Cachexia in rheumatoid arthritis is not explained by decreased growth hormone secretion


  • The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.



Patients with rheumatoid arthritis (RA) lose body cell mass (BCM) by unknown mechanisms. Since the loss of BCM in normal aging individuals parallels the characteristic age-related decline in growth hormone (GH) secretion, this study was carried out to determine whether further decreased GH secretion plays a role in the pathogenesis of this loss of BCM in RA patients, termed “rheumatoid cachexia.”


GH secretory kinetics were determined by deconvolution analysis in 16 patients with RA and 17 healthy controls matched for age (mean ± SD 45.4 ± 13.2 years and 47.1 ± 14.6 years, respectively), sex, race, and body mass index. Blood samples were obtained every 20 minutes for 24 hours. Body composition was ascertained using total-body potassium (TBK) as a measure of BCM and dual x-ray absorptiometry to determine fat mass.


BCM was reduced in patients with RA compared with healthy controls (mean ± SD gm TBK 79.5 ± 9.5 versus 94.9 ± 11.9; P < 0.0005), but there was no difference in fat mass. GH kinetic parameters in patients with RA did not differ from those in controls.


These findings suggest that GH kinetics are unaltered in RA patients compared with healthy subjects; thus, GH deficiency does not account for rheumatoid cachexia.

Patients with rheumatoid arthritis (RA) experience a loss of metabolically active tissue, or body cell mass (BCM), a condition “rheumatoid cachexia” (1). The consequences of BCM loss of this degree are physical inactivity, diminished strength, and decreased functional status, creating a cycle in which loss of lean mass is perpetuated, and fat mass tends to increase. Many factors appear to contribute to the loss of BCM in patients with RA, including catabolic cytokines, inadequate dietary intake, steroid effects, and decreased physical activity (2).

We also considered whether reduced growth hormone (GH) secretion might contribute to changes in body composition in RA. Studies of healthy aging, in which a loss of lean mass occurs, have shown an association between loss of lean mass and declining activity of the GH–insulin-like growth factor 1 (IGF-1) axis (for review, see ref. 3). Evidence of multiple endocrine alterations in RA (4), combined with the decreased lean body mass and increased mortality seen in this disease, raise the question of whether GH is reduced in patients with RA and whether a reduction in GH may contribute to some of the body composition changes in these patients.

The goal of the present study was to elucidate the possible role of altered GH secretion in loss of BCM in RA. To investigate this, we measured endogenous 24-hour GH kinetics among patients with RA and matched healthy control subjects. A secondary goal was to determine whether there were abnormalities in GH secretion related to interleukin-1 (IL-1), tumor necrosis factor (TNF), or IL-6 levels.



Sixteen women with RA (ages 24–70 years) and 17 women matched for age, race, weight, height, and menstrual status (ages 24–75 years) were studied. Premenopausal women were studied between days 3 and 7 after menses, during the follicular phase of the menstrual cycle, because of previous data indicating that estrogen status affects GH secretion (5). Patients with RA met the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) criteria for RA (6), and were stratified by functional status based on the ACR criteria (7). Disease was considered by the patients' rheumatologists to be under good control: all had low disease activity, but there was a range of disease severity. Patients with RA had to have been on a stable medication regimen, including a stable prednisone dosage, for 3 months prior to entering the study. Potential subjects were excluded if they were obese (body mass index [BMI] >30 kg/m2) or had cancer, renal disease, liver disease, coronary artery disease, endocrine disorders including diabetes, or autoimmune disease other than RA.

The research protocol was approved by the New England Medical Center/Tufts University Human Investigation Review Committee, and written informed consent was obtained from all participants.


BCM was measured by calculating total-body potassium (TBK) from measurements of endogenous 40K in a whole-body counter as reported previously (2). This measure is highly correlated with muscle mass (r = 0.9, P < 0.0001) (8). Body fat was measured by dual x-ray absorptiometry using a QDR2000 apparatus (Hologic, Waltham, MA) with the manufacturer's software (version 5.64A). Subjects underwent a whole-body scan with the array mode.

An indwelling catheter was placed in a forearm vein, from which 1 ml of blood was withdrawn every 20 minutes for 24 hours, for determination of GH secretory kinetics. Plasma was obtained at 7:00 AM, after a 12-hour fast, for measurements of IGF binding protein 1 (IGFBP-1), and IGFBP-3, which are known to modulate the action of IGF. GH was measured by chemiluminescence using a commercial assay (Nichols Institute Diagnostics, San Juan Capistrano, CA). IGF-1 was measured after acid–ethanol cryoprecipitation using a single Nichols IGF-1 extraction radioimmunoassay (RIA). IGFBP-3 was measured with a single RIA (Diagnostic Systems, Webster, TX). All measurements were made in duplicate and averaged.

Measurement of cytokine production.

Peripheral blood mononuclear cells were obtained from whole blood and cultured in 24-well flat-bottomed plates for measurement of IL-1, TNF, and IL-6, as previously described (9). Measurement of total cytokine synthesis was carried out with specific, non–cross-reacting enzyme-linked immunosorbent assays (ELISAs), using specific monoclonal antibodies for the analysis of the cytokine antigens (Quantikine ELISA; R&D Systems, Minneapolis, MN).

Calculations and statistical analysis.

Deconvolution analysis was applied to compute the number, amplitude, duration, and mass of significant GH secretory bursts and estimate the apparent endogenous GH half-life in blood sampled at 20-minute intervals over 24 hours. Weighted intraseries standard deviations for the GH series were used in the estimate of the variance and covariance matrices (10, 11). Terms used in the analysis have been summarized previously (12).

The distribution of each continuous variable was checked for normality, and non-normally distributed combined data were transformed prior to analysis. Comparisons between the groups were made by analysis of variance with Tukey's honest significant difference test. Prednisone use was added as a covariate to check for potential confounding due to this medication. Scatterplots and Pearson's correlation coefficients were used to examine the association among GH kinetic parameters, BCM, and fat mass. Associations were tested using univariate linear regression and were considered statistically significant when the 2-tailed P value was less than 0.05. All data analysis was performed using SYSTAT version 9.0 (SPSS, Chicago, IL).


Characteristics of the study subjects are shown in Table 1. Eight of the 16 patients were taking prednisone (mean ± SD dosage 2.4 ± 2.7 mg/day [range 2.5–7.5]), and 6 were taking methotrexate (mean ± SD dosage 15 ± 2.2 mg/day [range 7.5–15]). Body cell mass, calculated from measurements of TBK, was reduced in patients with RA compared with healthy control subjects (mean ± SD 79.5 ± 9.5 gm versus 94.9 ± 11.9 gm, respectively, P < 0.0005) (Table 2), confirming the presence of rheumatoid cachexia. BMI (kg/m2) and body fat did not differ significantly between the 2 subject groups (Table 2).

Table 1. Characteristics of the study subjects*
 RA (n = 16)Controls (n = 17)
  • *

    Values are the mean ± SD. RA = rheumatoid arthritis; ESR = erythrocyte sedimentation rate.

  • P < 0.05 versus controls.

Age, years45.4 ± 13.247.1 ± 14.6
ESR, mm/hour28.5 ± 14.018.4 ± 13.2
Duration of RA, years9.7 ± 10.2
Painful joints, no.4.5 ± 7.9
Swollen joints, no.4.7 ± 7.9
Pain, 0–15-cm scale5.8 ± 3.6
Prednisone, mg/day (n = 8)2.4 ± 2.7
Table 2. Body composition and endocrinologic parameters in the study subjects*
 RA (n = 16)Controls (n = 17)
  • *

    Values are the mean ± SD. RA = rheumatoid arthritis; TBK = total-body potassium; GH = growth hormone; IGF-1 = insulin-like growth factor 1; IGFBP-1 = IGF binding protein 1.

  • P < 0.0005 versus controls.

Body cell mass, gm TBK79.5 ± 9.594.9 ± 11.9
Body mass index, kg/m224.7 ± 4.223.7 ± 3.0
Body fat, %40.5 ± 10.336.0 ± 8.2
GH kinetic parameters  
 Basal secretion, μg/liter/minute0.005 ± 0.0040.004 ± 0.006
 Half-duration, minutes28.6 ± 11.029.6 ± 8.4
 Half-life, minutes16.5 ± 3.216.2 ± 3.1
 Bursts/24 hours, no.10.7 ± 2.512.3 ± 3.3
 Interburst interval, minutes134.4 ± 32.1119.6 ± 32.3
 Mass, μg/liter6.2 ± 3.96.5 ± 6.2
 Peak amplitude, μg/liter/minute0.29 ± 0.300.19 ± 0.15
 Production rate, μg/liter/day59.4 ± 30.375.7 ± 63.8
 Mean concentration, μg/liter1.39 ± 0.951.47 ± 1.21
 Integrated 24-hour area, μg/liter/minute1,982 ± 1,3462,084 ± 1,724
 Approximate entropy0.69 ± 0.190.72 ± 0.23
 IGF-1, ng/ml153.8 ± 63.9201.8 ± 86.6
 IGFBP-1, ng/ml41.6 ± 22.841.9 ± 20.2
 IGFBP-3, ng/ml3,007 ± 5883,324 ± 602

There were no significant differences in GH kinetic parameters between patients with RA and healthy control subjects, as evaluated either by univariate analysis or by multivariate analysis with adjustment for current prednisone dosage, BCM, and fat mass as covariates (Table 2). Mean IGF-1 levels tended to be lower in the patients with RA than in controls, but the difference was not significant (P = 0.08). There were no differences in mean IGFBP-1 or IGFBP-3 levels. These findings did not differ by prednisone status (P > 0.10; data not shown).

Among all subjects (controls and those with RA), several parameters of GH secretion were inversely correlated with fat mass: basal secretion rate (r = −0.51, P = 0.003), mass of secretion (r = −0.51, P = 0.003), production rate (r = −0.44, P = 0.01), mean concentration (r = −0.46, P = 0.01), and 24-hour integrated levels (r = −0.43, P = 0.02). Fat accounted for between 18% and 26% of the variability in GH, with increasing body fat being associated with lower GH kinetics. When healthy control subjects and patients with RA were considered separately, integrated 24-hour GH levels were inversely associated with body fat in the controls (r = −0.54, P < 0.03) but not in the patients (r = 0.30, P = 0.26). Peripheral blood mononuclear cell production of the inflammatory cytokines IL-1, TNF, and IL-6, did not correlate with any GH parameter (data not shown). We did observe an increase in TNF production among patients with RA compared with healthy control subjects (P < 0.03).


In this study of RA patients and healthy controls, we found no differences between the 2 groups in any of the GH kinetic parameters. Consistent with findings of previous studies (1, 2), patients with RA had significantly reduced BCM compared with healthy control subjects, but no difference in fat mass.

The present results contrast with those of 2 recent studies in which reduced GH responses were found among patients with RA after GH-releasing hormone stimulation (13) and insulin-induced hypoglycemia (14); however, the measures used to evaluate GH secretion in the present study differ from those used in the earlier investigations. It is difficult to draw comparisons between this study and previous studies of GH in RA since each study had a particular set of clinical criteria to define a group of patients with RA, and GH was not measured in the same manner (stimulated versus basal secretion). Therefore, each study contributes to the knowledge base regarding GH secretion in RA, and the findings are not necessarily contradictory.

The potential confounding effect of treatment with corticosteroids and other medications cannot be ignored. However, we have previously shown that body composition was abnormal in RA patients regardless of steroid treatment, suggesting that the low doses of corticosteroids used in RA do not materially affect physical status (2). Moreover, it has previously been shown that while high-dose “pulse” corticosteroids cause negative protein balance in patients with RA (15), low-dose oral corticosteroids, such as those used by the patients in the present study, may in fact protect against loss of lean body mass by reducing inflammation to a greater extent than they cause catabolism (1). With regard to methotrexate, it is unlikely that the relatively low doses used by patients with RA would have a significant impact on metabolism: it has been demonstrated recently that even much higher doses of this medication administered to cancer patients do not alter body composition (16). If anything, methotrexate use may tend to normalize body composition in RA, at least in terms of protein metabolism (17).

Only fat mass was found to be an important predictor of GH secretion, which emphasizes its important confounding influence. While the current study is the first to examine GH kinetics in RA, reduced GH secretion in obese humans has long been recognized (18). Obese individuals have a defect in GH dynamics involving both secretion and clearance, and the severity of the GH secretory deficiency varies with the degree of obesity (18). Although the subjects in the present study were not obese (BMI in both groups was ∼24 kg/m2), the association between fat mass and GH secretion was evident even within the relatively narrow range of body fat in these groups.

We observed no abnormalities in GH secretion related to in vitro–stimulated cytokine secretion. It is possible that if the RA patients' disease had been more active at the time of the study, we would have found an inverse relationship between GH and the inflammatory cytokines. In summary, our results clearly demonstrate that persistent GH deficiency is not the cause of rheumatoid cachexia.