AIDS Wasting Syndrome: Trends, Influence on Opportunistic Infections, and Survival

The authors examined data from a large cohort of HIV-infected persons to demonstrate recent trends in wasting syndrome, to examine the influence of wasting syndrome on the incidence of other opportunistic illnesses, and to explore if any of the commonly prescribed treatments for wasting are associated with improved survival. Kaplan-Meier analysis and multivariate left-truncated Cox models were used to estimate time to death after the first diagnosis of wasting syndrome and to quantify the association between the covariate and mortality, respectively. The incidence of wasting declined during 1992 through 1999, with the most marked rate of decline occurring after 1995. The incidence of AIDS- and non-AIDS-defining illnesses was generally high at or after a diagnosis of wasting syndrome. Factors significantly associated with improved survival include having a CD4⁺ count of ≥200 cells/L during the interval of the wasting syndrome diagnosis and antiretroviral therapy with two or more drugs at or after the diagnosis of wasting syndrome. Prescription of oxandrolone was associated with improved survival, but the results did not quite reach statistical significance. The authors' study provides supportive information that treatment of wasting syndrome may have a favorable impact on survival.

The incidence of AIDS-defining opportunistic illnesses (OIs) has markedly decreased since the introduction of highly active antiretroviral therapy (HAART) in the United States and elsewhere (^1-3). Although the incidence of wasting syndrome has decreased by approximately 50% (²), the percentage of persons with wasting syndrome at the time of their AIDS diagnosis increased during 1992 through 1997, with the highest percentage (12.2%) occurring in 1997.

A recent report on a cohort of HIV-infected persons enrolled in the Nutrition for Healthy Living Study found that nearly 14% of this cohort self-reported wasting (loss of >10% of usual body weight) at the time of enrollment and that many continued to lose weight despite the use of HAART (⁴). These investigators concluded that weight loss and wasting are common problems for HIV-infected persons despite the widespread use of HAART.

Because wasting has a significant impact on quality and duration of life and treatments for wasting may be quite expensive (⁵), we examined data from a large cohort of HIV-infected persons to demonstrate recent trends in wasting syndrome, examine the influence of wasting syndrome on the incidence of other OIs, and explore if any of the commonly prescribed treatments for wasting are associated with improved survival.

MATERIALS AND METHODS

Data for this study were obtained from the Adult and Adolescent Spectrum of HIV Disease Project (ASD). The ASD is a national surveillance project of the Centers for Disease Control and Prevention (CDC), in collaboration with 11 state and local health departments, in which information is abstracted from the medical records of HIV-infected patients at selected health care facilities. The facilities are composed of approximately 70% public and 30% private institutions (²). Initial abstraction of data from these records involves collection of information on demographics, the mode of exposure to HIV, previous occurrences of conditions listed in the surveillance case definition of AIDS (⁶) and other conditions, types of medications prescribed, and CD4⁺ cell counts recorded during the year before patient selection. The initial data abstraction is a review of the 12 months before enrollment, except for AIDS-defining conditions, which are recorded if available at any time before enrollment. The initial abstraction is followed by 6-month abstractions until death or loss to follow-up. Data collection began in 1990 (Atlanta, GA; Dallas, Houston, and San Antonio, TX; Denver, CO; Detroit, MI; Los Angeles, CA; New Orleans, LA; and Seattle, WA), 1991 (New York City, NY), and 1992 (Bayamon, PR). The more than 100 participating facilities include hospitals, outpatient offices, and emergency rooms. Data in this analysis were collected through December 1999. Sampling of patients for inclusion is conducted at some sites: 50% of black men in Atlanta since January 1994; 25% to 50% of white men in Dallas, San Antonio, and selected sites in Seattle since approximately 1990 through 1991; no white men (except injection drug users [IDUs] for some Los Angeles sites since 1992; 40% to 50% of all men at Detroit sites since 1993; and 16% (since January 1995) to 50% of white men at some New York City sites. Data are abstracted from multiple participating health care facilities for a given patient. Patients in this analysis were observed in the ASD from 1992 to 1999 and followed through December 1999.

Wasting syndrome was defined according to the CDC AIDS surveillance case definition (⁶): findings of profound involuntary weight loss of >10% of baseline body weight plus either chronic diarrhea (at least two loose stools per day for ≥30 days) or chronic weakness and documented fever (for ≥30 days, intermittent or constant) in the absence of a concurrent illness or condition other than HIV infection that could explain the findings (e.g., cancer, tuberculosis, cryptosporidiosis, other specific enteritis). Wasting syndrome may be the only AIDS diagnosis or may be reported concurrently with other AIDS conditions not responsible for the wasting. Collection of data on specific medications began at different times as the study was updated. For example, data on Megace were collected since the beginning of the study, but data on oxandrolone and protease inhibitors were collected as these treatments emerged. Collection of height and weight data began in 1998. Body mass index (BMI) was calculated for the preceding interval if it was not available for the interval of initial wasting syndrome diagnosis.

Incidence was calculated using only intervals of observation occurring after enrollment and excluded patients with a history of wasting syndrome before follow-up. The incidence rate was calculated by dividing the number of wasting syndrome cases in an interval by the total follow-up in person-years in that time period. Persons with more than one interval with wasting syndrome were included in more than one incidence calculation. The follow-up time for an individual was calculated as the number of months in the interval (minimum = 0, maximum = 6) that the person was in the ASD. Total follow-up time was the sum of each individual's follow-up time. Incidence calculations did not duplicate the methods used in a previous ASD report, which standardized the results to the national AIDS surveillance population (²). Therefore the numeric results are not directly comparable to that report.

Kaplan-Meier plots were used to estimate survival probabilities following the first diagnosis of wasting syndrome and were stratified by the years of diagnosis: 1992 through 1995 and 1996 through 1999. Left-truncated multivariate Cox proportional hazards models of time from the first diagnosis of wasting syndrome to death were fitted for demographic variables (sex, age [≤50 years or >50 years], mode of HIV transmission, race/ethnicity, and site), clinical conditions (CD4⁺ cell count [<200 cells/μL, ≥200 cells/μL, or missing] and Pneumocystis carinii pneumonia [PCP], disseminated Mycobacterium avium complex, cerebral toxoplasmosis, and extrapulmonary cryptococcosis [OIs for which prophylaxis, which could influence survival, is commonly prescribed]), treatment variables (testosterone, growth hormone, oxandrolone, Megace, and dronabinol), and combination antiretroviral therapy or antiretroviral monotherapy. Left-truncated proportional hazard models were used, because patients had different calendar times for their first diagnosis of wasting. Multivariate risk (hazard) ratios and 95% CIs from the left-truncated Cox models were used to quantify the association between the covariate and mortality. For determination of time to death after first diagnosis of AIDS-defining wasting syndrome, the Kaplan-Meier figure was truncated at the median follow-up (approximately 20 months). All statistical analyses were performed using SAS version 6.12 (SAS, Cary, NC) (⁷).

RESULTS

The study population was large and demographically diverse (N = 46,678). Of the total, 79.5% were male and 61.8% were nonwhite. The transmission mode for HIV infection included male-male sex (46.1%), injection drug use (18.8%), male-male sex and injection drug use (9.6%), and heterosexual contact (9.8%). The median age at diagnosis of wasting (or enrollment in the ASD for other patients) was 34 years. We identified 3289 patients with wasting syndrome among the 46,678 ASD patients followed during 1992 through the end of 1999; 2383 of these patients were diagnosed with wasting syndrome during follow-up (wasting syndrome diagnosed only after the initial abstraction), and incidence calculations were based on this latter number. The incidence of wasting syndrome decreased from 30.2 cases per 1000 person-years during the first half of 1992 to 11.9 cases per 1000 person-years during the first half of 1999 (Fig. 1). Poisson regression was used to model the wasting rate. Overall, the incidence of wasting declined, with the most marked rate of decline occurring after 1995 (χ² value of 12.51, p = .0004).

Fig. 1

The incidence of AIDS- and non-AIDS-defining OIs was generally high at or after a diagnosis of wasting syndrome (Table 1). The highest incidences were found for cytomegalovirus disease (including retinitis), infectious diarrhea (not AIDS-defining), disseminated M. avium complex, PCP, bacteremia with or without sepsis, and for a variable that combined all AIDS-defining OIs not already separately examined in the table.

Table 1

The Cox proportional hazards analysis demonstrated that factors associated with poorer survival after a diagnosis of wasting syndrome were a diagnosis of PCP, disseminated M. avium complex and cerebral toxoplasmosis, and prescription of dronabinol (Table 2). Factors associated with improved survival were lowest CD4⁺ count ≥200 cells/μL during the interval of the wasting syndrome diagnosis, age less than or equal to 50 years, prescription of two or more drugs of antiretroviral therapy at or after the diagnosis of wasting syndrome, and prescription of oxandrolone (although the results for oxandrolone did not quite reach statistical significance). Prescription of testosterone, growth hormone, and Megace were not associated with improved survival. Improved survival after the diagnosis of wasting syndrome was found for the interval of 1996 through the first half of 1999 compared with 1992 through 1995 (Fig. 2). Interactions between antiretroviral therapy and hormonal therapy, antiretroviral therapy and HIV transmission mode, and hormonal therapy and HIV transmission mode were examined and were not significant. When refitting the left-truncated Cox proportional hazards models adjusting for both CD4⁺ cell count and age as continuous variables, the significance of the other prognostic variables did not change; thus, the CD4⁺ cell count and age variables were left as categoric. In a separate analysis of 191 persons with wasting syndrome who had height and weight information, a BMI of ≥20 was associated with improved survival (RR = 0.26, 95% CI: 0.11, 0.60).

Table 2

Fig. 2

DISCUSSION

Our study found that the incidence of wasting syndrome in HIV-infected patients decreased during the 1990s in the United States. This trend was especially pronounced after the introduction of HAART in late 1995. Previous studies have attributed the decrease in OIs to the use of HAART (^1,8,9). These studies examine OIs overall or focus on the selected most common infections without specific attention to wasting syndrome, however. Our study demonstrates that the incidence of wasting syndrome has been decreasing since the early 1990s and suggests that this decrease, which preceded HAART in association with any antiretroviral therapy, was most pronounced after the introduction of HAART. The initial decline in wasting syndrome observed in the early 1990s may have been a result of the increased availability of antiretroviral medications and possibly the expanding use of prophylaxis for OIs such as M. avium complex.

Survival after a diagnosis of AIDS OIs improved during the HAART era, and our data suggest that for wasting syndrome, this improvement was a result of the combination of antiretroviral therapy and may have been influenced by the prescription of an endocrinologic treatment, oxandrolone. These findings are new; previously, it was reported that HAART had not been shown to be associated with weight gain or viral load to correlate with body weight changes (¹⁰). There are several studies that indirectly support our findings of improved survival associated with hormone-based treatments of wasting syndrome; however, no studies have directly examined survival. One report has shown that low levels of free serum testosterone correlate with loss of body cell mass (¹¹). Testosterone and nandrolone decanoate have been shown to increase lean body mass in HIV-infected men with wasting syndrome (^12,13). Megace (megestrol acetate), a progestin, may increase body weight and appetite (¹⁴). We have found no large studies of the effect of oxandrolone on HIV-infected patients with wasting syndrome. Recruitment of such patients into clinical trials may have been difficult because of the declining incidence of wasting syndrome.

We did not find a survival advantage to the prescription of growth hormone. Lean body mass has been shown to increase with the use of recombinant human growth hormone (^15,16). These effects depend on the dosage of growth hormone prescribed (a factor that we can not control for, because dosage was not collected in our study) (^15,16). Also, the effect of growth hormone on lean body mass is not maintained in many patients. Published studies of growth hormone treatment of wasting syndrome have not included prolonged follow-up and survival information. Our study suggests but does not prove that growth hormone use does not independently affect survival. We lacked the statistical power needed to show a benefit. In clinical practice, the use of growth hormone is often part of a regimen that may include appetite stimulants, anabolic steroids, a healthy diet, and exercise. This combination approach might be beneficial to survival but needs further study with long-term follow-up.

Our finding that dronabinol (Marinol) was associated with worse survival could be related to a selection bias. It has been reported that dronabinol did not increase body weight, although it may decrease nausea and increase appetite (¹⁴). Possibly, health care providers are more likely to prescribe it to patients who have a worse prognosis as a palliative measure.

The incidence of other OIs (including non-AIDS-defining conditions) was high at or after a diagnosis of wasting syndrome occurred. For example, the incidence of PCP (among those with wasting syndrome) was 269 cases per 1000 person-years compared with 87.9 cases per 1000 person-years from data published by Jones et al. (²) from the ASD cohort from 1992 through 1997. Similarly, the rate for AIDS-defining chronic herpes simplex (among those with wasting syndrome) was 61 cases per 1000 person-years in our study compared with 9.8 cases per 1000 person-years in the study by Jones et al. (²). Loss of weight, poor appetite, and the associated metabolic alterations and nutritional deficiencies may all lead to increased susceptibility to infections and could exacerbate immune deficiency, increasing the likelihood of reactivation diseases such as chronic herpes simplex. Also, because all ASD patients are in care and received evaluation leading to the diagnosis of wasting syndrome, heightened recognition of other infections and neoplasms concurrently or during follow-up is expected. Finally, some of these conditions could have been responsible for wasting in these patients, but diagnostic methods to rule out these diseases had either not been performed before the diagnosis of wasting syndrome or such methods were performed but yielded negative results and later yielded positive results. The high incidence of these HIV-associated conditions demonstrates the profound state of poor health and vulnerability these patients experience and underscores the need for improved recognition and treatment of wasting syndrome. The adverse impact that AIDS OIs may have on survival after the diagnosis of wasting syndrome also underscores the importance of preventing these illnesses (¹⁷).

Wasting in an HIV-infected patient is a complex pathophysiologic process that may be caused by inadequate oral intake, alteration in metabolic processes, malabsorption, or a combination of these (¹⁸). Often, it is a marker for a neoplasm such as lymphoma or an OI such as disseminated M. avium complex, cryptosporidiosis, or tuberculosis. When it occurs in the absence of a concurrent illness or condition that could explain the findings, however, it satisfies the CDC surveillance definition for wasting syndrome, a diagnosis of exclusion without diagnostic tests or sensitive and specific signs and symptoms. The definition is well suited for surveillance purposes because of its feasibility but is less optimal for clinical studies of the treatment of wasting syndrome. These limitations and alternative definitions have been published elsewhere (⁴). Therefore, our findings in our surveillance project of a decreasing trend in wasting syndrome diagnosis and improved survival are likely reliable. Nevertheless, our findings regarding the influence of treatment of wasting syndrome on survival should be interpreted cautiously and confirmed with clinical trials because of the inherent limitations of surveillance data to address the association between treatment and clinical outcomes. The association of oxandrolone treatment with improved survival in our population should not be extrapolated as a conclusive finding generalizable to all individuals. Notably, although our study included many women, data on efficacy and target doses for anabolic therapy in women are scarce. Also, treatments such as oxandrolone may cause liver toxicity and cholesterol changes (¹⁰); therefore, patients should be carefully selected for such treatments.

Because the ASD is an observational study, patients were not randomized to receive these treatments (hidden bias may be present related to their use and to observations about current or follow-up illnesses), and relatively few patients were prescribed these treatments for wasting syndrome. It is possible that patients prescribed oxandrolone were healthier than other patients. Our results represent an intention to treat, because we do not collect information on adherence or exact duration or dose of treatment. In the absence of clinical trials comparing survival benefits of these treatments for wasting, however, our results are useful in demonstrating a possible beneficial impact of hormonal treatment on survival following wasting syndrome in a large population. In addition, our trends data could overestimate the decline in wasting syndrome if health care providers are documenting its diagnosis less frequently. The ASD is large and diverse, with thousands of HIV-infected persons enrolled since its inception. It is not a population-based study, however, and our findings may not be generalizable to all HIV-infected persons. Also, previous studies of wasting syndrome have been limited to only or primarily male patients (^12-16), but our study had a relatively large percentage of women. The CD4⁺ cell count was missing for more than one third of these patients. Because missing a CD4⁺ cell count was associated with an increased risk for death, we suspect that these persons had low CD4⁺ cell counts and possibly other problems not recorded in the ASD (e.g., poor medication adherence) that may affect survival and likelihood of having a count measured. Additional analysis examining the role of BMI on our findings would be useful. The collection of height and weight data was added to the ASD protocol relatively recently; therefore, there are insufficient data to evaluate this issue fully. Finally, our study did not control for behavioral variables such as depression and its treatment; these variables could affect appetite or albumin levels, which have been shown to correlate with survival in AIDS patients (¹⁹).

Weight loss in HIV-infected persons is associated with risk for death (²⁰); therefore, the diagnosis of wasting syndrome is an important opportunity for intensive intervention. Early diagnosis of HIV infection and entry into care are essential to help prevent this generally late stage complication. When patients are diagnosed with wasting syndrome, a thorough examination for concurrent HIV-associated illnesses is warranted, and counseling regarding HIV treatment, nutrition, exercise, and treatment of wasting syndrome should be performed.