliminate patients with identifiable causes of wasting. Conditions for removal from study treatments included body weight loss of 10% or more of baseline body weight, possible life-threatening drug-related toxicity, interruption of therapy for longer than 2 weeks, subjective excessive weight gain, decision of physician investigator, or initiation of a treatment that might substantially affect appetite, body weight gain, quality of life, or HIV replication.

Evaluation of Patients and Follow-up

The study was approved by the institutional review board at each participating center. During the baseline period (week 0), the study was explained to patients and signed informed consent was obtained. Baseline body weight, height, and body frame size (as estimated by elbow width) were measured and ideal body weight was determined. Ideal body weight was considered the mean of the ideal body weight range for frame size according to the 1980 Metropolitan Height and Weight Tables [19]. A complete medical history and physical examination were done, and baseline Karnofsky status was established. If, on medical examination, no exclusionary criteria were identified, patients were enrolled and randomized with a blocking factor of four in each clinic.

All patients received bottles containing 20-mL portions of a liquid, lemon-lime flavored suspension of either placebo or 800 mg of megestrol acetate (Megace; Bristol-Myers Squibb, Princeton, New Jersey). Patients were instructed to take the drug orally once daily, 1 hour before or 2 hours after breakfast. Patients and clinicians were blinded to treatment groups throughout the study.

Patients were evaluated every 4 weeks for up to 12 weeks. Monthly measurements included body weight, mid-arm circumference, tricep skinfold, assessment of body composition, caloric intake, Karnofsky performance, appetite and patient well-being self-assessment surveys, medical assessment, and laboratory variables. A detailed description of these measurements is described below.

At each monthly visit, patients had a physical examination and a review of symptoms for edema, HIV-related constitutional manifestations, and indications of new or worsening opportunistic infections or malignancies. Patients with symptoms or signs suggestive of new or worsening infection or malignancy were referred to the study physician or physician assistant for further evaluation.

Monthly laboratory variables evaluated at each site included hemoglobin, hematocrit, leukocyte differential, and platelet counts; blood urea nitrogen, creatinine, electrolytes, and glucose concentrations; liver tests; and CD4⁺, CD8⁺, and total lymphocyte counts. Clinically significant changes in laboratory variables were not observed; therefore, these data are not shown. Skin antigen reactivity tests to Trichophyton, Proteus, glycerin, tetanus, diphtheria, Streptococcus, tuberculin, and Candida antigens were done at baseline, and at 12 weeks or last evaluation using the Multitest CMI test kit (Merieux Institute, Inc., Miami, Florida).

Monthly assessment of body weight gain and body composition was evaluated as described below. Patients were weighed and their height measured while dressed in ordinary clothing without shoes or coats. Body mass index, an index of adiposity, was calculated using the equation: weight/height² (kg/m²) [20]. Body composition was analyzed by bioelectrical impedance and anthropometry. Bioelectrical impedance analysis (BODYCOMP II—Version 1.5; RJL Systems, Inc., Detroit, Michigan) of body composition was done on patients in the supine position after all metals (jewelry, watches, keys, coins, belt-buckles, etc.) had been removed. For the anthropometric analyses, mid-arm circumference (MAC [cm]) and tricep skin-fold thickness (TSF [mm]) were measured on the right arm. Using the mid-arm circumference and tricep skinfold measurements, total body muscle and fat mass were estimated. Muscle mass was evaluated by calculation of mid-arm muscle circumference (MAMC) and mid-arm muscle area (AMA). Mid-arm muscle circumference (cm) was calculated from the equation: MAMC = (MAC –[ x TSF]), where all measurements are in centimeters. Mid-arm muscle area (mm²) was calculated from the equation: AMA = ([MAC –( x TSF)]^2/4 where all units are in millimeters. Total body fat mass was estimated by mid-arm fat area (mm²) AFA = ([(TSF x MAC)/2] -( x [TSF]²)/4]) where all units are in millimeters [20]).

The study participants were instructed by a dietitian to complete food intake diaries on three consecutive weekdays during the week before each monthly visit. To estimate the amount of food eaten, the use of a scale, measuring cups, and spoons was encouraged. Caloric intake was analyzed by Computrition, Inc. nutritional analysis software (1981, Chatsworth, California).

The patients' perceptions of well-being were assessed using nine linear analog scale questions about the effect of treatment on personal appearance, appetite, weight, overall health, benefits of treatment, and quality of life (Appendix Table). The nine questions were designed to evaluate the patients' perceived change in well-being during the course of therapy. Scores were transformed so that a positive number represented a positive response and a negative number represented a negative response. The highest possible score was 5 and the lowest was –5.

Statistical Analysis

Because severely ill patients were selected, patients who met all entry criteria were difficult to find and 42 patients completed the study. Patients who completed less than 4 weeks were excluded from all analyses (9 patients receiving placebo and 6 treated with megestrol acetate). Twenty-two patients were randomly assigned who did not meet the strict weight loss entry criteria; 9 of these patients were excluded from all analyses (6 patients receiving placebo and 3 treated with megestrol acetate). Thirteen of these 22 patients were within 1.8 kg of meeting the weight loss criteria and each of these patients was severely wasted. Data analyses were completed with and without these 13 patients. Data presented in the tables include these 13 patients; findings without these 13 patients for major outcome variables are presented in the Results section and show that inclusion of these patients did not change the conclusions.

Because of the drop-out rate and the inability to do follow-up measurements on those that dropped out, we analyzed the data using analysis of covariance. For the analysis of covariance, the general linear models procedure was used, with the baseline value of each dependent variable, group, and baseline x group interaction as covariates [21]. Baseline x group interactions were not observed, and this term was dropped from the model. Site effects could not be estimated because four clinical sites each had fewer than four patients. Plots of residual and predicted values showed no departure from homogeneity of variance for any outcome, and all analyses were done on untransformed data. Baseline results are expressed as mean ±SE; group effects are presented as the mean change from baseline for weeks 4, 8, and 12; the P values referenced in tables represent the P value for group differences as assessed by analysis of covariance.

In addition to the analysis of covariance, point estimates of the difference and their 95% CIs were calculated. The CIs in the text and tables estimate the differences in mean changes from baseline to weeks 4, 8, and 12 in the two groups; CIs are not adjusted for multiple comparisons. The number of patients in the treatment groups varies from variable to variable because of missing measurements. Pearson correlation coefficients were calculated to examine the relation between several dependent variables. The Fisher exact test was used to determine if the frequency of several categorical variables differed between treatment groups. Survival curves were estimated by the Kaplan-Meier method for censored data and were compared by the log-rank test. SAS software [22] was used for all analyses, and P values 0.05 were considered significant.

Results

Top Methods Results Discussion Author & Article Info References

Baseline Characteristics

Between 27 May 1989 and 25 April 1991, 100 patients at 13 centers (range, 1 to 28 patients per center) were randomly assigned to receive either placebo (n = 48) or megestrol acetate (n = 52). Selected patient characteristics, neoplasms, and opportunistic infections present at baseline are shown in Table 1. Patient characteristics were similar at baseline for all variables except the proportion with Kaposi sarcoma, which was higher in the patients who received placebo than in those treated with megestrol acetate. The patients ranged in age from 24 to 70 years and were predominately men. Ninety-five patients were taking zidovudine.

Seventy-eight patients completed the study to week 4; 64 patients completed the study to week 8; and 42 patients (20 treated with placebo and 22 treated with megestrol acetate) completed the full 12-week study. Table 2 lists the reasons for study discontinuation. Two patients in the megestrol acetate group withdrew because of excessive weight gain, and 6 patients in the placebo group withdrew because of insufficient weight gain. Three patients treated with megestrol acetate withdrew because of gastrointestinal disturbance that included nausea, vomiting, flatulence, and diarrhea. One patient treated with megestrol acetate had dyspnea and withdrew. One patient had a diffuse urticarial rash because of an allergic reaction to megestrol acetate and was immediately withdrawn from the study. One patient treated with megestrol acetate had potential "life-threatening toxicity" (defined as an event with immediate risk for death); this patient had a suspected thromboembolic event in the setting of Staphylococcus bacteremia and HIV-related dementia. One patient receiving placebo died of probable stroke induced by HIV-related encephalopathy. Four patients treated with megestrol acetate died of the following causes: Compound Q toxicity, cardiac arrest during seizure (Pneumocystis carinii pneumonia and central nervous system disease were present in this patient as well), pulmonary Kaposi sarcoma, and sepsis. The total number of patients discontinuing therapy prematurely was similar in the treatment groups.

Caloric Intake

The daily caloric intake at weeks 4 and week 8 relative to baseline was greater in patients treated with megestrol acetate than in those who received placebo (Table 3). By week 12, the placebo group decreased their daily caloric intake by 241 calories from baseline, whereas the megestrol acetate group increased their intake by 447 calories [difference, 688 calories; CI, –144 to 1521 calories]. When data were reanalyzed excluding the closely eligible 13 patients, daily caloric intakes were similarly increased in the megestrol acetate group. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for daily caloric intake were as follows: 1) week 4: difference, 568 calories [CI, 110 to 1026 calories]; 2) week 8: difference, 492 calories [CI, 27 to 958 calories]; and 3) week 12: difference, 731 calories (CI, –236 to 1697 calories).

The calories consumed per kg at weeks 4 and 8 relative to baseline were greater in patients treated with megestrol acetate than in patients who received placebo (Table 3). By week 12, the placebo group decreased the calories consumed per kg by 2.82 from baseline, whereas the megestrol acetate group increased calories consumed per kg by 8.85 [difference, 11.66 calories per kg; CI, –2.45 to 25.77 calories per kg]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for calories consumed per kg were as follows: 1) week 4: difference, 7.45 calories per kg [CI, 0.02 to 14.88 calories per kg]; 2) week 8: difference, 6.31 calories per kg [CI, –1.68 to 14.31 calories per kg] and 3) week 12: difference, 13.56 calories per kg (CI, –3.01 to 30.14 calories per kg).

The megestrol acetate group increased the intake (g) of protein consumed per kg from baseline at weeks 4, 8, and 12 when compared with the placebo group (Table 3). For example, at week 12, the placebo group decreased their intake of protein per kg by 0.20 g from baseline, whereas the megestrol acetate group increased by 0.54 g protein per kg [difference, 0.74 g protein per kg; CI, 0.05 to 1.44 g protein per kg]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for protein intake per kg were as follows: 1) week 4: difference, 0.41 g protein/kg [CI, 0.09 to 0.74 g protein/kg]; 2) week 8: difference, 0.40 g protein/kg [CI, 0.002 to 0.79 g protein/kg]; and 3) week 12: difference, 0.88 g protein/kg (CI, 0.07 to 1.69 g protein/kg).

Body Weight and Body Mass Index

Body weight change from baseline at week 4 was similar in the treatment groups (difference, 1.36 kg; CI, –0.27 to 2.99 kg) (Table 4). From baseline to week 8, body weight increased 3.86 kg in the megestrol acetate group, whereas body weight decreased 0.46 kg in the placebo group (difference, 4.32 kg; CI, 2.42 to 6.22 kg). Similarly, from baseline to week 12, body weight increased 4.16 kg in the megestrol acetate group and decreased 0.61 kg in the placebo group [difference, 4.77 kg; CI, 2.06 to 7.48 kg]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for body weight were as follows: 1) week 4: difference, 0.66 kg [CI, –1.04 to 2.37 kg]; 2) week 8: difference, 3.85 kg [CI, 1.86 to 5.85 kg]; and 3) week 12: difference, 5.73 kg (CI, 2.76 to 8.69 kg).

Body mass index change from baseline to week 4 was similar in the treatment groups (difference, 0.47 kg/m²; CI, –0.05 to 0.99 kg/m²) (Table 4). From baseline to week 12, body mass index increased by 1.36 kg/m (²) in the megestrol acetate group and decreased by 0.21 kg/m² in the placebo group [difference, 1.56; CI, 0.71 to 2.42 kg/m²]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for body mass index were as follows: 1) week 4: difference, 0.22 kg/m² [CI, –0.31 to 0.76 kg/m²]; 2) week 8: difference, 1.23 kg/m² [CI, 0.61 to 1.85 kg/m²]; and 3) week 12: difference, 1.86 kg/m² (CI, 0.92 to 2.80 kg/m²).

Thirty-four of 44 patients treated with megestrol acetate gained weight during the study. Baseline body weight, body mass index, lean mass, fat mass, and caloric intake were similar in those patients who gained weight (responders) and those 10 patients who did not gain weight (nonresponders). Nonresponders to megestrol acetate treatment tended to be more wasted (17.3% ±1.6% below ideal body weight; mean ±SE) and to have a lower CD4⁺ count (28.6 ±9.5 x 10⁶/L) than responders (13.5% ±1.6% below ideal body weight and CD4⁺ count of 56.4 ±13.0 x 10⁶/L) (P = 0.2 and P = 0.09, respectively).

Assessment of Body Composition

Bioelectrical Impedance

Body water content at weeks 4, 8, and 12 was similar in the treatment groups (Table 4). Lean mass changes from baseline were also similar in the two treatment groups at weeks 4, 8, and 12. For example, from baseline to week 12, lean mass decreased by 0.28 kg in the megestrol acetate group and by 0.34 kg in the placebo group [difference, 0.06 kg; CI, –2.07 to 2.19 kg]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for lean mass were as follows: 1) week 4: difference, 0.61 kg [CI, –1.17 to 2.39 kg]; 2) week 8: difference, 1.47 kg [CI, –3.76 to 6.70 kg]; and 3) week 12: difference, 0.08 kg (CI, –2.39 to 2.55 kg).

Fat mass increased by 1.66 kg from baseline to week 4 in the megestrol acetate group, whereas it decreased 0.29 kg in the placebo group (difference, 1.95 kg; CI, 0.78 to 3.12 kg) (Table 4). Similarly, from baseline to week 12, fat mass increased by 4.46 kg in the megestrol acetate group and decreased by 0.56 kg in the placebo group [difference, 5.02 kg; CI, 2.69 to 7.35 kg]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for fat mass were as follows: 1) week 4: difference, 1.69 kg [CI, 0.44 to 2.94 kg]; 2) week 8: difference, 4.17 kg [CI, 2.72 to 5.61 kg]; and 3) week 12: difference, 6.17 kg (CI, 3.81 to 8.53 kg).

Anthropometry

The megestrol acetate group increased mid-arm circumference from baseline to weeks 8 and 12 when compared with the placebo group (Table 5). From baseline to week 12, mid-arm circumference increased 0.82 cm in the megestrol acetate group, whereas mid-arm circumference decreased by 0.64 cm in the placebo group (difference, 1.46 cm; CI, 0.26 to 2.65 cm). Mid-arm circumference was positively correlated with body weight at week 4 (r² = 0.71), week 8 (r² = 0.56), and week 12 (r² = 0.75).

Tricep skinfold (an index of adiposity) increased from baseline to week 8 in the megestrol acetate group when compared with the placebo group (Table 5). From baseline to week 12, tricep skinfold increased by 1.46 mm in the megestrol acetate group, whereas it decreased by 0.92 mm in the placebo group (difference, 2.38 mm; CI, –0.13 to 4.90 mm). Mid-arm muscle circumference and mid-arm muscle area (indices of muscle mass) increased relative to baseline at all follow-up visits in the megestrol acetate group; these changes were similar to those seen in the placebo group for weeks 4 and 12 (Table 5). From baseline to week 8, mid-arm muscle area increased by 120 mm² in the megestrol acetate group and decreased by 181 mm² in the placebo group [difference, 301 mm²;^CI, –37 to 640 mm²]. Excluding the 13 closely eligible patients, the CIs for the differences in mean changes from baseline for mid-arm muscle area were as follows: 1) week 4: difference, 30 mm² [CI, –271 to 331 mm²]; 2) week 8: difference, 273 mm² [CI, –120 to 665 mm²]; and 3) week 12: difference, 310 mm² (CI, –37 to 658 mm²).

Similar to the increase in body fat as measured by bioelectrical impedance, mid-arm fat area (an index of body fat) increased from baseline at weeks 8 and 12 in patients receiving megestrol acetate (Table 5). From baseline to week 12, mid-arm fat area increased by 230 mm² in the megestrol acetate group and decreased by 124 mm² in the placebo group [difference, 355 mm²;^CI, –21 to 689 mm²]. Excluding the 13 closely eligible patients, the 95% CIs for the differences in mean changes from baseline for mid-arm fat area were as follows: 1) week 4: difference, 36 mm² [CI, –61 to 133 mm²]; 2) week 8: difference, 168 mm² [CI, –19 to 355 mm²]; and 3) week 12: difference, 319 mm² (CI, –39 to 678 mm²).

The body composition results obtained by bioelectrical impedance positively correlated with the anthropometric assessment of body composition. For example, mid-arm fat area was positively correlated with fat mass at week 4 (r² = 0.39), week 8 (r² = 0.31), and week 12 (r² = 0.47). Overall, the body composition results were similar whether anthropometry or bioelectrical impedance was used as an assessment tool.

Patients' Assessment of Well-Being

Durational cohort analysis showed that patients receiving megestrol acetate had more positive answers to the well-being assessment questions than did patients receiving placebo at weeks 4, 8, and 12 (data not shown). To summarize the well-being assessment questionnaires, analyses were done on the patients' last evaluation (Appendix Table). As in the durational cohort analysis, the mean score of the megestrol acetate group was higher than the mean score of the placebo group for each question (P < 0.05 except for questions 3 and 9). For each question, a strong correlation was noted between response and maximal weight change. For example, the question "Since beginning treatment has your quality of life become better or worse?" was correlated with maximal weight change (r² = 0.59).

The Karnofsky performance status at the patients' last evaluation was not statistically different from baseline in either treatment group. Improvement was noted in 6 patients in the placebo group and 5 in the megestrol acetate group; 18 patients in the placebo group and 21 patients in the megestrol acetate group were unchanged; and 15 patients in the placebo group and 20 patients in the megestrol acetate group had worse Karnofsky performance status at their last evaluation.

The changes from baseline in systolic and diastolic blood pressure and temperature were similar in the treatment groups throughout the study. Skin reactivity tests showed that patients in each treatment group were similarly anergic at baseline. No effect of megestrol acetate treatment on skin reactivity was noted (data not shown). Seven (15%) patients who received placebo and nine (17%) patients treated with megestrol acetate developed opportunistic infections during the study (P = 0.46). Pneumocystis carinii pneumonia was the most common infection; it occurred in three patients in each group.

The following adverse events were reported by patients: general pain (4 patients treated with placebo; 2 treated with megestrol acetate), diarrhea (3, placebo; 4, megestrol acetate), flatulence (3, megestrol acetate), asthenia (3, placebo; 2, megestrol acetate), and impotence (2, megestrol acetate). Because of the greater number of deaths in the megestrol acetate group, a follow-up search of patients who participated in the study was done. Investigators were asked to review clinical records for additional follow-up or death (or both) of those patients for whom data were censored. Pracon (Reston, Virginia) contacted sites about the status of those patients who were lost to follow-up. A date of death was recorded only if the date was known through medical records or a death certificate. For all other patients, a date of death was requested through the National Social Security Database. As seen in Figure 1, the median survival of patients treated with megestrol acetate was equal to the median survival of patients who received placebo.

View larger version (19K): [in this window] [in a new window]   Figure 1. Effect of megestrol acetate and placebo on survival from start of study. P > 0.82 for test of equality of survival curves. MA = megestrol acetate.

Discussion

Top Methods Results Discussion Author & Article Info References

Megestrol acetate has been shown to increase appetite and weight gain in patients with breast cancer [11], other cancers [12-16], and AIDS [17, 23, 24]. This study confirms these findings and shows that a new formulation of high-dose megestrol acetate given to cachectic and anorexic patients with AIDS with unidentified wasting is well tolerated, improves appetite, increases food intake, results in substantial weight gain, and increases the sense of well-being. Inclusion of patients who were within 1.8 kg of the strict weight entry criteria did not change the overall results.

Because megestrol acetate is a derivative of progesterone, concern exists about the potential effect of megestrol on the endocrine system. Recently, it was reported [25] that humans receiving 160 or 800 mg of megestrol acetate per day developed low serum cortisol concentrations caused by suppression of the pituitary-adrenal axis. Although body weight gain in the current study was primarily in the trunk area, only one patient treated with megestrol acetate had a "cushingoid" appearance. With regard to the potential effect of megestrol acetate on glucose metabolism, cats treated with megestrol acetate develop fasting hyperglycemia and suppression of basal and adrenocorticotrophic hormone-stimulated cortisol concentrations [26] as well as eosinopenia, glycosuria, and hepatocyte swelling from glycogen deposition [27]. Hyperglycemia is also a reported side effect of megestrol acetate therapy in humans [28]. It should be emphasized that in this 12-week study, no symptoms of diabetes mellitus or hyperglycemia were observed in patients treated with megestrol acetate.

Assessment of body composition by bioelectrical impedance showed that the megestrol acetate-stimulated weight gain was attributable to a gain in fat mass, not a shift in water content, which can occur rapidly in ill patients [10]. Bioelectrical impedance is a noninvasive technique used to assess body composition. Variables such as age, sex, and hydration of body compartments influence the accuracy and precision of body composition estimation. Although differences in hydration of individual body compartments of malnourished persons may result in over- or underestimation of body compartments and may have affected this study, the indirect estimation of body composition with bioelectrical impedance has been correlated with total body potassium, densitometry, anthropometry, and isotope-dilution methods in healthy and malnourished persons [29]. In this study, the body composition determined by bioelectrical impedance correlated with the body composition estimated by anthropometry; both techniques showed that megestrol acetate treatment caused a significant gain in fat mass. Therefore, bioelectrical impedance can be regarded as a valid tool for estimation of body composition, although differences in individual body compartment hydration cannot be excluded.

Despite an increased intake of protein, megestrol acetate promoted fat mass accrual more than lean mass at all time points. This suggests 1) that the progesterone-like effects of megestrol acetate may include stimulation of appetite and fat synthesis resulting in increased adipose stores; 2) that megestrol acetate may stimulate appetite alone and the resulting increased caloric intake is stored according to the existing hormonal milieu; or 3) that the relative inactivity of these patients contributed to the lack of lean body mass gain. Although an increase in lean body mass did not occur with megestrol acetate treatment, the enhanced sense of well-being and quality of life of the patients warrants serious consideration of using megestrol acetate treatment for cachectic patients with AIDS who have unidentifiable causes of wasting.

Although most patients treated with megestrol acetate gained weight, megestrol acetate failed to stimulate appetite and body weight gain in some patients. Comparison of the response of patients after megestrol acetate therapy shows that patients who were more wasted and had lower CD4⁺ counts were less likely to respond to megestrol acetate. Thus, earlier intervention with megestrol acetate therapy, before such severe weight loss and immune deficiency, might improve the chances for successful weight gain in cachectic patients with AIDS who have unidentified causes of wasting.

This study provides evidence supporting the observation that 800 mg of megestrol acetate per day enhances appetite, promotes weight gain, and improves quality of life in cachectic and anorexic patients with AIDS. Because patients who gained weight after megestrol acetate tended to be less wasted than nonresponders, the early use of megestrol acetate for cachectic patients with AIDS needs evaluation.