Absorption kinetics of levothyroxine is not altered by proton-pump inhibitor therapy.

Introduction
Iodothyronines are widely prescribed drugs in developed coun−
tries. Several studies have shown a high prevalence of hypothy−
roidism in unselected populations [1]. In view of the frequency
and comparatively small therapeutic range of thyroid hormones,
interactions with other drugs are of considerable importance.
The absorption of iodothyronines is known to be influenced by
numerous substances, including antacid drugs [2, 3]. Sufficient
gastric acidity has been assumed to be a prerequisite for effective
iodothyronine resorption as a result [4]. However, whether thy−
roid replacement therapy is disturbed in patients treated with
proton−pump inhibitors is still unknown, although this most ef−
fective form of acid inhibition is increasingly prescribed. To ad−
dress this question, we have conducted a study in order to an−
swer the question as to whether proton−pump inhibition affects
levothyroxine resorption.
Methods
Participants
Volunteers aged between 20 and 30 years who met certain inclu−
sion criteria (no history of thyroid disorders, normal thyroid
gland as ascertained by ultrasonography, BMI between 17 and
27 kg/m2, and euthyroid state as evidenced by normal levels of
thyrotropin (TSH) and peripheral thyroid hormones) were select−
ed for participation in a trial for resorption kinetics.
Study design
In the first screening visit history, thyroid ultrasonography
(Esaote Biomedica�, Genova, Italy, transducers at 5 and 10 MHz)
and initial laboratory investigations were performed in order to
screen for eligibility according to the inclusion criteria.
All subjects included were randomly assigned by picking straws
to one of two groups, with each group consisting of ten persons.
We chose an open crossover design with members of group A
participated in a resorption−kinetic trial (as described below) im−
mediately after inclusion and group B first receiving 40 mg of
pantoprazole (Pantozol�, ALT ANA Pharma AG, Konstanz, Germa−
ny) daily for one week before taking part in the kinetics trial. In
the second group, pantoprazole was then discontinued, and the
volunteers crossed over after a wash−out period of at least four
weeks with B subjected to the kinetics trial, while group A first
received pantoprazole for one week.
In the kinetics trial, levothyroxine (Euthyrox�, Merck KgaA,
Darmstadt, Germany) was administered in both groups orally in
the morning after overnight fasting. Following ingestion of 4 �g
L−thyroxin per kg body mass, we collected blood samples over
ten hours to obtain a time series for plasma TSH and levothyrox−
ine levels. Samples were taken every fifteen minutes during the
first three hours and every sixty minutes over the next seven
hours.
Absorption Kinetics of Levothyroxine Is Not
Altered by Proton−pump Inhibitor Therapy
J. W. Dietrich1, 2
K. Gieselbrecht2
R. W. Holl3
B. O. Boehm2
Affiliation
1 Medical Clinic I, Endocrinology and Diabetology, Bergmannsheil Hospitals, Ruhr University Clinics, Bochum,
Germany
2 Department of Internal Medicine I, Ulm University, Ulm, Germany
3 Central Institute for Biomedical Engineering, Ulm University, Germany
Correspondence
Prof. Bernhard O. Boehm, M.D. ´ Division of Endocrinology and Diabetes ´ Ulm University ´
Robert−Koch−Straße 8 ´ 89081 Ulm ´ Germany ´ Phone: +49 (731) 500−24 304 ´
Fax: +49 (731) 500−23 938 ´ E−Mail: bernhard.boehm@medizin.uni−ulm.de
Received 4 May 2005 ´ Accepted after revision 8 August 2005
Bibliography
Horm Metab Res 2006; 38: 57±59 � Georg Thieme Verlag KG Stuttgart ´ New York ´
DOI 10.1055/s−2006−924980 ´ ISSN 0018−5043
ShortCom
m
unication
57

Sample handling and laboratory investigation
All blood samples were frozen for simultaneous analysis. Speci−
mens were analyzed in a single batch with commercial immuno−
metric assays (Ortho−Clinical Diagnostics, Amersham, UK) for se−
rum TSH levels (analytical sensitivity 0.003 mU/l, intraassay CV
1.2± 4.5 %), thyroxine (T4, analytical sensitivity 4.0 nmol/l, in−
traassay precision 1.8 ±2.0 %), and free thyroxine equivalents
(FT4E, equivalent value calculated from T4 and T3 Uptake, with in−
traassay CV range from 2.4 ± 2.7 %) including fasting gastrin lev−
els (radioimmunoassay, MP Biomedicals, Inc., Orangeburg, NY,
USA, analytical sensitivity 3.3 pg/ml, intraassay precision 4.1 ±
5.9 %).
Mathematical and statistical analysis
Resorption kinetics values over ten hours were calculated as the
area under the concentration curve (AUC) with the trapezoid
method from baseline−corrected hormone levels. Peak levels for
thyroid hormones and minimum levels for TSH were also deter−
mined. Maximum and minimum values as well as AUCs were
compared using Wilcoxon’s signed−rank test, and gastrin levels
were compared using the t−test for paired samples. Statistical
analyses were performed using StatView 4.5 software for Macin−
tosh computers (Abacus Concepts, Berkeley, CA, USA). Signifi−
cance levels were corrected for the number of tests performed
according to Bonferroni and Sidak.
Ethical aspects
The study protocol was approved by the local Ethics Committee
of the University of Ulm; all participants gave their written in−
formed consent.
Results
Trial profile
The trial was conducted in the summer of 2003. One of the twen−
ty−two individuals participating in the study dropped out as a re−
sult of nodular goiter diagnosis, and one other was excluded due
to elevated basal TSH levels; the remaining twenty volunteers
(11 female, 9 male, all Caucasian) completed the study.
Baseline characteristics
There were no significant differences between the two groups in
baseline TSH (1.3 vs. 1.6 mU/l), T4 (99.8 vs. 96.5 nmol/l), FT4E (2.1
vs. 2.0), T3 (2.0 nmol/l), FT3E (35.3 vs. 33.1), thyroglobulin (6.0 vs.
4.9 �g/l), gastrin levels (63.2 vs. 91.8 pg/ml), thyroid volume (9.4
vs. 7.8 ml), age (24.6 vs. 25.5 years), body mass (71.7 vs. 71.1 kg),
body height (1.74 vs. 1.78 m) and BMI (23.6 vs. 23.3 kg/m2).
Resorption kinetics
After one week of treatment with pantoprazole, the participants
showed significantly higher fasting gastrin levels (70.8 vs.
50.1 pg/ml, p = 0.0011) as expected for an elevated pH in the
stomach. Contrary to expectations, AUCs for TSH, T4 and FT4E
did not differ significantly (Fig.1 and Table 1). Minimum levels
of TSH and peak levels of T4 and FT4E showed slightly greater dif−
ferences with and without pantoprazole treatment, but these re−
sults were also non−significant after correction for the number of
tests performed.
Discussion
Thyroid hormones are frequently prescribed drugs with a limited
therapeutic range. Therefore, interactions are an important issue
in replacement therapy with iodothyronines. Physicians used to
advice their patients to take their medication on an empty stom−
ach in order to avoid interference from food. Indeed, earlier stud−
ies revealed a strong influence of nutrition and drugs on the re−
Fig. 1 Time series for TSH, T4 and FT4E over ten hours.
Dietrich JW et al. Influence of PPIs on L−T4 Absorption ´ Horm Metab Res 2006; 38: 57 ± 59
ShortCom
m
unication
58

sorption of thyroid hormones [2, 3,5]. Singh et al. reported a
25.2 % decrease in the AUC for T4 in subjects treated with calcium
carbonate [6]. Furthermore, they saw peak T4 levels decreasing
by 33.8 % and a reduction of peak FT4 levels of as much as 44.3 %
by calcium carbonate administered at the same time. Even com−
paring different levothyroxine formulations showed significant
differences both of AUC and peak values. Berg and Mayor report−
ed a difference of 20.9% between subjects on Synthroid� and Le−
voxine�. In the same study, the peak concentration of T4 varied
by 20.5 % [7]. Krehan et al. compared L−Thyroxin Henning� and
Eferox�, reporting a difference of 26.3 % in the AUCs [8].
Conversely, several studies have found that proton−pump inhibi−
tion may impair the absorption of macronutrients [9] and vita−
mins such as cyanocobalamine [10].
Our study revealed that pantoprazole was not able to change the
resorption kinetics of levothyroxine over ten hours, although sig−
nificantly increased gastrin levels indicated a rise in gastric pH.
However, implications are limited by the fact that the relative ab−
sorption of levothyroxine is not independent of dose, mainly due
to saturation processes limiting resorption of thyroid hormones
in higher doses [11 ± 13]. Furthermore, the study cohort com−
prised healthy volunteers rather than hypothyroid patients re−
quiring substitution therapy. Both the normal state of the thyroid
glands and the relatively high thyroxin doses as specified by the
study protocol might have had an impact on their findings. Final−
ly, thyroid hormones require several days, if not weeks, of substi−
tution therapy to achieve steady−state levels. Within a number of
days, mechanisms such as levothyroxine degradation or reutili−
zation of thyroid hormones excreted through the bile might be
affected by proton−pump inhibitor therapy [14].
Nevertheless, compared to the earlier findings mentioned above,
which were partly based on trials sharing a similar design with
our study, the influence of pantoprazole on thyroid−hormone re−
sorption seems to be surprisingly limited. In conclusion, it is
highly unlikely that the acidity of gastric content plays a major
contributing role in the absorption of thyroid hormones; it
seems that factors other than gastric acidity, in particular ad−
sorption and competition processes for saturable transport
mechanisms, would be responsible for the known phenomenon
of nutrition influencing the resorption of thyroid hormones.
This observation is of considerable clinical importance since
both intermediate blockade of gastric acid output or concomi−
tant presence of achlorhydria and thyroid−hormone replacement
therapy are common.
Acknowledgments
The authors thank C. Rudtke, M. Fick, R. Sing, and S. Schilling for
performing the hormone assays. Drugs were a kind gift from AL−
TANA Pharma AG, Konstanz, Germany and Merck KgaA, Darm−
stadt, Germany. No sponsors were involved in study design,
data collection, data analysis, data interpretation, writing or sub−
mission of the report. There was no potential financial conflict of
interest. Supported by GRK 1041 “Molecular Diabetology and
Endocrinology in Medicine” to BoB.
References
1 Canaris GJ, Manowitz NR, Mayor G, Ridgway EC. Arch Intern Med
2000; 160: 526 ± 534
2 Surks MI, Sievert R. New Engl J Med 1995; 333 (25): 1688 ± 1694
3 Sperber AD, Liel Y. Arch Intern Med 1992; 152: 183± 184
4 Klein I, Danzi S. Thyroid 2003; 13: 1127 ± 1132
5 Campbell NR, Hasinoff BB, Stalts H, Rao B, Wong NC. Ann Intern Med
1992; 117: 1010 ± 1013
6 Singh N, Weisler SL, Hersman JM. Thyroid 2001; 11: 967± 971
7 Berg JA, Mayor GH. J Clin Pharmacol 1993; 33: 1135± 1140
8 Krehan A, Dittmar M, Hoppen A, Lichtwald K, Kahaly GJ. Medizinische
Klinik 2002; 97: 522 ± 527
9 Evenepoel P, Claus D, Geypens B et al. Aliment Pharmacol Ther 1998;
12: 1011 ± 1019
10 Marcuard SP, Albernaz L, Khazanie PG. Ann Intern Med 1994; 120:
211 ± 215
11 Hays MT. Endocrine Research 1988; 14: 203 ± 224
12 Krenning EP, Docter R, Visser TJ, Hennemann G. Acta medica Austriaca
1988; 15 (Suppl 1): 15± 17
13 Hennemann G, Docter R, Friesema EC, de Jong M, Krenning EP, Visser
TJ. Endocr Rev 2001; 22: 451 ± 476
14 Harmon SM, Seifert CF. Ann Intern Med 1993; 118: 112 ± 113
Table 1 Pharmacokinetic properties of levothyroxine without and after treatment with pantoprazole
Subjects in untreated state Subjects on Pantoprazole Rel. Difference p
Gastrin (pg/ml) 50.1  5.9b 70.8  10.9b + 41.3 % 0.001g
AUC TSH (mU ” sec/l)a,c ± 39 360.8  6811.0b ± 37 422.2  5440.0b + 4.9 % 0.65f
AUC T4 (mmol ” sec/L)a,c 1.09  0.06b 0.96  0.10b ± 12.4 % 0.60f
AUC FT4E (sec)a,c 28 829.9  1252.0b 25 944.1  1791.0b ± 10.0 % 0.09f
Minimum TSH (mU/l)d 1.599  0.228b 1.496  0.791b ± 6.4 % 0.47f
Maximum T4 (nmol/l)e 35.01  2.44b 27.80  2.70b ± 20.6 % 0.02f,h
Maximum FT4Ee 0.94  0.06b 0.78  0.06b ± 17.0 % 0.01f,h
aTSH: thyrotropin, T4: total thyroxine, FT4E: free thyroxine equivalent; bMean value  SEM; cBase−corrected area under the curve; dBase−corrected minimum activity
([minc] = [base] ± [min]); eBase−corrected peak concentration ([maxc] = [max] ± [base]); fWilcoxon signed rank test; gt−test for paired samples; hNot significant with
Bonferroni’s or Sidak’s adjustment for multiple testing.
Dietrich JW et al. Influence of PPIs on L−T4 Absorption ´ Horm Metab Res 2006; 38: 57 ± 59
ShortCom
m
unication
59