Is vascular mortality increased in hypopituitarism?

Hitherto four relevant cohort studies have been published on the life expectancy of patients with hypopituitarism, two from the UK and two from Sweden. In all four studies patients with acromegaly or Cushing's disease were excluded, and patients with GH replacement were not included. A crude meta-analysis, not considering differences in validity of the study designs, gives an overall Relative Risk (RR) of 1.47, with a relatively small 95% confidence interval of 1.27-1.70. The aim of this critical review was to assess whether lacking validity in the design of one or several of these studies indicates that the true RR might be higher or lower than indicated by meta-analyses. A high frequency of cranial irradiation in one of the Swedish cohorts may explain part of the very high observed RR for cerebrovascular mortality, but there remained a 40% risk increase for cardiac diseases that cannot be explained by irradiation. Some other validity problems, especially seen in the UK studies, will result in under-estimation of the true RR, but for others the direction of bias is uncertain due to lacking information in the original publications. According to our view, it is more probable that the observed overall 50% risk increase for vascular mortality is an under-estimation than an over-estimation of the true RR. This analysis is of a qualitative nature, and it is not possible to quantify to what extent the lacking validity may have affected the RR.