Endocrine (2018) 62:432–439

https://doi.org/10.1007/s12020-018-1686-1

ORIGINAL ARTICLE

Effect of timing of levothyroxine administration on the treatment of

hypothyroidism: a three-period crossover randomized study
Marko Skelin 1 Tomo Lucijanić2 Ana-Marija Liberati-Čizmek3 Sanja Majanović Klobučar4 Marko Lucijanić5
● ● ● ● ●

Lejsa Jakupović6 Miro Bakula3 Jelena Vučak Lončar7 Srečko Marušić2 Tomas Matić2 Željko Romić8
● ● ● ● ● ●

Jerka Dumić9 Dario Rahelić2

●

Received: 11 April 2018 / Accepted: 12 July 2018 / Published online: 24 July 2018
© Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract
Aim Hypothyroidism is a common clinical problem that is successfully treated with hormone substitutes in the form of
levothyroxine (LT4). LT4 is a drug with a narrow therapeutic index and is usually administered by strict rules, standardly at
least half an hour before breakfast. The aim of this study was to investigate a possible effect of different timings of
administration on thyroid function status and lipid profile.
1234567890();,:
1234567890();,:

Methods The study included patients with the diagnosis of primary hypothyroidism, which were using a stable dose of
levothyroxine. They were randomized into three different groups regarding the timing of LT4 administration in a crossover
fashion. Each timing regimen lasted for at least 8 weeks; timing regimen A—half an hour before breakfast; timing regimen B
—an hour before the main meal of the day; timing regimen C—at bedtime (minimally 2 h after dinner). The hormones (TSH,
fT3, fT4) and lipid profile (triglycerides, HDL-, LDL-, and total cholesterol) were measured before the study, at the
beginning of every timing regimen and at the end of the study.
Results Altogether, 84 patients finished the study. Different timings of LT4 administration were non-inferior in comparison
to the standard one and between each other. Median differences in TSH level between baseline and timing regimens were:
baseline vs. A = −0.017 95% C.I. (−0.400–0.192); baseline vs. B = −0.325 95% C.I. (−0.562–0.023); baseline vs. C =
−0.260 95% C.I. (−0.475–0.000). There were no statistically significant differences in either TSH, fT4, or fT3 when
compared between all three timing regimens of LT4 administration and the baseline. There were no statistically significant
differences in any of the lipid profile parameters (triglycerides, HDL-, LDL-, and total cholesterol) when compared between
all three timing regimens of LT4 administration and the baseline.
Conclusion The three investigated timing regimens of LT4 administration were equally efficient and offer additional options
regarding the treatment individualization.
Keywords Levothyroxine Timing of administration Individualization of pharmacotherapy Adherence Hypothyroidism.
● ● ● ●

* Marko Skelin 5
Hematology Department, Dubrava University Hospital,
marko.skelin@bolnica-sibenik.hr Zagreb, Croatia
6
Healthcare institution of community pharmacies, Slavonski Brod,
1
Pharmacy Department, General Hospital Šibenik, Šibenik, Croatia
Croatia
2
Department of Endocrinology, Diabetes and Metabolic Disorders, 7
Internal Medicine Department, General Hospital Zadar,
Dubrava University Hospital, Zagreb, Croatia
Zadar, Croatia
3
Department of Endocrinology, Diabetes and Metabolic Disorders, 8
Department of Laboratory Diagnostic, Dubrava University
General Hospital “Sv. Duh”, Zagreb, Croatia
Hospital, Zagreb, Croatia
4
Rijeka Clinical Hospital Centre, Faculty of Medicine, Department 9
Faculty of Pharmacy and Biochemistry, Department of
of Endocrinology, Diabetes and Metabolic Diseases, University of
Biochemistry and Molecular Biology, University of Zagreb,
Rijeka, Rijeka, Croatia
Zagreb, Croatia
Endocrine (2018) 62:432–439
Abbreviations
BMI Body mass index
fT3 Free triiodothyronin
fT4 Free thyroxine
HDL High-density lipoprotein
IQR Interquartile range
LDL Low-density lipoprotein
LT4 Levothyroxine
PPI Proton pump inhibitor
SmPC Summary of product characteristics
TSH Thyrotropin
Introduction
Hypothyroidism is a condition of reduced thyroid function
estimated to affect 3.7% of the population according to US
National Health and Nutrition Examination Survey, with
overt presentation of the disease in 0.3% of population [1].
It is nearly 10 times more frequent in women than in men
and incidence increases with age [2]. Clinical manifesta-
tions of hypothyroidism are nonspecific, variable, and
depend on duration and severity of thyroid hormone defi-
ciency. Fatigue, hoarseness, weight gain, and cold intoler-
ance are commonly present. Due to non-specificity of signs
and symptoms, diagnosis is based on laboratory testing,
with TSH values being the best diagnostic marker [2].
Hypothyroidism can have a negative effect on lipid profile
values resulting in increased cardiovascular risk [3]. How-
ever, an improvement of cardiovascular risk factors can be
achieved with the adequate treatment [4].
Since the 1960s, hypothyroidism has been treated with
thyroid hormone substitutes, most commonly levothyroxine
(LT4). LT4 is optimally absorbed under fasting conditions,
with 62–82% of the dose absorbed during the first 3 h after
the ingestion [5]. It is usually administered in low doses [6],
which makes it especially sensitive to the factors affecting
absorption. Consequently, due to its narrow therapeutic
index and chronic mode of use, interferences with absorp-
tion are mainly clinically relevant [7]. Rigid rules of
application could potentially interfere with factors that have
a negative effect on the absorption, and therefore a higher
flexibility in timing of administration could result in
improved individualization of pharmacotherapy, and better
treatment outcomes. Current summary of product char-
acteristics (SmPC) recommends taking the drug half an hour
before breakfast with a glass of water [8]. Since patient
lifestyle has been dramatically changed from the time when
LT4 was introduced into the market, many patients find
taking the drug strictly by the instructions of the SmPC
433
inconvenient. The increase of flexibility in timing of
administration for LT4 could result in a more individualized
treatment, with a possibility to reduce the frequency of
tablet administration during the day (referring to poly-
pharmacy due to comorbidities commonly observed in
patients with hypothyroidism), which has been described as
a negative factor considering adherence [9].
Increased absorption of LT4 during the switch from
morning to bedtime administration has been suggested [10].
On the other hand, opposite observation was reported as
well [11]. In addition, reduced absorption has been reported
with LT4 administration 1 h before dinner [12]. Due to
conflicting reports on this issue, it is not yet clear whether
different timing regimens have a clinically significant
impact on LT4 absorption estimated by the thyrotropin
(TSH) values.
Therefore, we aimed to investigate the impact of chan-
ging the timing of LT4 administration on the values of TSH,
free thyroxine (fT4), free triiodothyronin (fT3), and lipid
profile of the patients with primary hypothyroidism.
Patients and methods
Study population
Patients, 18 years or older, with a diagnosis of primary
hypothyroidism and stable dose of LT4 were eligible for the
study. After inclusion, patients continued to take the same
dose of LT4.
The exclusion criteria were pregnancy, celiac disease,
thyroid cancer, or Addison’s disease. Furthermore, the
examinees that were under increased risk of developing
symptoms of hypo or hyperthyroidism per physician’s
opinion were excluded from the study as well.
This study was conducted in the Republic of Croatia
from November 2014 to September 2017 in four hospitals.
The Ethical Committees of the participating hospitals
approved the study and all procedures were in accordance
with the Declaration of Helsinki. After confirmation of
eligibility, written informed consent was acquired from all
eligible patients.
Study design
This was a multicentre randomized open label three-period,
crossover study. The patients included in the study were
subjected to three different timing regimens of LT4
administration a half an hour before breakfast, b an hour
before the main meal of the day, and c at bedtime (mini-
mally 2 h after dinner). The randomization scheme was
generated by using the Web site Randomization.com
〈http://www.randomization.com〉. The patients were
434
Check 0 Check I.
R Regime I. R Regime II.
0 4 8 12
R - randomizaon
Check - Control medical exam and assessment of thyroid funcon status and lipid profile
Each regime consisted of one of the following randomly appointed ming of administraon:
(A) half an hour before breakfast,
(B) an hour before the main meal of the day
(C) at the bedme (minimally 2 hours aer dinner)
Fig. 1 The scheme of the study R — randomization, Check – Control
medical exam and assessment of thyroid function status and lipid
profile. Each regime consisted of one of the following randomly
randomized to one of the regimes at the beginning of the
study; afterwards they were switched from one regimen to
another according to the initial randomization, until they
passed all three of them, which can be seen on Fig. 1.
Duration of each timing regimen (a, b, and c) was 8 weeks
in order to achieve steady state between the possible
changes in LT4 absorption and TSH values. Control med-
ical exams were scheduled at the beginning of each studied
timing regimen and at the end of the study. At the control
visits, patients were instructed how to take LT4 according to
the next randomly assigned regimen. The patients who
needed to change the LT4 dose according to the physician’s
opinion were excluded from the study and switched back to
the regular LT4 timing of administration (half an hour
before breakfast).
Patients’ thyroid function status and lipid profile were
assessed according to standardized methods in each of local
hospital laboratories.
Tablets were not withheld on the day of the control visit.
Generic tablets of LT4 were used in this study, with
majority of patients (72%) using tablets containing lactose
as an excipient (Euthyrox®). Patients’ medical histories,
medication use, compliance, and physical examinations
were checked by a physician. Clinical and laboratory data
were used to evaluate patients’ safety and eligibility to
further participate in the study.
Endpoints of the study
The primary objective of this study was to determine
whether the change in timing of LT4 administration has an
impact on the serum parameters of therapeutic adequacy. In
addition, we investigated whether these changes in the
timing of administration, could have an impact on the
patients’ lipid profile. Clinical thyroid function status con-
sisted of the serum TSH, fT3, and fT4 levels, while the lipid
profile consisted of triglycerides, HDL-, LDL-, and total
cholesterol.
Endocrine (2018) 62:432–439
Check II. Check III.
R Regime III. End
16 20 24 weeks
appointed timing of administration: a half an hour before breakfast, b
an hour before the main meal of the day, c at the bedtime (minimally
2 h after dinner)
Statistical methods
Normality of distribution of measured numerical variables
was determined using the Shapiro–Wilk test. Since thyroid
function status and lipid profile parameters were non-
normally distributed, non-parametric tests were used for all
analyses. Non-normally distributed numerical variables
were presented as median and interquartile range (IQR),
normally distributed numerical variables were presented as
arithmetic mean ± standard deviation, while the categorical
variables were presented as a proportion and percentage.
For comparisons between the subgroups at baseline the
Mann–Whitney U test, the Student's T-test and the Χ2 test/
the Fisher test were used where appropriate. Correlation
between numerical variables was assessed using Spearman
rank correlation and described with the Spearman Rho
correlation coefficient.
Thyroid function status and lipid profile parameters were
compared between different time points (different treatment
regimens) using the Friedman ANOVA test for repeated
measures. P-values < 0.05 were considered to be statisti-
cally significant in this setting.
Variability of particular measurements at particular time
points was expressed as the coefficient of variation (ratio
between standard deviation and arithmetic mean) and were
compared between time points using the Brown–Forsythe
variant of the Levene test. P-values < 0.016666 according to
the Bonferroni correction for three simultaneous compar-
isons were considered to be statistically significant in this
setting. MedCalc statistical program, version 18 (MedCalc
Software bvba, Ostend, Belgium) was used for all analyses.
Power analyses were based on a comparison of TSH
between different treatment regimens using paired samples
T-test. Difference in TSH of 1.0 mIU/L was considered
clinically significant according to the prior studies [10]. Due
to three simultaneous comparisons between three treatment
regimens, type I error was set at 0.05/3 = 0.016666 in the
manner of the Bonferroni correction. Type II error was set
Endocrine (2018) 62:432–439 435

Table 1 Baseline characteristics

Total Male Female P-value

Number of participants 84 10.70% 89.30% —

Age 57 IQR (44.5–67.5) 59 IQR (24–63) 56.5 IQR (45–68) 0.514
<65 71.10% 77.80% 70.30% 1
≥65 28.90% 22.20% 29.70%
BMI (kg/m2) 28.1 ± 4.9 26.7 ± 4.7 28.3 ± 4.9 0.401
BMI (kg/m2)
<18.5 1.30% 0% 1.40%
18.5–24.9 25% 50% 22.20% 0.383
25–29.9 41.30% 25% 43.10%
≥30 32.50% 25% 33.30%
Statin in therapy 20.20% 33.30% 18.70% 0.378
Other lipid lowering agent in 8.30% 22.20% 6.70% 0.162
therapy
PPI or H2 receptor antagonist in 9.50% 11.10% 9.30% 1
therapy
Antipsychotic in therapy 13.10% 11.10% 13.30% 1
Average dose of LT4 mcg /BMI 2.6 IQR (1.7–3.3) 2.8 IQR (1.8–3.5) 2.6 IQR (1.7–3.3) 0.887
m2/kg
Length of LT4 intake (years) 6 IQR (3–9) 4 IQR (2–12.5) 6 IQR (3–9) 0.856
Length of LT4 intake (years)
<1 yrs 5.10% 0% 5.70%
1–4 yrs 8.20% 50% 25.70% 0.192
4–9 yrs 43.60% 12.50% 47.10%
>10 yrs 23.10% 37.50% 21.40%
TSH (mIU/L) 2 IQR (1.2–2.8) 2.3 IQR (1.6–4) 2 IQR (1.2–2.8) 0.269
fT4 (pmol/L) 15.6 IQR 15.2 IQR 15.6 IQR 0.718
(13.6–17.3) (14–16.3) (13.5–17.4)
fT3 (pmol/L) 4.5 IQR (4.1–5) 5.6 IQR (4.8–5.8) 4.4 IQR (4.1–4.9) 0.002*
Triglycerides (mmol/L) 1.4 IQR (1–2.1) 1 IQR (0.9–1.5) 1.4 IQR (1–2.3) 0.055
Total cholesterol (mmol/L) 5.1 IQR (4.6–6.3) 4.6 IQR (4–6.1) 5.2 IQR (4.7–6.3) 0.194
HDL-cholesterol (mmol/L) 1.4 IQR (1.2–1.7) 1.3 IQR (1.1–1.4) 1.4 IQR (1.2–1.7) 0.121
LDL-cholesterol (mmol/L) 3.2 IQR (2.7–4.3) 3.5 IQR (2.7–3.8) 3.2 IQR (2.8–4.3) 0.833
*Statistically significant at P < 0.05. The Mann–Whitney U test, the Student T-test and the Χ2 test were used

at 0.1 level (90% power). Standard deviation of differences
was 2.28 according to our pilot data. Based on aforemen-
tioned input parameters, minimum required number of
patients was 74.
(44.5–67.5), while the majority of patients have been
females (75/84; 89.3%). Patients’ characteristics stratified
by gender are shown in Table 1.
Primary endpoints

Results There were no statistically significant differences in either

Patients’ characteristics
A total number of 84 patients were included in the final
analysis (per protocol), while 13 patients were excluded
because they did not finish all three timing regimes of the
LT4 administration, 4 patients were excluded due to chan-
ging of LT4 dose and 3 patients were excluded because of
protocol violation. Median age was 57 years, IQR
TSH, fT3, or fT4 when compared between all three timing
regimens of LT4 administration and baseline (P > 0.05 for
all comparisons) as shown in Table 2 and Fig. 2. Median
differences in TSH between baseline and timing regimens
were: baseline vs. A = −0.017 95% C.I. (−0.400–0.192);
baseline vs. B = −0.325 95% C.I. (−0.562–0.023); base-
line vs. C = −0.260 95% C.I. (−0.475–0.000). Since 95%
C.I.s of median differences did not reach 1 mIU/L, all
timing regimens were non-inferior to the baseline in regard
436
Table 2 Thyroid hormones Baseline
during different modalities of
LT4 administration TSH (mIU/L) 2 IQR (1.2–2.8)
fT4 (pmol/L) 15.6 IQR (13.6–17.3)
fT3 (pmol/L) 4.5 IQR (4.1–5.0)
The Friedman ANOVA test for paired samples was used.
Fig. 2 Thyroid hormones during different modalities of LT4 admin-
istration. Special symbols (dot, triangle, square) represent outside/far
out measurements and are automatically plotted by statistical program
to the TSH values. However, serum TSH levels were
insignificantly lower during regimen A compared with the
other two regimens (B and C) with interquartile ranges
shifted upward in the regimens B (1.4–3.3) and C (1.3–3.4)
compared with regimen A (1.1–3.2).
There was no statistically significant difference in
variability of either of TSH, fT3, or fT4 when compared
between all timing regimens of LT4 administration (P >
0.05/ > 0.0166 for all comparisons), Table 3.
Secondary endpoints
There were no statistically significant differences in any of
the lipid profile parameters (triglycerides, HDL-, LDL-, and
total cholesterol) when compared between all timing regi-
mens of LT4 administration and to the baseline (P > 0.05
for all comparisons) as shown in Table 4 and Fig. 3.
There was no statistically significant difference in
variability of either triglycerides, HDL-, LDL-, or total
cholesterol when compared between all timing regimens of
LT4 administration (P > 0.05/ > 0.0166 for all compar-
isons), Table 3.
In addition, we detected no significant differences
between the timing regimens in neither thyroid function
status nor lipid profile parameters (P > 0.05 for all com-
parisons), concluding that sequence of study periods
through the study did not affect measured parameters.
Endocrine (2018) 62:432–439
A B C P-value
1.9 IQR (1.1–3.2) 2.3 IQR (1.4–3.3) 2.2 IQR (1.3–3.4) 0.193
15.1 IQR (12.9–16.6) 14.3 IQR (12.4–16.5) 14.9 IQR (13.4–16.9) 0.145
4.5 IQR (4.0–5.2) 4.6 IQR (4.1–5.0) 4.5 IQR (4.0–5.0) 0.793
Analysis of specific subgroups
To account for possible differences in LT4 absorption in
specific subgroups of patients, we performed additional
analyses of primary and secondary endpoints in the fol-
lowing subgroups: patients older than 65 years; patients
taking/not taking statins; patients taking/not taking other
lipid lowering agents (omega-3-acid, fenofibrate); patients
taking/not taking proton pump inhibitors (PPI) or H2
receptor antagonist; patients with BMI > 30/BMI ≤ 30.
In a subgroup of patients with BMI ≤ 30 we observed
statistically significant difference in HDL—cholesterol (P
= 0.027; lowest values observed during bedtime adminis-
tration of LT4) and triglycerides (P = 0.041; highest values
observed during bedtime administration of LT4), suggesting
that these patients might not benefit from bedtime LT4
administration. No other statistically significant differences
in primary or secondary endpoints were observed in this and
other analysed subgroups of patients (P > 0.05 for all ana-
lyses). It should be noted that these analyses were
exploratory, limited by small number of patients in specific
subgroups, and suffer from the problem of multiple com-
parisons (inflation of type I error/increased chance of false
positive findings), therefore should be interpreted with
caution.
Discussion
According to our knowledge and reviewed literature, this
study is the first randomized crossover study, which directly
compares the effects of all three previously investigated
times of LT4 administration (half an hour before breakfast,
an hour before the main meal of the day, and at bedtime) on
thyroid function status and lipid profile. The differences
between the values of TSH, fT3, and fT4, as well as tri-
glycerides, HDL-, LDL-, and total cholesterol were mea-
sured after each of the three different timing of
administration. We chose crossover design because each
patient serves as his/her own control, which reduces bias
and consequently abolishes the need for control group.
A previous study [10] has shown that a 6-weeks period is
a minimally sufficient time frame for measuring the effect of
the changes in LT4 absorption on TSH values. We decided
to extend the necessary 6-week period with an additional
2 weeks of randomly assigned timing of administration,
Endocrine (2018) 62:432–439 437

Table 3 Coefficient of variation A B C P-value A vs. B P-value A vs. C P-value B vs. C

of thyroid function status and
lipid profile parameters during TSH (mIU/L) 71.74% 69.35% 71.21% 0.652 0.463 0.771
different modalities of LT4 fT4 (pmol/L) 19.91% 22.19% 19.68% 0.551 0.928 0.558
administration fT3 (pmol/L) 20.09% 15.72% 18.81% 0.055 0.162 0.714
Triglycerides (mmol/L) 47.27% 48.75% 44.03% 0.474 0.476 0.386
Total cholesterol (mmol/L) 22.98% 23.53% 22.57% 0.612 0.688 0.723
HDL-cholesterol (mmol/L) 22.38% 22.07% 21.83% 0.510 0.497 0.627
LDL-cholesterol (mmol/L) 31.98% 32.00% 30.70% 0.628 0.660 0.606

The Brolwn–Forsythe variant of the Levene test was used.

Table 4 Parameters of lipid metabolism during different modalities of LT4 administration

Baseline A B C P-value

Triglycerides (mmol/L) 1.4 IQR (1–2.1) 1.6 IQR (1.1–2.2) 1.5 IQR (1–2.1) 1.5 IQR (1.1–2.1) 0.213
Total cholesterol (mmol/L) 5.1 IQR (4.6–6.3) 5.4 IQR (4.5–6.3) 5.2 IQR (4.5–6.2) 5.2 IQR (4.6–6.2) 0.506
HDL-cholesterol (mmol/L) 1.4 IQR (1.2–1.7) 1.4 IQR (1.2–1.7) 1.4 IQR (1.2–1.6) 1.4 IQR (1.2–1.7) 0.177
LDL-cholesterol (mmol/L) 3.2 IQR (2.7–4.3) 3.2 IQR (2.8–4.1) 3.4 IQR (2.8–4.1) 3.3 IQR (2.7–4.1) 0.696
The Friedman ANOVA test was used for paired samples.

night, which can have an influence on the greater exposure

Fig. 3 Parameters of lipid metabolism during different modalities of
LT4 administration. Special symbols (dot, triangle, square) represent
outside/far out measurements and are automatically plotted by statis-
tical program
thus the time frame for each regimen (A, B, and C) was
8 weeks. The same period has been used before in a similar
study, which investigated influence of timing of adminis-
tration on LT4 absorption [11]. Based on the obtained
results, non-inferiority of all three regimens to the current
standard has been shown.
Several studies have suggested that the timing of LT4
administration could be important for its absorption. A
randomized double-blinded placebo-controlled crossover
study [10] suggested that specific circadian properties of
gastrointestinal tract resulted in increased LT4 absorption
when administered at the bedtime. This conclusion was
explained by the reduced motility of the gut during the
of the LT4 on the intestinal wall. Since there is the evidence
of beneficial effect of gastric acid on LT4 absorption [13]
and basal acid secretion is the greatest during the night [14]
this could also contribute to the greater exposure of the LT4
on the intestinal wall. These results also implied that the
interval of 30 min between administration of LT4 and
having breakfast in not long enough to prevent the inter-
ference of food with absorption.
Another randomized three-period crossover study that
included patients with thyroid cancer, documented the
better LT4 absorption when it was taken an hour before
breakfast comparing to the bedtime administration, which
seems to be a challenging time of administration consider-
ing the modern way of life [11]. Although breakfast seems
to be an important interfering factor regarding LT4
absorption [15], recent evidence suggests that this problem
could be solved with newer formulations of LT4 [16–18].
Interestingly, a crossover placebo-controlled study, docu-
mented a slightly increased TSH values when LT4 was
administered before dinner [12]. In addition, there was no
difference observed between the evening LT4 administra-
tion as compared to the morning administration in the drug
naive patients [19]. While these studies have shown that the
timing of LT4 administration may affect its absorption, they
did not provide clear answers which timing of administra-
tion is the most appropriate, and whether different timing
regimens are non-inferior to each other considering clini-
cally relevant change in TSH.
In the current study, we showed non-inferiority of dif-
ferent timings of LT4 administration in regard to the TSH
regulation. Furthermore, there were no statistically sig-
nificant differences in neither thyroid function status nor
438
lipid profile between three different timings of LT4
administration. However, we note that the small non-
significant tendency towards better absorption was docu-
mented in the timing regimen of the administration
according to the SmPC (half an hour before breakfast).
Also, subgroup analyses suggest that patients with BMI ≤
30 might not benefit from bedtime LT4 administration
regarding HDL-cholesterol and triglycerides regulation.
These findings are interesting but cannot be interpreted as
evidence due to limitations of subgroup analyses as stated
previously in the results section.
Considering the current results, our data also suggests
that administration of LT4 an hour before the main meal
was no different than administration half an hour before
breakfast. Although our results are not directly comparable,
they seem to be contrary to the results observed in the
previous study that found a difference comparing LT4
administration half an hour and 1 h before breakfast [15].
The difference in results can be attributed to the different
doses of LT4 (regular vs 1000 mcg), duration of the studies
(8 weeks vs single ingestion) and administration of LT4 at
different time points during the day.
One of the main aims of LT4 therapy is to maintain the
TSH values in the relatively narrow range and to avoid
significant oscillations. Results of our trial (adequately
powered to investigate effect of different timings of LT4
administration on TSH regulation) question strict recom-
mendations for administration of LT4 [8]. Since we did not
notice any differences in thyroid function status, we believe
that LT4 can be administered in accordance with patient’s
preference with avoidance of concomitant use with food,
leaving the space for more individualized treatment options.
The need for more individualized treatment with LT4 was
seen in the randomized double-blinded placebo-controlled
crossover trial [10], where half of the patients after com-
pletion of the study continued to take LT4 at bedtime.
For the patients whose habits interfere with morning LT4
administration or for those who have difficulties with
attaining normal TSH values with morning administration,
an alternative timing of administration may be more con-
venient. It is important to highlight that 20.6% of patients
receiving thyroid hormone replacement therapy do not
achieve treatment goals [1]. The more individualized
approach can result in a better adherence and consequently
might have a positive impact on the treatment outcomes. It
is also important to note that these issues might be resolved
with newer formulations of LT4. New formulations of LT4
in form of liquid and soft gel capsules seem to have a better
pharmacokinetic profile, due to the fact that they are less
prone to the factors that affect LT4 absorption in tablet
formulation [16–18]. In addition, soft gel LT4 capsule
absorption seems to be improved even in patients without
proven malabsorption [20]. However, the results of our
Endocrine (2018) 62:432–439
study cannot be extrapolated to the other formulations of
LT4 that are not in tablet form.
There are several limitations to our study. We did not
apply placebo-controlled blinded design as we believe this
would filter the patients with extremely good adherence due
to the more demanding protocol (taking tablets three times a
day for a 24 weeks), and consequently decrease the gen-
eralizability of our findings. Yet, our study still may suffer
from the adherence problem, because participants recruited
into the study might be selected as patients with good
adherence due to demanding nature of our study. Bedtime
regimen of LT4 administration with minimally 2 h after the
dinner could be subject of criticism, as this might not be
sufficient time frame for food to leave stomach, which could
consequently affect LT4 absorption. However, our results
are easier to compare with previous studies with this timing
of administration, and as we show, it seems to be non-
inferior to other regimens of administration. In addition, we
did not exclude patients with conditions or drugs that might
impair gastric acid secretion, which can affect LT4
absorption [21–23]. However, our subgroup analyses of
patients with/without PPI or H2 receptor antagonist in
therapy did not identify differences in study endpoints.
Patients with lactose intolerance were not excluded as well,
although a subset of patients received LT4 tablets contain-
ing lactose as an excipient. Nevertheless, some benefits of
our study are that it has an adequate number of included
patients and a multicentre randomized crossover design.
In conclusion, timing of LT4 administration an hour
before the main meal of the day and at bedtime (2 h after
dinner) were shown in our study to be non-inferior to the
standard administration of LT4 by SmPC (half an hour
before breakfast). Therefore, multiple timing options with
similar treatment efficacy are possible with the aim of an
individualized treatment approach and achievement of
higher compliance.
Funding This study did not receive any external funding.
Authors’ contribution All authors contributed to preparation of this
manuscript.
Compliance with ethical standards
Conflict of interest DR has served as principal investigator or co-
investigator in clinical trials of AstraZeneca, Eli Lilly, MSD, Novo
Nordisk, Sanofi Aventis, Solvay, and Trophos. He has received hon-
oraria for speaking or advisory board engagements and consulting fees
from Abbott, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Eli
Lilly, Lifescan–Johnson & Johnson, Novartis, Novo Nordisk, MSD,
Merck Sharp & Dohme, Pfizer, Pliva, Roche, Salvus, Sanofi Aventis,
and Takeda. MS has recevied honoraria for lectures from Roche. Other
co-authors have no conflict of interest. TL has served as principal
investigator or co-investigator in clinical trials of Bayer HealthCare
AG, Sanofi Aventis, Institut de Recherches Internationales Servier,
Zavante Therapeutics, and Novo Nordisk. He has received honoraria
Endocrine (2018) 62:432–439
for speaking engagements from Boehringer Ingelheim, Eli Lilly, Novo
Nordisk, MSD.
Ethical approval All procedures performed in studies involving human
participants were in accordance with the ethical standards of the
institutional and/or national research committee and with the 1964
Helsinki declaration and its later amendments or comparable ethical
standards.
Informed consent Informed consent was obtained from all individual
participants included in the study.
References
1. Y. Aoki, R.M. Belin, R. Clickner, R. Jeffries, L. Phillips, K.R.
Mahaffey, Serum, TSH and total T4 in the United States popu-
lation and their association with participant characteristics:
National Health and Nutrition Examination Survey (NHANES
1999-2002). Thyroid 17, 1211–1223 (2007)
2. M. Devdhar, Y. Ousman, K. Burman, Hypothyroidism. Endocri-
nol. Metab. Clin. North. Am. 36, 595–615 (2007)
3. C.V. Rizos, M.S. Elisaf, E.N. Liberopoulos, Effects of thyroid
dysfunction on lipid profile. Open Cardiovasc. Med. J. 5, 76–84
(2011)
4. B. Biondi, I. Klein, Hypothyroidism as a risk factor for cardio-
vascular disease. Endocrine 24, 1–13 (2004)
5. L. Liwanpo, J.M. Hershman, Conditions and drugs interfering
with thyroxine absorption. Best. Pract. Res. Clin. Endocrinol.
Metab. 23, 781–792 (2009)
6. S.J. Mandel, G.A. Brent, P.R. Larsen, Levothyroxine therapy in
patients with thyroid disease. Ann. Intern. Med. 119, 492–502
(1993)
7. M. Skelin, T. Lucijanić, D. Amidžić Klarić, A. Rešić, M. Bakula,
A. Liberati-Čizmek, H. Gharib, D. Rahelić, Factors affecting
gastrointestinal absorption of levothyroxine: a review. Clin. Ther.
39, 378–403 (2017)
8. Agency for Medicinal Products and Medical Devices. (2015)
Euthyrox 100 mg tablets http://www.halmed.hr/upl/lijekovi/SPC/
UP-I-530-09-14-01-218.pdf. Accessed 10 January 2017.
9. A.J. Claxton, J. Cramer, C.A. Pierce, A systematic review of the
associations between dose regimens and medication compliance.
Clin. Ther. 23, 1296–1310 (2001)
10. N. Bolk, T.J. Visser, A. Kalsbeek, R.T. Van Domburg, A.
Berghout, Effects of evening vs morning thyroxine ingestion on
serum thyroid hormone profiles in hypothyroid patients. Clin.
Endocrinol. (Oxf.). 66, 43–48 (2007)
439
11. T.G. Bach-Huynh, B. Nayak, J. Loh, S. Soldin, J. Jonklaas,
Timing of levothyroxine administration affects serum thyrotropin
concentration. J. Clin. Endocrinol. Metab. 94, 3905–3912 (2009)
12. S. Ala, O. Akha, Z. Kashi, H. Asgari, A. Bahar, N. Sasanpour,
Dose administration time from before breakfast to before dinner
affect thyroid hormone levels? Casp. J. Intern. Med. 6, 134–140
(2015)
13. M. Centanni, L. Gargano, G. Canettieri, N. Viceconti, A. Franchi,
G. Delle Fave, B. Annibale, Thyroxine in goiter, Helicobacter
pylori infection, and chronic gastritis. N. Engl. J. Med. 17,
1787–1795 (2006)
14. J.G. Moore, E.Jr Englert, Circadian rhythm of gastric acid
secretion in man. Nature 226, 1261–1262 (1970)
15. S. Benvenga, L. Bartolone, S. Squadrito, F. Lo Giudice, F. Tri-
marchi, Delayed intestinal absorption of levothyroxine. Thyroid 5,
249–253 (1995)
16. R. Vita, P. Fallahi, A. Antonelli, S. Benvenga, The administration
of L-thyroxine as soft gel capsule or liquid solution. Expert. Opin.
Drug. Deliv. 11, 1103–1111 (2006)
17. G. Ianiro, F. Mangiola, T.A. Di Rienzo, S. Bibbò, F. Franceschi,
A.V. Greco, A. Gasbarrini, Levothyroxine absorption in health
and disease, and new therapeutic perspectives. Eur. Rev. Med.
Pharmacol. Sci. 18, 451–456 (2014)
18. A.M. Formenti, G. Mazziotti, R. Giubbini, A. Giustina, Treatment
of hypothyroidism: all that glitters is gold? Endocrine 52, 411–413
(2016)
19. R. Rajput, S. Chatterjee, M. Rajput, Can levothyroxine be taken as
evening dose? Comparative evaluation of morning versus evening
dose of levothyroxine in treatment of hypothyroidism. J. Thyroid
Res. 2011, 505239 (2011). https://doi.org/10.4061/2011/505239
20. P. Trimboli, C. Virili, M. Centanni, L. Giovanella, Thyroxine
treatment with softgel capsule formulation: usefulness in hypo-
thyroid patients without malabsorption. Front. Endocrinol. (Lau-
sanne). 9, 118 (2018)
21. E. Lahner, C. Virili, M.G. Santaguida, B. Annibale, M. Centanni,
Helicobacter pylori infection and drugs malabsorption. World J.
Gastroenterol. 20, 10331–10337 (2014)
22. M. Cellini, M.G. Santaguida, C. Virili, S. Capriello, N. Brusca, L.
Gargano, M. Centanni, Hashimoto’s thyroiditis and autoimmune
gastritis. Front. Endocrinol. (Lausanne). 8, 92 (2017)
23. I. Sachmechi, D.M. Reich, M. Aninyei, F. Wibowo, G. Gupta, P.J.
Kim, Effect of proton pump inhibitors on serum thyroid-
stimulating hormone level in euthyroid patients treated with
levothyroxine for hypothyroidism. Endocr. Pract. 13, 345–349
(2007)