Introduction

Hypothyroidism is usually accompanied by raised serum cholesterol levels and hypertension, and is known to be associated with cardiovascular disease. It has not been established to date, however, whether subclinical hypothyroidism, which is common in elderly women, carries the same risk. Now an analysis of data from the Rotterdam Study has provided some evidence.

Method

The Rotterdam Study is a large population-based cohort study of older persons living in a suburb of the Dutch city. For the purpose of this study, a random sample of 1,149 postmenopausal women was selected from the entire study population. Baseline data collected between 1990 and 1993 included height, weight, smoking status, medication history, blood pressure, resting ECG, nonfasting serum total cholesterol, high-density lipoprotein cholesterol, total serum protein, serum albumin, thyroid stimulating hormone (TSH), serum free thyroxine (if TSH level was abnormal), and serum antibodies to thyroid peroxidase.

A TSH level above 4.0 mU/L in the presence of a normal free thyroxine level (11 - 25 pmol/L) was indicative of subclinical hypothyroidism. If the THS was above 4.0 mU/L and the free thyroxine level decreased, clinical hypothyroidism was diagnosed.

Visible calcification in the lumbar aorta on radiography was graded to provide an assessment of aortic atherosclerosis; mild, moderate and severe degrees were subsequently grouped to allow categorization into "present" and "absent". Myocardial infarction (MI) at baseline was assessed by self-reporting and from the resting ECG.

During the follow-up period, which ran until 1996, Information on fatal and nonfatal MIs was gathered from the family physicians, and verified by checking the patients' medical records. Additional data came from hospital discharge reports.

Results

Of the 1,149 women evaluated at baseline, 94 (8%) were found to have clinical hypothyroidism, and were excluded. Of the remaining 1,055 women, 931 were euthyroid and 124 (11.8%) had subclinical hypothyroidism. The two groups did not differ from one another significantly except for significantly lower levels of serum total cholesterol in the subclinical hypothyroidism subjects. Their mean age was 69 years.

At baseline, 53% of the women had aortic atherosclerosis and 7.5% had a history of MI. In women with these conditions the prevalence of subclinical hypothyroidism was increased - 14% and 21.5%, respectively. Conversely, women with subclinical hypothyroidism had an increased risk of having aortic atherosclerosis (odds ratio 1.7) and a history of MI (odds ratio 2.3). The risk was slightly greater in the presence of autoantibodies - odds ratios of 1.9 (CI, 1.1 - 3.6) and 3.1 (CI, 1.5 - 6.3), respectively.

During the follow-up period, which averaged 4.6 years, 16 women had a first incident MI - using a Cox proportional hazards regression analysis, it was found that the adjusted relative risk for women with subclinical hypothyroidism was 2.5 (CI, 0.7 to 9.1, i.e. not statistically significant).

The authors computed an attributable risk percentage of 60 for subclinical hypothyroidism associated with MI - i.e. the thyroid condition contributed to 60% of cases of MI in patients with subclinical hypothyroidism. In the same population, the attributable risk percentages were 58 for hypercholesterolemia, 38 for hypertension, 50 for current smokers, and 58 for diabetes.

Comment

The results of this study appear sound. The authors justify the use of radiography to determine aortic atherosclerosis, and give satisfactory percentage figures for recovery of medical data from their original sources. The prevalence of elevated TSH in this study closely resembles that reported in the Framingham study and other major epidemiological studies. Finally, the relative risk of incident MI occurring during the 4.6-year follow-up (2.5), although not statistically significant, corresponded closely to the odds ratio for a history of MI at the baseline analysis (2.3).

There was no association between the occurrence of autoantibodies and aortic atherosclerosis or MI in the subjects in this study, although there was a slightly stronger association between subclinical hypothyroidism and these conditions when autoantibodies were present. The authors argue that the presence of antibodies is associated with a more severe or persistent form of subclinical hypothyroidism, and hence a higher odds ratio for cardiovascular disease.

This study shows conclusively that subclinical hypothyroidism is quite prevalent in elderly women, and is strongly associated with aortic atherosclerosis and myocardial infarction. It is clear that this thyroid condition must be added as an equal partner to the better-known risk factors for these cardiovascular diseases - smoking, high cholesterol, diabetes, and hypertension.