Original Article
EFFECT OF PROTON PUMP INHIBITORS
ON SERUM THYROID-STIMULATING HORMONE LEVEL
IN EUTHYROID PATIENTS TREATED WITH
LEVOTHYROXINE FOR HYPOTHYROIDISM
Issac Sachmechi, MD, FACE, FACP,1,2 David M. Reich, MD, FACE,1,2
Michael Aninyei, MD,2 Francisca Wibowo, MD,2 Garima Gupta, MD,2
and Paul J. Kim, MD, FACE1,2
ABSTRACT
Objective: To examine retrospectively the effect of
proton pump inhibitors (PPIs) on thyrotropin (thyroid-
stimulating hormone or TSH) values in patients with
hypothyroidism and normal TSH levels receiving levothy-
roxine (LT4) replacement therapy.
Methods: The data collection was done by retrospec-
tive review of electronic medical records from the period
of December 2002 to August 2005 from patients with
hypothyroidism who were receiving at least 25 µg of LT4
replacement daily at Queens Hospital Center. The first 92
patients meeting all inclusion and exclusion criteria were
included in the study. The study group (N = 37) patient
data were collected by selecting euthyroid patients who
had received stable LT4 replacement for at least 6 months
and in whom PPI therapy (lansoprazole) was later initiat-
ed. TSH levels were collected before and at least 2 months
after the PPI treatment was started. The control group (N
= 55) patient data were collected by reviewing TSH levels
among euthyroid patients with a history of hypothy-
roidism receiving stable LT4 therapy and not receiving a
PPI during the period of data collection. The statistical
analysis was done by comparing the mean change in TSH
level in each group with use of the Student t test.
Results: In the study group, the mean change in the
TSH level from before to at least 2 months after initiation
of PPI therapy, 0.69 ± 1.9 µIU/mL, was statistically sig-
nificant (P = 0.035). In the control group, the mean change
in the TSH level during the study period, 0.11 ± 1.06
µIU/mL, was not statistically significant (P = 0.45).
Conclusion: To our best knowledge, this is the first
study in humans with hypothyroidism demonstrating the
Submitted for publication July 12, 2006
Accepted for publication December 12, 2006
From the 1Division of Endocrinology and the 2Department of Medicine,
Queens Hospital Center/Mount Sinai School of Medicine, Jamaica, New
York.
Address correspondence and reprint requests to Dr. Paul J. Kim, Division
of Endocrinology, Department of Medicine, Queens Hospital Center/
Mount Sinai School of Medicine, 82-68 164th Street, Jamaica, NY 11432.
© 2007 AACE.
ENDOCRINE PRACTICE Vol 13 No. 4 July/August 2007 345
effect of PPIs on serum TSH levels. PPIs should be added
to the list of medications affecting the level of thyroid hor-
mone in patients with hypothyroidism treated with LT4
replacement. Patients with hypothyroidism and normal
TSH values during LT4 replacement therapy may need
additional thyroid function testing after treatment with
PPIs and may need adjustment of their LT4 dose. (Endocr
Pract. 2007;13:345-349)
Abbreviations:
LT4 = levothyroxine; PPIs = proton pump inhibitors; T3
= triiodothyronine; T4 = thyroxine; TSH = thyroid-stim-
ulating hormone (thyrotropin); UGT = uridine diphos-
phate-glucuronosyltransferase
INTRODUCTION
Hypothyroidism is a common medical condition that
is easily treated with thyroid hormone replacement, and
thyrotropin (thyroid-stimulating hormone or TSH) is the
laboratory study that is monitored to assess the adequacy
of replacement.
Several medications are known to affect the level of
serum thyroid hormone in patients with hypothyroidism
receiving levothyroxine (LT4) replacement, either by
reducing gastrointestinal absorption of LT4 (1-3) or by
increasing its metabolic clearance (4). Proton pump
inhibitors (PPIs) are commonly used to treat conditions
such as gastroesophageal reflux disease and Helicobacter
pylori infection. In a recent study, the dose of LT4 required
to obtain suppressed levels of TSH in patients with multi-
nodular goiter was higher in patients with H. pylori-relat-
ed gastritis and atrophic gastritis in comparison with the
patients without H. pylori-related gastritis and atrophic
gastritis (5). In the same study, a small subgroup of
patients with multinodular goiter and gastroesophageal
reflux who were treated with omeprazole and LT4 sup-
pression required a 37% increase in dose of LT4 to achieve
suppressed levels of TSH (5). In addition, a study in rats
showed that PPIs increased metabolic clearance of LT4
(6).
346 Protein Pump Inhibitors and TSH, Endocr Pract. 2007;13(No. 4)
We have observed an apparent increase in TSH levels
in patients with hypothyroidism receiving LT4 replace-
ment after initiation of therapy with a PPI. The aim of the
current study was to perform a retrospective analysis of
the effect of lansoprazole, a PPI, on TSH levels in patients
with hypothyroidism who had normal TSH levels while
receiving LT4 replacement.
PATIENTS AND METHODS
Subjects and Study Design
Patient Population
We collected our data using electronic review of med-
ical records, which enabled us to access data regarding
patients who were treated for primary hypothyroidism at
Queens Hospital Center, Jamaica, New York, during the
period between December 2002 and August 2005.
Approval for the study was granted from the Institutional
Review Board of Queens Hospital Center and Mount Sinai
School of Medicine. The first 92 patients meeting all
inclusion and exclusion criteria (see subsequent material)
were included in the study.
Inclusion Criteria
The following factors were required for inclusion in
the study:
1. Patients with primary hypothyroidism who had
TSH levels assessed at the beginning of the study
period and between 2 and 6 months later and who
had had a stable TSH level (0.34 to 4.9 µIU/mL)
for at least 6 months before the baseline TSH level
collection
2. Patients with primary hypothyroidism who had
been taking a stable dose of LT4 (between 25 µg
and 200 µg daily) for at least 6 months without a
change in LT4 dose
Exclusion Criteria
The following factors disqualified subjects from
inclusion in the study:
1. Patients with a history of depression or other psy-
chiatric disorders, pituitary disorders, cirrhosis,
chronic renal insufficiency with nephrotic syn-
drome, hospitalization, or corticosteroid use during
the data collection period
2. Patients receiving medications known to affect thy-
roid hormone metabolism or absorption during the
study period, including 6 months preceding the
baseline TSH measurement
3. Patients with a documented history of noncompli-
ance or substance abuse
The study group consisted of 37 (30 female and 7
male subjects) of the 92 patients fulfilling the inclusion
and exclusion criteria, who had been treated with lanso-
prazole (30 mg daily) during the study period. These
patients had TSH levels measured before and between 2
and 6 months after lansoprazole therapy was started. The
remaining 55 patients (45 female and 10 male subjects)
(not treated with lansoprazole) were assigned to the con-
trol group.
Assay
Serum TSH levels were measured by a hypersensitive
2-site immunoenzymatic assay (Beckman Coulter, Inc)
(reference range, 0.34 to 5.50 µIU/mL). This assay
exhibits performance consistent with the definition of
“third-generation functional sensitivity” (0.01 to 0.02
µIU/mL, with an interassay coefficient of variation of 20%
or less) for TSH measurement.
Statistical Analysis
The statistical analysis was performed by comparing
the mean change in TSH level in each group by means of
a paired or 2-sample t test with use of the Minitab system,
version 14.
RESULTS
The baseline variables in the 2 groups, including
mean TSH levels, age, body mass index, and LT4 dose, did
not differ statistically between the 2 groups (Table 1).
In the study group, the baseline mean TSH value was
2.34 ± 1.3 µIU/mL, and the mean TSH value after taking
lansoprazole for 2 to 6 months and the same dose of LT4
was 3.0 ± 2.2 µIU/mL. This represented a mean change of
0.69 ± 1.9 µIU/mL (P = 0.035) (Fig. 1), with the post-PPI
TSH levels ranging from 0.09 to 8.0 µIU/mL (Table 2). In
comparison, the mean change in TSH value in the control
group, assessed 2 to 6 months after baseline measurement
during treatment with the same dose of LT4, was not sta-
tistically significant (P = 0.45) (Table 2). To avoid skew-
ing of data, we subsequently excluded from analysis one
patient who had a TSH level exceeding 20 µIU/mL after
starting lansoprazole treatment.
Seven of 37 patients (19%) in the study group had
TSH levels higher than 5 µIU/mL after initiation of PPI
therapy, which necessitated LT4 dosage adjustment. The
mean adjustment in dose for these patients was 20 µg/day,
which represented a 35% mean increase in LT4 dose.
DISCUSSION
To our best knowledge, this is the first study in
humans with primary hypothyroidism receiving LT4
replacement therapy to demonstrate that a PPI (in this
study, lansoprazole) affects the TSH level in such patients.
Our patients were euthyroid, clinically and biochemically,
during treatment with a stable dose of LT4 for at least 6
months preceding the initiation of lansoprazole therapy.
This study showed that the mean TSH level increased after
 Protein Pump Inhibitors and TSH, Endocr Pract. 2007;13(No. 4) 347

Table 1
Baseline Demographic Characteristics of
the Study and Control Groups*

Control Study P
Factor group group value†

Number of patients 55 37 …
Sex (M:F) 10:45 7:30 …
TSH (µIU/mL)‡ 1.95 ± 1.2 2.34 ± 1.3 0.16
Age (yr)‡ 58.4 ± 12.1 59.4 ± 7.5 0.61
Body mass index (kg/m2)‡ 30.9 ± 5.2 30.6 ± 8.2 0.8
Levothyroxine dose (µg)‡ 96.6 ± 24.3 82.8 ± 40.3 0.068

*TSH = thyroid-stimulating hormone (thyrotropin).

†Group differences of variables were compared with use of the Student t test.
‡Data are shown as mean values ± SD.

2 to 6 months of lansoprazole therapy. In the study group,
19% of the patients required an adjustment of LT4 dose
after lansoprazole therapy was initiated, and the mean
increase in LT4 dose was 35%.
A review of the medications used by the patients in
the study and control groups showed that no other agents
known to affect thyroid hormone metabolism or absorp-
tion were introduced during the study period.
The proposed mechanism for increased TSH levels
during PPI therapy is either an increase in metabolic clear-
ance of LT4 or a reduction in gastrointestinal absorption of
LT4 (or both factors).
According to a recent study, reduction of gastric acid
secretion with a PPI (omeprazole) modifies the absorption
of LT4 (5). Thus, reduction of gastric acidity could have
been a factor modifying absorption of LT4 in our study
patients. Studies in rats, however, have shown that the
clearance of thyroxine (T4) by the liver is enhanced by
many chemical agents, such as 3-methylcholanthrene,
polychlorinated biphenyls, and hexachlorobenzene,

Fig. 1. Mean thyroid-stimulating hormone (TSH) changes in the control and study groups. See
text for further details. CI = confidence interval; PPI = proton pump inhibitor.
348 Protein Pump Inhibitors and TSH, Endocr Pract. 2007;13(No. 4)

Table 2
Mean Thyroid-Stimulating Hormone Changes
in the Study and Control Groups*

µIU/mL)
TSH (µ P
Group Baseline Post-PPI† Change value‡

Study (N = 37)
Mean ± SD 2.34 ± 1.3 3.0 ± 2.2 0.69 ± 1.9 0.035
Range 0.6 to 4.5 0.09 to 8.0
Control (N = 55)
Mean ± SD 1.95 ± 1.2 1.84 ± 1.1 0.11 ± 1.06 0.45
Range 0.4 to 5.1 0.37 to 4.5

*PPI = proton pump inhibitor; TSH = thyroid-stimulating hormone (thyrotropin).

†Or equivalent study period for control group (not receiving a PPI).
‡Group differences were compared with use of the Student t test.

through inducement of hepatic uridine diphosphate-glu-
curonosyltransferase (UGT) enzymes (7-9). In addition,
animal studies involving rats have demonstrated that lan-
soprazole at high doses reduces plasma T4 levels signifi-
cantly by increasing the biliary clearance of LT4 through
induction of UGT enzymes (6). Phase I (functionalization)
and phase II (conjugation) drug metabolism lead to con-
version of drugs into highly water-soluble products that
can be excreted. Cytochrome P-450 enzymes introduce a
functional group, such as hydroxyl, into the substrate dur-
ing phase I drug metabolism (10). During phase II, many
enzymes, including UGT enzymes, use this functional
group as a “handle” for conjugation with such moieties as
glucuronic acid, sulfate, and glutathione. UGT enzymes
catalyze the conjugation of drugs during the phase II drug
metabolism (10,11). Deiodination is the major route of
thyroid hormone metabolism in humans, with up to 85%
of T4 and about 50% of triiodothyronine (T3) and reverse
T3 disposed by deiodination. Glucuronidation is a major
metabolic pathway for thyroid hormone in rats, but much
less is known about the importance of thyroid hormone
metabolism by glucuronidation in humans (12).
The suggested mechanism for alteration of thyroid
hormone homeostasis in humans by antiepileptic drugs is
probable induction of UGT enzymes or some other
unknown mechanism (12). The exact mechanism for alter-
ation of thyroid hormone homeostasis in humans by
antiepileptic drugs, however, needs to be investigated fur-
ther. Surks and DeFesi (13) demonstrated that therapeutic
levels of phenytoin and carbamazepine displace T4 and T3
from serum binding proteins with increased levels of free
T3 and free T4 hormone fractions, but unchanged free T4
and free T3 levels; the result is normal TSH levels on ultra-
filtration assay. That study also demonstrated that routine
commercial tests still show decreased free T4 concentra-
tions in patients taking phenytoin and carbamazepine (13).
Because PPIs have been shown to be inducers of UGT
enzymes in rats, any enhanced function of UGT enzymes
or other inducers in humans by PPIs may increase metab-
olism and clearance of LT4.
A direct effect of PPIs on pituitary function in humans
has not been demonstrated. In a study reported by
Dammann et al (14), 8 weeks of treatment with lansopra-
zole (30 mg daily) did not exert effects on serum TSH or
other pituitary hormones in healthy volunteers.
Centanni et al (5) recently suggested that PPIs could
interfere with the absorption of LT4 and result in increas-
es in TSH levels, necessitating the use of higher LT4 doses
to achieve TSH suppression in patients with multinodular
goiter. We agree that altering the acidic environment in the
gastrointestinal tract could reduce LT4 absorption. Our
study thus confirms the results reported by Centanni et al
(5) in a different group of patients—those with primary
hypothyroidism receiving LT4 replacement. We believe,
however, that another mechanism (or mechanisms), such
as induction of UGT or other enzymes, could have had a
role in accelerating the metabolism of LT4 that our study
group patients were receiving.
CONCLUSION
To the best of our knowledge, this is the first study in
humans with hypothyroidism receiving LT4 replacement
demonstrating the effect of a PPI on serum TSH levels. On
the basis of our findings, PPIs should be added to the list
of medications potentially affecting the level of thyroid
hormone in patients with hypothyroidism treated with LT4
replacement. Patients with hypothyroidism and normal
TSH values during LT4 replacement therapy may need to
undergo additional thyroid function testing after treatment
with PPIs and may need adjustment of their LT4 dose.
Moreover, because our study was a retrospective analysis,
 Protein Pump Inhibitors and TSH, Endocr Pract. 2007;13(No. 4) 349
further investigations—particularly randomized controlled
intervention trials, need to be conducted to verify our cur-
rent findings.
ACKNOWLEDGMENT
We acknowledge and thank Jihye Ann, PharmD, for
data collection and Ruth Hoffenberg, the director of
Queens Hospital Center Medical Library, for providing
reference resources.
DISCLOSURE
The authors have no conflicts of interest to disclose.
REFERENCES
1. Sherman SI, Tielens ET, Ladenson PW. Sucralfate caus-
es malabsorption of L-thyroxine. Am J Med. 1994;96:531-
535.
2. Singh N, Singh PN, Hershman JM. Effect of calcium car-
bonate on the absorption of levothyroxine. JAMA. 2000;
283:2822-2825.
3. Benvenga S, Bartolone L, Squadrito S, Lo Giudice F,
Trimarchi F. Delayed intestinal absorption of levothyrox-
ine. Thyroid. 1995;5:249-253.
4. Surks MI, Sievert R. Drugs and thyroid function. N Engl
J Med. 1995;333:1688-1694.
5. Centanni M, Gargano L, Canettieri G, et al. Thyroxine
in goiter, Helicobacter pylori infection, and chronic gastri-
tis. N Engl J Med. 2006;354:1787-1795.
6. Masubuchi N, Hakusui H, Okazaki O. Effects of proton
pump inhibitors on thyroid hormone metabolism in rats: a
comparison of UDP-glucuronyltransferase induction.
Biochem Pharmacol. 1997;54:1225-1231.
7. Barter RA, Klaassen CD. UDP-glucuronosyltransferase
inducers reduce thyroid hormone levels in rats by an
extrathyroidal mechanism. Toxicol Appl Pharmacol. 1992;
113:36-42.
8. van Raaij JA, Kaptein E, Visser TJ, van den Berg KJ.
Increased glucuronidation of thyroid hormone in hexa-
chlorobenzene-treated rats. Biochem Pharmacol. 1993;45:
627-631.
9. Saito K, Kaneko H, Sato K, Yoshitake A, Yamada H.
Hepatic UDP-glucuronyltransferase(s) activity toward thy-
roid hormones in rats: induction and effects on serum thy-
roid hormone levels following treatment with various
enzyme inducers. Toxicol Appl Pharmacol. 1991;111:99-
106.
10. Burchell B, Nebert DW, Nelson DR, et al. The UDP glu-
curonosyltransferase gene superfamily: suggested nomen-
clature based on evolutionary divergence. DNA Cell Biol.
1991;10:487-494.
11. Bosch TM, Meijerman I, Beijnen JH, Schellens JH.
Genetic polymorphisms of drug-metabolising enzymes and
drug transporters in the chemotherapeutic treatment of can-
cer. Clin Pharmacokinet. 2006;45:253-285.
12. Benedetti MS, Whomsley R, Baltes E, Tonner F.
Alteration of thyroid hormone homeostasis by antiepileptic
drugs in humans: involvement of glucuronosyltransferase
induction. Eur J Clin Pharmacol. 2005;61:863-872.
13. Surks MI, DeFesi CR. Normal serum free thyroid hor-
mone concentrations in patients treated with phenytoin or
carbamazepine: a paradox resolved. JAMA. 1996;275:
1495-1498.
14. Dammann HG, von zur Muhlen A, Balks HJ, et al. The
effects of lansoprazole, 30 or 60 mg daily, on intragastric
pH and on endocrine function in healthy volunteers.
Aliment Pharmacol Ther. 1993;7:191-196.