Since the 1970s levothyroxine (LT4) has become the standard of care for thyroid hormone replacement in subjects unable to produce their own thyroid hormones due to congenital, autoimmune or iatrogenic causes. Levothyroxine has become one of the most widely used drugs worldwide, it is the most commonly prescribed drug in the United States (US), the third most in the United Kingdom (UK), and there is evidence that its use is steadily increasing [1, 2]. Despite this, there remains controversy as to the best way to therapeutically replace thyroid hormones, with a small group of people on LT4 monotherapy not feeling as though they have achieved their premorbid well-being [3, 4] and hence a significant push for other therapies including combination LT4/liothyronine (LT3) therapy, LT3 alone or extracts from animal thyroids.

Possible reasons for some patients not responding to levothyroxine monotherapy and evidence for efficacy of mono- and combination therapy have been well covered in recent systematic reviews by both the American Thyroid Association (ATA) [5] and European Thyroid Association (ETA) [6]. Both guidelines suggest that in select cases, a trial of therapy including LT3 can be considered, carefully supervised by an expert in the field. The purpose of this review, therefore, is not to repeat the analysis in these recent reports, but to provide a more practical approach to the use of combination therapy in clinical practice including safety aspects and cost, where a trial is considered appropriate.

The target audience for this review is clinicians who incorporate or would consider incorporating trials of combination LT4/LT3 therapy in their practice, or who have patients who have questions about this therapy. The review was performed by literature search in each area, review of published articles and studies and practical experience.

Although the thyroid gland produces T4 and T3, LT4 monotherapy has been the mainstay of thyroid hormone replacement since the 1970s replacing desiccated thyroid extract (DTE) which had been used for many years prior. This is because it is easily administered, well absorbed by the oral route and its long half-life allows for once daily dosing with very stable serum levels. It was also shown to be converted to T3 within the body [7] alleviating the need to add LT3 which does not have the same stable pharmacological profile. Furthermore, it has also been shown that the majority of circulating T3 comes from peripheral conversion of T4 to T3 and not secretion of T3 from the thyroid [8], hence a T4:T3 secretion ratio of approximately 14:1 appears average in humans, suggesting only a small role for secreted T3. However, there is a small group of patients who do not feel back to their euthyroid well-being despite having thyroid function tests suggestive of adequate replacement on LT4 [3, 9, 10]. There are also several studies showing that on LT4 monotherapy serum T3 levels are significantly lower for the same TSH in euthyroid patients [11–16], although the clinical significance of this is unknown. Another study showed it was not possible to normalise serum TSH, T3 and T4 levels or tissue T3 levels in laboratory animals giving them LT4 monotherapy [17].

Because of this there have now been at least 13 randomized controlled trials (RCT) comparing efficacy of combination LT4/LT3 therapy versus LT4 monotherapy for thyroid hormone replacement [18–30]. There has also been one trial comparing DTE to LT4 monotherapy [31]. The trials differ significantly in study population size, method of substituting LT3 for LT4, dose, length of study and outcomes measured. There have now been 4 systematic reviews/meta-analyses of these studies attempting to clarify the findings [32–35]. Overall these meta-analyses have suggested there is no significant benefit of combination LT4/LT3 therapy compared to LT4 monotherapy in terms of mood, health-related quality of life or cognitive function. This provides reasonable evidence that at a population level there is no benefit of using combination therapy over monotherapy. However, the different methods of replacing LT4 with LT3 (some resulting in raised TSH suggestive of under replacement) and small size and hence power of some studies should be considered. There was no increase in adverse events in the combination groups in this admittedly short period of follow up. There is one randomized cross-over trial comparing DTE with levothyroxine monotherapy in 70 patients for 16 weeks each arm [31]. This did not show any significant benefit of DTE over LT4 in symptoms and neurocognitive markers. There was an increase in preference for DTE over LT4 however it is unclear what this signifies given an overall lack of improvement in symptoms. The study was too small to pick up subgroups responding to T3 and future studies in this area may consider having a synthetic T4/T3 arm also to determine if any benefit is due to the addition of T3 or DTE itself.

Whilst this evidence is convincing there remains a possibility that a small subset of hypothyroid patients will do better on combination therapy; this group may get lost in the larger group of patients with no benefit. Practicing clinicians will be able to identify a group of patients not satisfied on LT4 monotherapy which makes up a small subset of all their patients on LT4. There is also one study suggesting that a polymorphism in the DIO2 gene which codes for the deiodinase 2 enzyme (Thr92Ala), important for conversion of T4 to T3 in many tissues including the brain may suggest the group which will respond to therapy (rare homozygotes approximately 12% of the Caucasian population) [36]. These findings remain controversial however, particularly with inconsistent findings of the functional effect of this polymorphism [37–39].

After extensive review of available literature on this topic guidelines were produced by both the ATA and ETA related to thyroid hormone replacement with LT4 and alternatives. Both guidelines suggest there is insufficient or no strong evidence of superiority of combination LT4/LT3 therapy over LT4 monotherapy in patients with hypothyroidism. However, both acknowledge that there are patients who have persistent symptoms or sub-optimal health despite LT4 therapy guided by normal thyroid biochemistry and suggest possible reasons for this include inadequacy of LT4 monotherapy to normalise serum and tissue T4 and T3 levels. Thus, both suggest that in an appropriate clinical setting (see below) combination therapy may be trialled to determine if it is beneficial for the individual patient [5, 6]. The remainder of this review will consider practical issues related to such a trial/long term therapy.

As in all clinical practice, clinicians should only offer treatments that they are comfortable using. The experience is, however, that patients wishing for a trial are often frustrated at the lack of specialists willing to provide such, and therefore often obtain combination therapy from health practitioners with little training in the area, who do not monitor for complications, can give incorrect doses/dose ratios and offer ongoing treatment without assessing the benefits. In some instances, patients self-treat with medications obtained online. Until more is known of benefits and risks, LT3 should not be offered routinely. LT3 should only be trialled in pts. who specifically request it, who have marked persisting symptoms, and who fully understand and accept (with written/documented information if possible) the unknown potential for long-term harm. Endocrinologists are clearly best suited to provide this.

Patient selection is important to determine pre-trial whether the patient is likely to gain benefit from the treatment and prevent harm to the patient. There are circumstances in which a trial may delay another treatable diagnosis or put the patient at significant risk without possibility of benefit in which it may be inappropriate.

Studies looking at the pharmacology of LT3 replacement all show a significant peak of serum T3 2–4 h after dose and wearing off after 12 h in those on a single daily dose, these include hypothyroid patients on combination therapy [63] (Fig. 1), LT3 monotherapy [64, 65], and even euthyroid subjects taking LT3 only [66]. The profiles are very different to those in patient with normal endogenous thyroid function, and depending on the dose of LT3 the peak level is often above the reference range, and/or the serum TSH is raised/suppressed compared to T4 treatment [67]. Figure 1 suggests that to sample the peak serum T3 level a test taken 2–4 h after ingestion of the T3 would be appropriate. There is a study showing equivalent TSH responsiveness to TRH in patients on either T3 or T4 monotherapy [68], suggesting the rise in TSH seen in some of these studies is an indication of under replacement. Almost all the combination trials used once daily LT3 dosing with one using twice daily.

To truly mimic the normal production of T3 patients would have to split the dose of T3 and take it two or three times a day, however, the large dose size of the available products may preclude this in patients with a lower requirement for thyroid hormone. Comparison of the figures from Saravanan et al. [63] and Russell et al. [62] would suggest that to mimic serum T3 levels in euthyroid individuals the LT3 dose should be split with the second dose given approximately 8 h after the first. This may prevent the insomnia reported by some patients when they take LT3 prior to bed which is presumably secondary to a serum T3 peak whilst trying to sleep. What effect this will have on the steady state of the drug is unclear as there are no studies looking at serum levels in multiple daily dosing, and even the half-life of T3 is debated with a wide range of opinions. Saravanan et al. did not find any difference in cardiovascular parameters (pulse rate and blood pressure) between their groups on T4/T3 and T4 only despite the clear peak in T3 levels at approximately 4 h. This interesting finding reminds us that serum hormone levels do not necessarily reflect action in tissues due to the presence of thyroid hormone transporters and deiodinases in different tissues which may influence the effect of these hormones in individual tissues [69]. Furthermore, most actions of thyroid hormone take several hours to have effect as they require the synthesis of new mRNA and protein.

Given the relatively small number of patients using combination T4/T3 therapy safety data has until recently been lacking. None of the 13 RCTs comparing combination therapy to T4 monotherapy showed any increase in adverse events in the combination group, however the follow up was generally short ranging from 5 to 52 weeks. Leese et al. have recently reported safety data in their observational cohort in the TEARS study [42]. In this study a subgroup of patients on thyroid hormone placement were considered with outcomes for those who had ever been taking LT3 (n = 400) compared to those who had only ever taken LT4 (n = 33,955). The study showed no increase in cardiovascular disease, atrial fibrillation or fractures (outcomes previously shown to be associated with levothyroxine overtreatment [78]) with a median follow-up of 9 years. There was an increased risk for new prescriptions of anti-psychotic medications of unclear significance. There was also a possible association with an increased risk of breast cancer in subjects who had taken LT3. Although there is some data on T3 augmentation of breast cancer cell line proliferation and higher endogenous T3 being associated with more aggressive breast cancers, the authors point out that this association was of borderline statistical significance and not related to number of prescriptions of LT3, arguing against a causal relationship. This potential association should be assessed in other large cohorts. The data from Leese et al. are from an observational study and so carry a risk of bias and self-selection; in addition the duration of LT3 use was not quantified. However, this may be the largest and longest duration study of its type for some time and offers some reassurance that the risks of taking T3 are not greater than expected.

Many patients and alternative physicians prefer to use forms of dessicated thyroid extract (DTE) for thyroid hormone replacement. DTE is described as “the cleaned, dried, and powdered thyroid gland previously deprived of connective tissue and fat. It is obtained from domesticated animals that are used for food by humans” by the United States Pharmacopeia. Extracts of animal thyroids have been used for hypothyroid symptoms for many centuries in different cultures, and as a form similar to what is used today for over 110 years [79]. Most formulations are from porcine thyroid; however some from bovine thyroid or a combination of the two exist. Initially it was titrated to improvement in clinical symptoms and avoidance of hyperthyroid symptoms. The development of Levothyroxine and ability to assay serum TSH, T4 and T3 levels in the 1980’s led to calls for removal of DTE for treatment of hypothyroidism because of supraphysiological levels of T3 post dose, hyperthyroid symptoms and complications, and fluctuating levels of T3 as compared to very stable TSH, T3 and T4 levels with levothyroxine dosing [80–83]. Despite continuing concerns about these issues and also the consistency of the various preparations it continues to be used. Whilst the dramatic cases of thyroxtoxicosis on these preparations are mainly historical, there remain concerns about frequency of adverse events and calls for greater standardization of these preparations [84].

Amour thyroid is the most commonly used formulation, also the most expensive, and in the US costs around $1 per 60 mg grain.. Price varies across the UK mostly around £1 per grain. These formulations are not available in Australia and therefore most users import them from the US at similar cost plus postage.

Although there is convincing evidence that there is no benefit of combination T4/T3 therapy over T4 monotherapy for management of hypothyroidism at a population level, there remains a population of patients who do not feel well on T4 monotherapy. There are several possible reasons for this, one of which is an inability to use T4 effectively in a group of patients who may respond better to combination T4/T3 therapy. This will remain a possibility until large RCTs in an appropriately targeted population can confirm or refute it, and whilst it does a trial of combination therapy in such patients may be indicated. It is important that this be performed by a clinician with adequate knowledge and experience in the area, with appropriate patient selection, clear explanation of risks and benefits to the patient for consent and careful monitoring and follow-up. Possible ways to do this are covered in this review although it is clear that further research into this area and possible methods of delivering T3 are required.