Drugs That Interact With Levothyroxine

An Observational Study From the Thyroid Epidemiology, Audit and Research Study (TEARS)

Savannah A. Irving; Thenmalar Vadiveloo; Graham P. Leese


Clin Endocrinol. 2015;82(1):136-141. 

In This Article


This study shows the importance and magnitude of drug interactions with oral thyroxine at a population level and adds to the previous information from small select case series. In addition, it has demonstrated interactions with drugs not previously noted, for example PPIs. Iron, PPIs, calcium and oestrogens co-administered with thyroxine caused increased TSH concentration possibly due to decreased thyroxine absorption or through interfering with drug metabolism. It is known from previous studies[9–13] that some of these drugs may affect thyroxine absorption and/or metabolism. The median change in population TSH was small and nonsignificant (−0·005–0·7 mU/l) and unlikely to be important for the majority of patients. The relative stability of TSH levels over the course of a year prior to starting a study drug serves as a control for this patient population and leads to the conclusion that the change in TSH during study drug administration is due to the study drug affecting thyroxine.

For patients on a steady dose of thyroxine for 2 years, some experienced an increase in serum TSH of over 5 mU/l, which is likely to be of clinical significance as this places them into a biochemically hypothyroid state. At least 7·5% of patients taking iron, 4·4% of those on calcium and 4·3% of those on oestrogen experienced this clinically significant rise in serum TSH. Our population-based study shows that this is a clinically significant problem that could have public health issues for some patients.

Strengths of the study include the large cohort of patients (n = 6482), concurrently treated with thyroxine and a study drug in a clinical scenario, which makes the results clinically relevant. The number of patients included in this study is much greater than previous studies. Other studies used healthy volunteers or small numbers of patients, and this many not adequately represent the population of patients on chronic thyroxine therapy. A clinical study of patients could be undertaken to show direct causality between a study drug and thyroxine absorption. This should be followed by pharmacokinetic trials and case-controlled studies. Weaknesses of the study include the retrospective nature of analysis. We were not able to address patient compliance issues and so do not know whether patients consistently took their tablets at the same time or in the same way. Patients with primary hypothyroidism are more likely to develop coeliac disease and pernicious anaemia which could affect thyroxine absorption and affect gastric acid concentrations. We were unable to correct for these confounding variables, but if they were to influence the results, it would have been expected to affect all drugs studied and not just a selection of them. These issues would have affected the patient population prior to initiation of a study drug, but there was no significant change in TSH during this control interval. There may be other drugs and over the counter medications that we have not examined that interact with thyroxine, such as coffee[6,20] and other products that are impossible to measure in a population-based study. In addition, the time of day thyroxine was ingested and the serum TSH measurements were taken was not standardized such that we did not account for diurnal variation. It is, however, unlikely that this would have introduced a consistent bias.

Iron is known to cause a reduction in thyroxine absorption, and guidelines[19] recommend taking iron at least 2 h after thyroxine administration. Calcium and antacids are also listed[19] as drugs that affect the absorption of thyroxine. Our study of current clinical practice suggests that either this advice is ignored or it is inadequate to prevent drug interaction for a significant number of patients.

Previous studies that have looked at the effect of PPIs on thyroxine absorption have been inconclusive. Pabla et al.[14] used an in vitro model to show that dissolution of thyroxine decreased with an increase in pH, suggesting a link between patients on concurrent thyroxine and PPI therapy. Further studies[15,16] concluded that both omeprazole and lansoprazole therapy increased serum TSH in patients also treated on thyroxine. Two studies[17,18] noted no difference in thyroxine absorption after PPI administration. These studies concluded that acidity of the stomach does not play a role in absorption of thyroxine, but rather other factors, such as competition and adsorption processes, may affect absorption. These studies were potentially flawed as they studied healthy volunteers,[18] the volunteers were given larger doses of thyroxine than are normally prescribed, and they were also not on concurrent therapy for prolonged periods of time as is normal in clinical settings. Our study suggests that there is an interaction between PPIs and thyroxine.

PPIs may affect the absorption of thyroxine by increasing the pH of the stomach, but there is currently a paucity of conclusive evidence to substantiate this theory. Studies[21,22] have indicated that PPIs may interact with other drugs, such as clopidogrel, due to shared metabolism by cytochrome P450. The clinical significance of this interaction[23] is still not fully understood as many patients taking PPIs also have comorbidities. Studies into the mechanism of PPI interactions with thyroxine warrant further consideration as suboptimal dosing can exacerbate patient symptoms and quality of life, resulting in poor management of hypothyroidism. For 5·6% of patients concurrently taking thyroxine and a PPI, the interaction is clinically significant, causing an increase in serum TSH of over 5 mU/l and placing them back into a biochemically hypothyroid state.

H2 receptor antagonists did not have a statistically significant effect on thyroxine absorption. This could be explained because the antagonists work by a different mechanism than PPIs. H2 antagonists also increase stomach pH, and this finding supports the likelihood that PPIs interact with thyroxine by mechanisms other than lowering stomach acidity. Glucocorticoids and DMARDs also did not cause a statistically significant change in TSH levels. This indicates that not all drugs affect thyroxine levels.

A 2011 study of (N = 41)[24] suggested statins did not have any effect on levothyroxine absorption, but patients were only followed for a period of 3 months. Interestingly, in this population-based study, statins caused a statistically significant decrease in serum TSH levels. Metformin has previously been shown to have this effect, but the mechanisms are unclear.[25,26] Further studies could examine whether metformin causes a statistically significant decrease in TSH in a large patient population in comparison with the decrease in TSH caused by statins. The mechanism of action causing this decrease could also be further investigated.

Suboptimal treatment can directly affect quality of life, and it may be important to monitor patients more closely when they are taking thyroxine and another drug is initiated, especially iron, calcium oestrogens or PPIs. Our study findings support the conclusion of other studies[16] that PPI therapy should be added to the list of medications affecting a patient's level of thyroid hormone in those patients being treated chronically with thyroxine.

For patients on long-term thyroxine, it may be best that it is swallowed either one hour before or four hours after food or concomitant iron and calcium[8,27] to ensure that thyroxine absorption is not affected by co-administration. This may also be the case for PPIs. If PPIs and oestrogen mainly affect thyroxine hepatic metabolism, then when these drugs are prescribed and a patient is on long-term thyroxine, the TSH needs to be carefully monitored with consideration to increasing the thyroxine dose.