Current issues of ACP Journal Club are published in Annals of Internal Medicine


Older age, impaired lung or renal function, anemia, and decreased total serum protein predicted shortened survival in systemic sclerosis

ACP J Club. 1991 Nov-Dec;115:90. doi:10.7326/ACPJC-1991-115-3-090

Source Citation

Altman RD, Medsger TA Jr, Bloch DA, Michel BA. Predictors of survival in systemic sclerosis (scleroderma). Arthritis Rheum. 1991 Apr;34:403-13. [PubMed ID: 1901491]



To identify predictors of survival in patients with systemic sclerosis.


Analysis of clinical and demographic characteristics of an inception cohort followed for an average of 5.2 years.


Patients from 29 centers were referred between 1973 and 1977 to the Scleroderma Criteria Cooperative Study done by the Subcommittee for Scleroderma Criteria of the American College of Rheumatology.


264 patients with recently diagnosed definite systemic sclerosis were entered. Mean age was 49.1 years; study entry was a mean of 1.9 years after diagnosis and 5.7 years after the first symptoms of systemic sclerosis. 32 patients (12%) were lost to follow-up.

Assessment of prognostic factors

For each study entrant, 484 demographic, clinical, and laboratory variables were recorded and a serum specimen was stored. Patients were subclassified according to major organ system involvement and were followed at 1 and 3 years and then between 1983 and 1985.

Main outcome measures

Outcome was determined on 133 variables. Information was collected from patients, their relatives, physicians, charts, death certificates, and autopsy reports. Deaths were categorized as definitely related to systemic sclerosis if there was a clear medical association with the disease.

Main results

2-year survival was < 80%, 8.5-year survival was 50%, and 12-year survival was 30%. 131 patients (50%) died. 89 deaths (68%) were definitely related to systemic sclerosis, with renal involvement predominating in the causes of death (39%). Patients with renal involvement (n = 10) had the poorest median survival (3 months), followed by cardiac (n = 15, median 32 months), pulmonary (n = 104, median 78 months), gastrointestinal (n = 84, median 99 months), and no system involvement (n = 51, median 108 months). Using a survival tree model with Cox proportional hazards analysis, the variables predicting shortened survival were older age at study entry (> 64 years, P < 0.02); lower forced vital capacity (< 80% of predicted with hemoglobin > 14 gm/dL or forced vital capacity < 65% with hemoglobin ≤ 14 gm/dL, P< 0.001) and reduced carbon monoxide diffusion capacity (≤ 50% of predicted, P = 0.01); higher blood urea nitrogen (> 16 mg/dL, P < 0.001); and lower hemoglobin (≤ 11 g/dL, P = 0.003); and lower total serum protein (≤ 6 g/dL, P = 0.005). Neither male sex nor race predicted poorer survival.


Several variables predicted shortened survival in systemic sclerosis: older age, impaired lung or renal function, anemia, and decreased total serum protein.

Sources of funding: National Institutes of Health; United Scleroderma Foundation; RGK Foundation; American College of Rheumatology; Perlman Scleroderma Fund; Arthritis Foundation; Miami Veterans Affairs Medical Center.

Address for article reprint: Dr. R.D. Altman, Department of Medicine, University of Miami School of Medicine, P.O. Box 016960 (VA 111), Miami, FL 33101, USA.


The effects of systemic sclerosis can range from only minimal skin changes to a rapidly progressive course that involves multiple organs and causes early death. It clearly is of great value to physicians to have reliable predictors for patients who may have a poor prognosis. This major multicenter study is prospective, has a wide geographic representation, uses uniform classification and diagnostic criteria, and has a long-term follow-up. Appropriate statistical methods used are used.

A total of 484 information items were obtained on each patient, and the paper details 72 selected variables subdivided into 9 appropriate subsets. The data clearly show that the patients with the worst survival were those over 64 years of age with renal disease and younger patients with anemia. Other predictors included severe pulmonary disease and reduced total serum protein levels. These observations alert the physician to target studies at the function of the kidney, lung, heart, and hemopoietic systems.

It should be noted, however, that because patients were probably from tertiary care centers, those with more minimal symptoms at onset may have been excluded. Although valiant attempts were made to have as complete follow-up as possible, 12% were still lost. Previous studies have shown a worse prognosis for black women with systemic sclerosis. Although the univariate analysis of the variables did not show statistical significance for sex or race, we do not know the total of white or nonwhite patients in the study. The patients were all entered during 1973 to 1977, and the treatments now available for systemic sclerosis were lacking. It is possible that the prognosis of patients being entered into a similar study today would be much improved. This disease, however, remains a challenge to physicians, and the information in this article is of considerable help.

Evelyn V. Hess, MD
University of CincinnatiCincinnati, Ohio, USA