Elsevier

The Lancet

Volume 357, Issue 9254, 10 February 2001, Pages 425-431
The Lancet

Articles
Association between premature mortality and hypopituitarism

https://doi.org/10.1016/S0140-6736(00)04006-XGet rights and content

Summary

Background

Four retrospective studies have reported premature mortality in patients with hypopituitarism with standard mortality ratios (SMRs) varying between 1·20 and 2·17. Patients with hypopituitarism have complex endocrine deficiencies, and the mechanisms underpinning any excess mortality are unknown. Furthermore, the suggestion has emerged that endogenous growth-hormone deficiency might account for any excess mortality. We aimed to clarify these issues by doing a large prospective study of total and specific-cause mortality in patients with hypopituitarism.

Methods

We followed up 1014 UK patients (514 men, 500 women) with hypopituitarism from January, 1992, to January, 2000. 573 (57%) patients had non-functioning adenomas, 118 (12%) craniopharyngiomas, and 93 (9%) prolactinomas. SMRs were calculated as the ratio of observed deaths to the number of deaths in an age-matched and sex-matched UK population.

Findings

The number of observed deaths was 181 compared with the 96·7 expected (SMR 1·87 [99% CI 1·62–2·16], p<0·0001). Univariate analysis indicated that mortality was higher in women (2·29 [1·86–2·82]) than men (1·57 [1·28–1·93], p=0·002), in younger patients, in patients with an underlying diagnosis of craniopharyngioma (9·28 [5·84–14·75] vs 1·61 [1·30–1·99], p<0·0001), and in the 353 patients treated with radiotherapy (2·32 [1·71–3·14] vs 1·66 [1·30–2·13], p=0·004). Excess mortality was attributed to cardiovascular (1·82 [1·30–2·54], p<0·0001), respiratory (2·66 [1·72–4·11], p<0·0001), and cerebrovascular (2·44 [1·58–4·18], p<0·0001) causes. There was no effect of hormonal deficiency on mortality, except for gonadotropin deficiency, which, if untreated was associated with excess mortality (untreated 2·97 [2·13–4·13] vs treated 1·42 [0·97–2·07], p<0·0001). Multiple regression analyses identified age at diagnosis, sex, a diagnosis of craniopharyngioma, and untreated gonadotropin deficiency as independent significant factors affecting mortality.

Interpretation

Patients with hypopituitarism have excess mortality, predominantly from vascular and respiratory disease. Age at diagnosis, female sex, and above all, craniopharyngioma were significant independent risk factors. Specific endocrine-axis deficiency, with the exception of untreated gonadotropin deficiency, does not seem to have a role.

Introduction

Four retrospective studies have examined mortality in patients with hypopituitarism and all have confirmed increased mortality compared with age-matched controls.1, 2, 3, 4 The exact cause of this increased mortality remains unclear and has been the subject of much debate. Respiratory, cardiovascular, and cerebrovascular diseases have all been cited as being responsible for the increase in mortality. 1, 2 However, there are major limiting factors with all the studies published to date. They include differences in treatments, inclusion criteria, and duration of follow-up, and above all, none has had sufficient power in terms of numbers of patients to disentangle the complexities of this disorder and to determine exactly which factors are most important in influencing mortality. Patients with hypopituitarism have compound endocrine deficiencies, any of which may have an important role. They also undergo various treatments, surgery, drug therapy, and radiotherapy, all of which might contribute to outcome. Finally, to label all patients as simply having a diagnosis of hypopituitarism is probably unjustified, since this definition encompasses many underlying diagnoses. Acromegaly and Cushing's disease are known to confer excess mortality because of their association with high concentrations of growth hormone and cortisol, respectively.5, 6, 7, 8 Having excluded such cases, whether differing causes of hypopituitarism convey a good or bad prognosis is uncertain.

In this prospective study, we examined total and specific-cause mortality in more than 1000 patients with a diagnosis of hypopituitarism, and attempted to identify factors contributing to the excess mortality. We did subgroup analyses to explore many of the questions that still exist about mortality in hypopituitarism—eg, the role of growth-hormone deficiency, sex-steroid status, and corticosteroid replacement therapy, all of which have been implicated in mortality.2, 4, 9, 10 We also examined the use of radiotherapy amid increasing concern over its long term safety and complications.11, 12

Section snippets

Patients

The West Midlands Hypopituitary database was established in 1990 and has the approval of local hospital ethics committees. The database comprises patients with underlying pituitary disease from 16 referral centres across the West Midlands region in the UK (population about 5·5 million, representing about 10% of the UK population). All patients with a diagnosis of hypopituitarism who were alive on Jan 1, 1992, were entered onto the database the analysis for this study was done 8 years later in

Patients

1014 patients had a diagnosis of hypopituitarism, of whom 514 were men. Median age at diagnosis in men was 46·2 years (range 4–81) and in women 45·3 years (range 4–86). Median duration of disease (from diagnosis) was 12·1 years (range 0·5–49) in men and 12·7 years (range 0·5–49) in women. 573 (57%) patients were diagnosed as having non-functioning adenomas, 118 (12%) craniopharyngiomas, 93 (9%) prolactinomas, and 92 (9%) were labelled as having idiopathic hypopituitarism (table 1). Median

Discussion

Since the seminal publication of Rósen and Bengtsson,1 four retrospective studies have examined mortality in patients with hypopituitarism. All have shown higher mortality than the general population, with an excess mortality among women. However, they all have limitations and have produced conflicting data. First, the cohorts of patients are not identical, have had varying periods of follow-up, and comprise relatively small numbers of patients (172–344). Second, there have been wide variations

References (30)

  • SM Orme et al.

    Mortality and cancer incidence in acromegaly: a retrospective cohort study

    J Clin Endocrinol Metab

    (1998)
  • JC Flickinger et al.

    Incidence of cerebral infarction after radiotherapy for pituitary adenoma

    Cancer

    (1989)
  • KE Murros et al.

    The effect of radiation on carotid arteries

    Arch Neurol

    (1989)
  • J Bowen et al.

    Stroke after pituitary irradiation

    Stroke

    (1992)
  • M Bitzer et al.

    Progressive cerebral occlusive disease after radiation therapy

    Stroke

    (1995)
  • Cited by (0)

    View full text