The mean age of participants in our study was 38.57 (± 12.56) years with a mean duration of hypothyroidism of 6.01 (± 4.86) years suggesting that the study participants were already on LT4 for a substantial period of time before recruitment. There was a female gender preponderance in our study group that conforms to the results from the Indian prevalence data showing higher prevalence of hypothyroidism in females as compared to males [1]. The gender distribution is also consistent with other studies which have looked at the prevalence of hypothyroidism [15]. Though a combination of subnormal FT4 and elevated TSH, usually above 10 µIU/mL, characterizes overt primary hypothyroidism, we took a TSH cut-off of 20 µIU/mL as TSH may increase up to 20 µIU/mL during the recovery phase of thyroiditis or non-thyroidal illnesses.

Injudicious PPI use is exceedingly common in clinical practice either prescribed by general physicians or as OTC agent, particularly for functional dyspepsia [16]. In India, pantoprazole is the most commonly prescribed PPI [17]. Of the available PPIs, lansoprazole and omeprazole have been found to interfere with LT4 absorption, and thus increase TSH in individuals who are biochemically euthyroid on stable dose of LT4. Two retrospective studies have documented a median increase of TSH by 0.69 (± 1.9) µIU/mL and 0.18 µIU/mL after 2–6 months of PPI intake in such patients (p = 0.035 and 0.001, respectively) [11, 18]. Retrospective review of the Italian general practice HSD also showed increased TSH levels with concomitant PPI intake (adjusted incidence ratio rate: 1.02; 95% CI: 1.01–1.03) [10]. Participants in our study were given daily PPI for initial 6 weeks and then for another 6 weeks following a washout period of 1 week. With morning dose of pantoprazole TSH rose from 2.70 µIU/mL to 3.78 µIU/mL, and with night dose TSH increased from 2.20 µIU/mL to 3.30 µIU/mL during the initial 6 weeks. We documented that daily intake of PPI even for 6 weeks also lead to significant impairment in LT4 absorption irrespective of time of administration, as evidenced by a TSH elevation of 1.07 µIU/mL in group M, 1.10 µIU/ mL in group N and 1.09 µIU/ml in the entire study population. With longer duration of PPI therapy (12 weeks), the changes were even more pronounced. The mean increment in TSH from baseline was 1.06 µIU/mL in M group, 2.33 µIU/mL in N group and 1.78 µIU/mL overall at the end of the study (Table 2; Fig. 3A and B).

Co-prescription of LT4 and usual dose of PPI also leads to clinically significant elevation in TSH due to inadequate LT4 absorption, which resulted in 35–37% higher dose of LT4 to achieve target TSH in different cohorts of patients with euthyroid nodular thyroid disease or hypothyroidism [6, 18]. However, we did not look into the percentage increase in LT4 dosage to achieve biochemical euthyroidism in our patients.

The effect of PPIs on LT4 absorption, however, has been discordant. While some studies documented reduced absorption of LT4 tablet with lansoprazole (30 mg daily), omeprazole (40 mg per day) or intravenous esomeprazole (80 mg) [6, 18, 19], others did not find similar effects with omeprazole, esomeprazole or pantoprazole. Potential dose dependent reduction in LT4 absorption with two different doses of omeprazole (20 mg and 40 mg) were studied in primary hypothyroid subjects who were biochemically euthyroid on LT4 tablets for at least 1 year (n = 19). There was no statistically significant difference in TSH levels before and 3 months after 20 mg (median levels: 2.24 vs. 2.42 µIU/ml, p = 0.62) or 40 mg (median level: 2.28 and 2.30 µIU/mL, p = 0.82) of omeprazole therapy [20]. Similar observation was noted in serum thyroid hormone concentrations after 600 µg of LT4 administration following 1-week therapy with esomeprazole (and famotidine) [21]. However, this study was biased by a short sampling time (8 hour), which is much lower than the half-life of LT4, making a reliable assessment of the pharmacokinetic profile somewhat difficult. The differences in dose, route of administration or in PPI molecule might explain, at least partly, this discrepancy in TSH changes. Though gastric acid suppression is attributed to a class effect of PPIs, differences in potency and pharmacokinetic profile among different molecules may lead to variable effect on LT4 absorption. We used a single dose of 40 mg of pantoprazole. A crossover study involving 20 healthy subjects, however, found that pantoprazole had no effect on LT4 absorption. TSH was measured following a single dose of LT4 (4 µg/kg) at baseline and after 1 week of 40 mg pantoprazole therapy [13]. An important drawback of both these studies (esomeprazole & pantoprazole) was the short course (1 week) of PPI therapy. Steady state is usually achieved after 4–6 weeks of LT4 therapy; hence, any alteration in circulatory TSH concentration may not become apparent after 1 week of PPI therapy. Though a plateau in gastric acidity has been described following 4 days of PPI use in some studies, and suppression of gastric acid secretion was documented through increased gastrin level in the pantoprazole study, the investigators used a very high dose of LT4 (4 µg/kg), which is much higher than the usual dose required in athyreotic individual (1.6–1.8 µg/kg/day) [13]. Such an unusual high dose may overcome PPI interference on LT4 absorption. Long-term PPI treatment is likely to alter gastrointestinal transit time, mucosal architecture, and local milieu along with alteration in enterohepatic metabolism of orally administered LT4. The other potential confounders of this study were short sampling time (10 hour), which is insufficient to alter TSH, and to accurately determine T4 kinetics profile, and dose of LT4; a dose of 600 µg (which is much higher than 4 µg/kg) is needed for accurate assessment of LT4 pharmacokinetics and precise discrimination in potential differences in LT4 absorption. We did not find any difference between time of administration of pantoprazole as far as increase in serum TSH were concerned (Table 3). Serum TSH concentrations were equally elevated both with morning and bed time dose of pantoprazole.

Oral pantoprazole has a bioavailability of 77% and its absorption is not affected by food or antacids. It is absorbed from the small bowel, resulting in a maximum serum concentration 2 to 3 hours post-ingestion. Unlike other PPIs, the serum concentration of pantoprazole is not dose-dependent; serum concentration after the first dose is similar to that following multiple doses. The metabolism of pantoprazole is independent of route of administration, with a half-life of approximately 1 hour [22]. Gastric acid suppression is a class effect of the PPIs. Though the half-life of the available PPIs (other than tenatoprazole) is 1–2 hours, the duration of acid inhibition is almost 48 hours because of irreversible binding to the H+K+-ATPase [23]. Once the steady sate is achieved, the duration of action of pantoprazole lasts for about a week. Duration of acid suppression with ranitidine, a H2RB is 4–6 hours. LT4-treated patients, who require gastric acid suppressive therapy, thus, should better be offered H2RB after 3–4 hours of LT4 administration to avoid elevation in TSH. PPI use even for 6 weeks, irrespective of timing of administration, is likely to result in increment in ongoing LT4 dosage to maintain the TSH level.

Intra-individual fluctuation in serum TSH is physiological and differences in TSH values, measured in two different time points does not necessarily indicate abnormal test result. Individual reference ranges of TSH is much narrow than the population-specific reference intervals, and it has been suggested that changes in serum TSH of as low as 0.75µIU/ml (range: 0.2–1.6 µIU/ml) may take a particular individual out of their ‘personal reference range’ [24]. However, a ‘true difference’ between two TFTs is said to be present if there is 40% variation in TSH and 15% in FT4 [25]. It needs to be remembered that these cut-offs were derived from longitudinal follow-up of patients with stable untreated subclinical hypothyroidism for 1 year. The intra-individual variations in TFT in patients with LT4-treated overt hypothyroidism within 13 weeks are expected to be much lower. Following lansoprazole initiation, 19% patients noticed TSH of more than 5 µIU/mL, which necessitated LT4 dose increment [18]. In our cohort, TSH went up the upper limit of the reference range (and also more than 5µIU/ml) in 26.7% after 6 weeks, and in 33.3% after 12 weeks of pantoprazole therapy. In addition, a large proportion of our cohort (40% at 6 weeks and 50% at end of study) had increment in TSH, beyond the upper limit of expected intra-individual variation, described earlier. Though the findings were not statistically significant, our study was not powered to look for such changes in serum TSH concentrations. Duration of PPI therapy could possibly result in ‘clinically meaningful’ elevation in serum TSH as well.

Interestingly, absorption of LT4 formulations, other than tablet, seems unaffected by PPIs. LT4 absorption kinetics were assessed following intravenous esomeprazole (80 mg), and coadministration of 600 µg of LT4 either in tablet form or in gel capsule form [19]. Under esomeprazole infusion, Cmax of LT4 decreased by 13.0% with the tablet, but only 2.6% with the gel capsule. Gelatine coating of the gel capsules possibly helped in absorption and obliviated the need for low intra-gastric pH [19].

Compared to tablets, absorption of liquid formulation of LT4 seems to be unaffected by simultaneous PPI use. In an earlier study, LT4 dose requirement decreased when patients were switched to liquid formulation from tablets while on any of the 4 different PPI molecules (omeprazole, pantoprazole, lansoprazole, esomeprazole) [14]. As discussed earlier, aqueous form of LT4 is required for its effective absorption. Impaired gastric acid secretion following PPI use interfere with the aqueous formation of administered LT4 tablets resulting in reduced absorption. Gel-capsule and liquid formulations of LT4 do not require low intra-gastric pH for absorption, hence unaffected by coadministration of PPI. However, these formulations are not widely available.

This study was conducted in a prospective, 2-arms, randomized, cross-over pattern, which is inherently a strong study design. To the best of our knowledge, this is the first prospective study in the literature, which assessed the effect of PPI administration in two different point of times on change in TSH concentration in a cross-over pattern. We opted for a very strict inclusion criteria as far as variation in TSH was concerned (variation not more than 1.6 µIU/mL or 40% whichever is lower for at least 6 months). The baseline characteristics were similar between both the groups. Furthermore, medications which are known to affect LT4 absorption or metabolism were excluded; thus, removing any potential confounder. We ensured 100% compliance both for PPI and LT4. In addition, we tried to look at ‘clinically meaningful’ changes in serum TSH concentrations with PPI use.

Our study had certain limitations. Due to the Covid-19 pandemic, we could recruit only 30 patients. This may have been enough to render the study adequately powered; however, more number of participants should have been enrolled to provide further strength. Four patients were lost to follow-up in group M, which led to a female gender predominance in group M. This was addressed by assessing the changes in TSH between the two groups after removing all male participants in both the groups. The results from this analysis were concordant with our overall findings. The study participants were under strict vigilance of the investigators and their relatives to ensure compliance with study medications (LT4 and pantoprazole). This may have inadvertently led to the Hawthorne effect leading to enhanced adherence. This also explains decreased in TSH concentration from baseline in some of the participants. Such strict adherence is often not seen in clinical practice. A placebo arm to the cross-over design could have further strengthened the study. However, we could not arrange for ‘matching placebo’ tablets. One potential limitation might be a short wash-out period of 1 week for pantoprazole. The duration of acid inhibition following PPI use is 48 hours; one week, thus, seems sufficient for wash-out [23]. Another limitations of the study is single centre recruitment of patients. We did not assess anti-thyroid antibodies in the study participants. Patients with chronic lymphocytic thyroiditis may require progressive increment of LT4 doses during follow-up, as remaining thyroid function gradually declines. However, the duration of hypothyroid was 6 years before recruitment and participants had stable TSH during the preceding 6 months. It is unlikely that the residual gland function declined during that 13 weeks of the study resulting in increased TSH with stable dose of LT4.