Does treatment of acromegaly affect life expectancy?

Abstract

Previous studies have demonstrated a twofold to threefold increase in mortality in acromegalic patients largely due to increased cardiovascular disease, but also in some series due to increased malignant disease. In a retrospective analysis of 79 patients from one large hospital center in the United Kingdom, we sought to determine (1) whether the increased mortality persists and (2) what is the relationship between mortality and growth hormone (GH) levels achieved following treatment. Hospital records and death certificates were scrutinized and serum GH data collected from patients treated between 1967 and 1991. GH assessments were performed on average annually during follow-up evaluation and consisted of 3-hour measurements during waking hours in a GH day series. The average value was calculated from five readings. We used the lowest value, usually the most recent, for average GH achieved during the follow-up period. There were 29 males and 50 females, and no numerical difference between sexes was found in living and deceased groups. The mortality of the cohort was compared with that of the general population, and expected mortality was calculated from Office of Population Census Statistics data. The observed to expected ratio was determined. There were no differences between living and deceased groups in the proportion of subjects with hypertension, diabetes, or visual field defects at the time of dianosis. Most patients were treated with radiotherapy ± bromocriptine as primary therapy. There was a marked shift to lower mean GH values with treatment such that 48 of 79 had values less than 10 mU/L. Those subjects with mean GH levels less than 5 mU/L had significantly lower serum insulin-like growth factor-I (IGF-I) values than subjects with values above this level. The observed to expected mortality ratios (with 95% confidence intervals in parentheses) were as follows: males, 2.55 (1.4–4.3), P = .002; female, 2.83 (1.6–4.8), P = .001; total, 2.63 (1.8–3.9), P < .001; lowest GH less than 10 mU/L, 2.01 (0.9–3.83), P = .039; and lowest GH less than 5 mU/L, 1.42 (0.46–3.31), P = .28. These results show that overall mortality in both males and females with acromegaly is still two to four times that of the general population. However, the important novel finding is that if GH levels of less than 5 mU/L can be achieved, there is no increased mortality. The median age of death (62 years) was similar to that found previously; 57% of deaths were due to vascular causes, 25% to respiratory disease, and 11% to malignancy. Small numbers of deaths precluded statistical comparison with the general population. The implication of these preliminary data is that if serum GH levels can be reduced to less than 5 mU/L by treatment, the long-term outlook for acromegalic patients is very good. These data need to be confirmed with a larger cohort of patients.