Hypothyroidism is a common chronic disease caused by insufficient production of thyroid hormones [1]. Thyroid hormones produced by the thyroid gland regulate metabolism in all tissues, including the heart, liver, brain, muscles, and bones. Any interruption of the thyroid hormone balance can lead to metabolic dysfunction. The overall prevalence of hypothyroidism is approximately 5–10% of the general population in the United Kingdom (UK) [2] and is greater in females and people aged over 60 years [3,4,5]. Subclinical hypothyroidism (SCH) refers to serum thyroid stimulating hormone (TSH) levels higher than the reference range and levels of free thyroxine (fT4) remaining in the range [1, 6]. Overt hypothyroidism refers to TSH levels being elevated and fT4 levels below their respective reference ranges [7].

Cardiovascular disease remains the leading cause of death worldwide, with ischaemic heart disease and stroke accounting for 8.9 million and 6.1 million deaths, respectively, in 2019 [8]. Each year there are 160,000 deaths in the UK attributed to cardiovascular events [9], accounting for approximately 23.9% of deaths in the UK [10]. Therefore, unnecessary cardiovascular events must be reduced in the population. In addition, the UK population is increasing and ageing; the Office for National Statistics has predicted that by 2043, the number of people aged over 85 will have nearly doubled to 3 million [11]. Due to the ageing and growing population in the UK, a rise in pharmaceutical prescribing can be expected. In 2019, the National Health Service (NHS) distributed 32,956,754 prescriptions of LT4. This is a significant increase compared to only 20,426,378 LT4 prescriptions distributed in 2008 [12]. Moreover, the estimated number of individuals prescribed LT4 in the UK has increased from 2.8 to 4.1% over 11 years. Reducing the number of patients prescribed LT4 will not only potentially improve patient outcomes and quality of life in the ageing population but will reduce prescription wastage, hence cost to the NHS.

Mildly elevated TSH levels become more prevalent as people age [13]. The National Health and Nutrition Examination Survey, based in the United States (US), examined 16,533 people who were carefully selected to represent the general US population without thyroid disease and demonstrated that a significant proportion of older adults had high serum TSH (> 4.5 mIU/L) [4]. In addition, the Thyroid Epidemiology, Audit, and Research Study (TEARS) found that, based on a population of 153,127 people in Scotland, the 97.5th centile for TSH concentration steadily rose within age deciles, to an upper limit of 5.9 mIU/L in the over 90-year-olds [13]. This suggests that it is normal for the elderly population to have a TSH level within the range of 0.4–5.9 mIU/L rather than within the range of 0.4–4.0 mIU/L that is typically used. Furthermore, the longitudinal all-stars study, on 843 participants not taking thyroid medication, indicated that the 97.5th percentile TSH concentration could be as high as 8mIU/L in over 90-year-olds [14]. Longitudinal studies have also shown increased TSH levels in the same individuals over time, with no significant changes in fT4 concentrations [14,15,16]. Therefore, there is significant evidence to suggest that TSH concentration rises with age.

Slightly elevated TSH levels may be beneficial to the older person. Simonsick et al. [17] demonstrated better mobility and cardiorespiratory fitness functioning amongst adults over 70 years, with a mildly elevated TSH between 4.5 and 7.0 mIU/L compared to those with TSH levels of 0.4–4.5 mIU/L [17]. In the Leiden 85 + Study, 599 participants over 85 years old were followed up after four years. Those with a TSH level greater than 4.8 mIU/L had lower mortality than euthyroid individuals with a TSH level within 0.3–4.8 mIU/L; higher levels of fT4 were associated with increased all-cause mortality in the same cohort [18]. On the contrary, the Newcastle 85 + study found no association between higher TSH levels and all-cause mortality in a cohort of 643 patients aged 85 years followed up for 9 years [19]. In addition, a recent narrative review in 2021 found no unfavourable cardiovascular or musculoskeletal outcomes in patients with TSH levels between 4.5 and 7.0 mIU/L [6]. However, a case-control study found a significant mortality increase in patients prescribed LT4 aged over 65 with TSH below 10 mIU/L [20]. Much of the literature, including reviews that are powerful due to tertiary data, suggests that TSH levels between 4.5 and 10 mIU/L may benefit older adults.

However, TSH levels above 10 mlU/L have been found to be detrimental in all patients. A systematic review of 36 studies found an increased incidence of heart failure amongst patients aged over 65 years with TSH levels above 10 mlU/L [21]. Similarly, another systematic review of cohort studies found that in individuals with SCH and TSH levels above 10 mIU/L, there is an increased risk of coronary heart disease events, coronary heart disease mortality and heart failure events [22]. Current evidence indicates that TSH levels above 10 mlU/L can adversely affect older individuals. This is in line with the LT4 prescribing guidelines, stating that the majority of adults should only be prescribed LT4 with TSH levels above 10 mlU/L [23].

Previous studies that addressed the link between cardiovascular outcomes and LT4 treatment have failed to find a conclusive outcome. A cohort study on 1,642 patients aged over 70 years found no difference in cardiovascular events between treated and untreated groups [24]. The Thyroid Hormone Therapy for Older Adults with Subclinical Hypothyroidism (TRUST) study, a double-blind randomised placebo-controlled parallel-group trial of 737 adults aged 65 years and older with SCH, found no association between LT4 use and cardiovascular outcomes (hazard ratio 0.89; 95% confidence interval 0.47–1.69) [25]. The TRUST group concluded that LT4 should not be routinely prescribed to older people with SCH. Pooled results of the TRUST study and another randomised control trial found no significant difference in the cardiovascular risk of patients treated with LT4 versus those untreated [26]. However, both randomised controlled trials were underpowered due to small sample sizes. Most studies in older populations have not shown a clear association between SCH and cardiovascular events in patients aged over 65. The effect of mild thyroid failure on cardiovascular health becomes less evident with increasing age [27]. There is need for further large-scale studies investigating the impact of LT4 on cardiovascular health in patients over 50, particularly those aged between 50 and 65 years.

It is accepted that patients with overt hypothyroidism should be prescribed LT4 due to the association between the untreated condition and increased morbidity and mortality [28, 29]. However, the treatment of SCH is unclear due to a lack of reliable evidence in the past [30]. Over-treatment with thyroid hormones harms health, including poorer cardiovascular outcomes, atrial fibrillation, fractures, and osteoporosis [27, 31, 32]. However, over-treatment is common in older people and is associated with a longer time on replacement therapy and increasing age [33,34,35]. The European Thyroid Association guidelines recommend that SCH patients be prescribed LT4 when symptoms are shown and a TSH level is above 10 mIU/L [23]. It is also recommended that LT4 doses are personalised according to age, co-morbidities, and life expectancy [6]. However, research proves that these recommendations are not always followed [33].

The current literature suggests that older patients with mildly raised TSH concentrations may benefit from not being prescribed LT4 medication. The evidence is increasing and compelling to introduce age-specific TSH reference ranges for hypothyroidism regarding diagnostic thresholds and treatment targets to avoid patients being misclassified as having SCH and inappropriate treatment with thyroid hormones [13,14,15, 36, 37]. The distribution of TSH levels at the time of LT4 initiation has been described by Taylor et al., 2014 [33]. The median TSH when a patient is prescribed thyroid hormones has decreased from 8.7 mIU/L in 2001 to 7.9 mIU/L in 2009 in the UK, suggesting the likelihood a patient is prescribed LT4 treatment is increasing [33]. However, there has been a rising debate and emerging evidence suggesting that assessing borderline thyroid dysfunction may be more accurately characterised by fT4 levels rather than relying solely on TSH measurements [38, 39].

The literature is conflicting surrounding LT4 use and cardiovascular and bone health risk in the older population. A questionnaire answered by 408 individuals with SCH and taking LT4 found an increased risk of fractures in individuals over 50 years, particularly in the forearm [41]. It has also been found that the over-treatment of LT4 can increase the risk of cardiovascular morbidity, cardiovascular mortality and osteoporosis [42]. There is also conflicting evidence between the use of LT4 and mortality. The recent Baltimore Longitudinal Study of Ageing in 2021 found no association between all-cause mortality and LT4 use in over 65 year olds with normal TSH levels [43], while a much larger database study on 2,007,528 patients found that all-cause mortality was significantly higher in LT4-treated patients [44]. This suggests that LT4-based outcomes are dependent on both age and TSH levels.

The current research group has completed a National Institute for Health and Care Research (NIHR) Research for Patient Benefit study (Study of Optimal Replacement of Thyroxine in ElDerly (SORTED) – a feasibility study: Ref: PB-PG-0610-22139) and has demonstrated that older people can tolerate a lower dose of LT4 and hence higher TSH levels [45]. The work done in 2016 explored the relationship between the three variables: diagnosis of hypothyroidism, TSH measurement, and LT4 treatment. A feasibility analysis at one large rural general practice in Northumberland, England, with a practice list size of 11,752 found that 465 patients had been diagnosed with hypothyroidism. The prevalence of hypothyroidism at the practice was 4%, and the highest prevalence was in the 81–90 years age category, at 9.85%. The female-to-male ratio with hypothyroidism at the practice was 4.4:1.0, such that 379 females had been diagnosed. Despite 465 patients diagnosed with hypothyroidism (Read Coded), 622 patients received treatment with LT4. This is an excess of 157 patients receiving LT4 without a diagnosis of hypothyroidism. This may represent a proportion of SCH patients that are not Read Coded as hypothyroid but are receiving treatment. It is also possible that a small proportion of these patients are hyperthyroid patients receiving block and replace treatment or patients with pituitary disorders. Alternatively, some patients may report symptoms suggestive of hypothyroidism but still have normal serum thyroid function tests and receive LT4 therapy in contravention of accepted clinical practice and guidelines. For this reason, our search strategy centres upon the TSH values recorded in the database rather than a diagnosis. Between 2003 and 2015, 215 (1.9%) patients were recorded as having an age-specific index TSH as mildly elevated, in line with the TEARS study. Of those, 39% were untreated, and 61% were treated with varying degrees of treatment length. This feasibility work concludes that 53% of the treated patients may have been over-diagnosed due to their TSH level being within the range of 4.0 mIU/L and the upper TSH limit given by the TEARS study [13].