Thyroid replacement therapy might undermine hypothyroidism-related longevity

The observation that thyroid replacement therapy increases mortality in elderly subjects with subclinical hypothyroidism [ ] has, as its converse, the documentation of an association between hypothyroidism and increased longevity in the animal model of hypothyroidism [ ]. The corollary, in human subjects, is that a serum thyroid stimulating hormone (TSH) level which is at the upper level of the normal range also appears to be one which is associated with increased longevity [ ]. In the latter study serum TSH levels were shown to be significantly (p < 0.001) higher in subjects of median age 98 than in younger controls of median age 72 [ ]. Serum TSH levels also appear to be characterised by an age-related increase, with the consequence that, among 14,376 individuals (aged 12 or more) who were free of thyroid disease and had no thyroid antibodies, 9.9% of those who were in the 70–79 age group had serum TSH above the normal range vs 3.5% of counterparts in the 60–69 age group. For counterparts in the age group of 80 years or more the proportion of those with serum TSH above the normal range was as high as 12% [ ]. The age-related increase in serum TSH is validated, not only by cross sectional studies [ , ], but also by longitudinal studies [ , ]. In a longitudinal study which enrolled 843 subjects aged 65 or more, when comparing serum TSH levels taken during the period 1992–1993 with those from 2005 to 2006(in the same set of subjects), a 13% increase (p < 0.01) in TSH was documented, concurrently with a 1.7% increase (p = 0.01) in free thyroxine (FT4) level, and a 13% decrease (p < 0.01) in serum total triiodothyronine [ ]. None of the subjects were taking thyroid medication [ ]. In another longitudinal study, there were 908 participants of mean age 45 at the beginning of the study. On 13 year follow up of those 908 participants mean serum TSH had increased from 1.49 mU/L to 1.81 mU/L (p < 0.001) whereas mean serum FT4 had not changed. The greatest increase in TSH occurred in the oldest participants. After adjustment for gender the difference in TSH between the two time periods increased, on average, by 0.08 mU/L with each additional decade of age at baseline. The fact that there was no concurrent fall in FT4 in any age group was interpreted as being suggestive of an age-related alteration in TSH set point [ ].