Efficacy of a levothyroxine dosage regimen based on serum thyrotropin level, for primary hypothyroidism. An open label dose finding study

ABSTRACT

Background:

There are no guidelines on individualized initial levothyroxine dosage in primary hypothyroidism. This prospective observational study was done to assess whether a predetermined dose of levothyroxine based on Thyroid Stimulating Hormone (TSH) levels would be able to make the patient euthyroid during a period of six weeks and to find other factors which influence the levothyroxine requirement.

Materials and Methods:

Newly diagnosed patients with primary hypothyroidism or those patients who were not on levothyroxine therapy were divided into TSH-based groups—Group 1, 5–9.99, Group 2, 10–29.99, Group 3, 30–99.99 and Group 4, >100 μIU/ml and treated with an initial levothyroxine dose of 25,50,75 and100 μg/day for next six weeks. Factors correlating with levothyroxine requirement were determined.

Results:

Of the 171 patients who were included 142 completed the study, 34,46,28 and 34 patients were included in groups 1 to 4, respectively. Normalization of TSH with the above criteria was achieved in 111 (78.7%) out of 141 patients, and 91%, 67%, 75%, and 82% respectively in the 4 groups. Among adequately replaced patients pre-treatment TSH level (r = 0.81), T4 level (r = 0.61), and body weight (r = 0.19) correlated with the levothyroxine requirement. Based on these factors predicted initial dose (μg/day) was found to be 0.54 (Body Weight [Kg]) +0.47 (TSH [μIU/m]) - 1.4 (Total T4 [μg/dl]) +17.79 or 0.27 (Body Weight) +0.553 (TSH) +21.

Conclusion:

Serum thyrotropin-based categorization for initial levothyroxine dose leads to euthyroidism in nearly four of five patients with primary hypothyroidism. The dose required for adequate replacement of levothyroxine has correlation with pre-treatment serum TSH levels serum thyroxine levels and body weight.

Introduction

Primary hypothyroidism is a common endocrine disorder affecting approximately 5% of the world population.[1] Hormone replacement with oral levothyroxine alone constitutes the treatment in the overwhelming majority of patients with primary hypothyroidism. If untreated, undertreated, or overtreated, primary hypothyroidism can lead to many complications. Whereas, undertreatment can lead to non-resolution of symptoms with treatment discontinuation, overtreatment can lead to palpitation and predispose the patient to cardiac tachyarrythmias including atrial fibrillation. Due to the lack of adequate studies, guidelines on the treatment of primary hypothyroidism recommend either an upfront full replacement dose (1.6 mcg/kg/day) of levothyroxine for treatment of primary hypothyroidism or a starting dose of 25 mcg/day up titrated every 4–6 weeks for patients with subclinical hypothyroidism who have indication for treatment, e.g., Thyroid peroxidase (TPO) antibody positivity, symptoms of hypothyroidism or periconceptional.[2,3,4] In practice, however, it is observed that many patients do not require a full replacement dose of levothyroxine for normalization of Thyroid Stimulating Hormone (TSH) levels.[5] Studies have shown that the levothyroxine requirement depends upon the lean body mass, pre-treatment TSH levels, age of the patient, and possibly the etiology of primary hypothyroidism.[6,7] Still the guidelines (e.g., American Thyroid Association guidelines for the management of hypothyroidism) are not clear on the approximate initial levothyroxine dose to be started in patients suffering from primary hypothyroidism based on various determinants of the same.[2] No previous studies have evaluated the efficacy of a pre-treatment TSH-based initial levothyroxine dose in hypothyroid patients. Moreover, there are very few prospective studies from South Asia regarding the initial dosage requirement of levothyroxine.[5] This prospective observational study was done to assess the proportion of patients who achieve euthyroidism if treated with a predetermined dose of levothyroxine based on TSH levels over a period of six weeks.

Materials and Methods

This prospective observational study was done in the Endocrinology outpatient clinic of a teaching hospital in South India. Institutional Human Ethics Committee approval was obtained (HEC 01/02/2020 dated 07.02.2020). The study was done over a period of two and half years starting from March 2020. Patients both male and female of 18 years or older were included in the study after taking informed consent. Newly diagnosed treatment naïve patients with primary hypothyroidism or those patients with primary hypothyroidism who were not on levothyroxine therapy for the last two months were included. Primary hypothyroidism was defined to be present in patients with TSH values more than normal with the serum thyroxine level being normal or low.[1] Patients with critical illness, and those unwilling for the study were excluded. Similarly, patients with chronic diarrhoea or any known chronic gastrointestinal disorder were excluded. Patients with central (secondary—pituitary/hypothalamic) hypothyroidism, those on any other medication known to alter levothyroxine absorption and those with ischemic, structural, or rhythm disorders of the heart were also excluded. Pregnant patients were excluded. Patients of differentiated thyroid carcinoma (requiring TSH suppression) were excluded from the study. Patients taking vitamin B complex supplements containing biotin were also excluded from the study.

The sample size calculation was done based on a pilot study on 10 patients, which revealed TSH normalization in 8 of them. The required sample size was estimated to be 130 patients based on an absolute precision of 7% at a confidence level of 95%. All patients attending the endocrine outpatient clinic fulfilling the inclusion criteria and exclusion criteria during the study period were included, sampling from within them was not done. Thyroid Stimulating Hormone was done for all patients, serum thyroxine level (T4), and anti-thyroid peroxidase antibody titer were advised for those without a prior report of the same to confirm primary hypothyroidism and autoimmunity. The patients were divided into four groups based on TSH levels. Those with TSH levels of 5–9.99 μIU/ml (subclinical hypothyroidism) were assigned to Group 1, 10–29.99 μIU/ml with mild primary hypothyroidism to Group 2, 30–99.99 μIU/ml with moderate primary hypothyroidism to Group 3 and those with >100 μIU/ml with severe primary hypothyroidism to Group 4 for the purpose of this study [see Figure 1]. These categories of patients were treated with 25 μg (Group 1), 50 μg (Group 2), 75 μg (Group 3), and 100 μg (Group 4) as the initial empirical levothyroxine dose. In Group 1, the treatment was initiated only if the patients had anti-TPO antibody positive or if they were women planning to conceive or if they had symptoms suggestive of hypothyroidism. They were advised to take a branded levothyroxine tablet of the same brand, an empty stomach in the morning after a minimum fasting period of four hours, and refrain from food or beverage intake for the next one hour after the consumption of the tablet as per international guidelines.[2] The patients were instructed to report back to the clinic after six weeks of levothyroxine intake with proper compliance to therapy. At the second visit to the clinic, repeat estimation of serum TSH was done. The patients were classified as adequately replaced if the follow-up TSH value fell between 0.1 and 5 mIU/ml and were clinically euthyroid, under-replaced if the value remained more than 5 mIU/ml, and over-replaced if the TSH value fell even below 0.1 mIU/ml or if they had any clinical features of over replacement with a TSH of 0.1–0.5 mIU/ml. All hormonal measurements were done between 8 am and 10 am irrespective of fasting state. Hormonal analysis for T4 and TSH and antibody titers were estimated on Roche Cobas e-411 using commercially available kits from Roche (Germany). The intra and interassay coefficient of variation for the estimated parameters were <7%. The obtained data was entered into Excel 2010 and analyzed with the help of Excel 2010 and IBM Statistical Package for the Social Sciences statistics version 28.0.1.0 (NY, USA). Categorical variables were expressed as frequency and percentage, normally distributed continuous variables as mean and standard deviation, and those with a skewed distribution as median and interquartile range. The strengths of association were assessed using Chi-square, Student t-test, and Mann–Whitney U test for the above categories of variables respectively. Pearson’s Correlation coefficient was used to show the strength of correlation between levothyroxine dosage and the analyzed parameters. Linear regression was used to derive a model for predicting appropriate initial levothyroxine dosage among those with normalization of TSH using the variables that showed correlation with the levothyroxine dose with which the TSH levels were controlled.

Figure 1.

Flow of patients and procedures in the study

Results

A total of 171 eligible patients attended the out patient department during the study period who were willing to the study [Table 1]. Pre-treatment and post-treatment TSH values were available for 142 patients and the rest did not follow-up. Out of these 142 patients, 108 (76%) were females. The mean (SD) age of the study subjects was 32.7 (14.1) years [Table 1]. Newly diagnosed cases of primary hypothyroidism constituted 126 (88.7%) and the rest were those with stoppage of levothyroxine replacement since at least 8 weeks. Nine patients had a history of thyroid surgery as a cause of primary hypothyroidism. Anti-TPO report was available for 89 participants of whom 72 (80.8%) were positive. Family history of primary hypothyroidism in first-degree relatives was found in 13 (0.09%) participants. Mean (SD) body mass index (BMI) of the participants was 23.3 (5.5) kg/m². Median (interquartile range) pre-treatment TSH was 21.6 μIU/ml (10.11–92.145). Median Total T4 values before treatment was 6.1 μg/dl (4–7.3). 32 out of 92 (34.8%) subjects had a T4 below the reference range. Baseline parameters are given in Table 1.

Table 1.

Baseline characteristics of studied subjects

Parameter	n (%)/Mean (SD)/Median (IQR)
Females	108 (76.0%)
Age (Years)	32.7 (14.1)
Height (cm)	157.5 (10.73)
Weight (Kg)	58.8 (15.6)
BMI (kg/m2)	23.(5.5)
Newly diagnosed	126 (88.73%)
TSH	21.6 (10.1-92.1)
Total T4	6.07 (3.75-7.3)

Of the 142 patients included 34 had a TSH of 5–9.99 (Group 1) μIU/ml, 46 had a TSH of 10–29.99 μIU/ml (Group 2), 28 had a TSH of 30–99.99 (Group 3) and 34 had TSH of ≥100 μIU/ml (Group 4). These patients were treated according to the TSH level with 25, 50, 75, and 100 mcg per day. The mean dose per log₁₀ TSH was 41.8 ± 11.46 (mcg/log₁₀ μIU/ml). The dose per log TSH was 33.48 ± 11.22 (mcg/log₁₀ μIU/ml) in group 1, 41.24 ± 5.76 (mcg/log₁₀ μIU/ml) in Group 2, 44.35 ± 16.21 (mcg/log₁₀ μIU/ml) in Group 3, and 49.41 ± 5.32 (mcg/log₁₀ μIU/ml) in Group 4.

Response to therapy was assessed as having achieved TSH <5 μIU/ml or not at six weeks after starting therapy. Safety was assessed based on the presence of prespecified adverse effects i.e., symptoms of thyrotoxicosis (palpitation or tremor) or the presence of tachycardia (pulse rate more than 100 bpm) or tremor on examination after starting of therapy. None of the patients had any symptoms of malabsorption like diarrhoea, weight loss, or oily stools while on treatment. All patients reported good compliance to therapy and the method of levothyroxine intake. Normalization of TSH with the above criteria was achieved in 111 (78.7%) out of 141 patients. In different groups, the normalization was achieved in 31/34 (91.2%), 31/46 (67.4%), 21/28 (75%), and 28/34 (82.4%) patients respectively in groups 1, 2, 3, and 4, respectively [Table 2]. The difference between the groups with respect to normalization of TSH was significant (P < 0.05). None of the patients reported any prespecified adverse effects. Non-normalization of TSH with the above doses was studied by comparing the baseline characteristics and dose relation to other parameters like age, sex, family history, weight, BMI, TPO positivity, and pre-treatment T4 Levels. None of the factors were found to be significantly associated with non-response to the TSH-based predetermined dose of levothyroxine [Table 3].

Table 2.

TSH normalization according to pre-treatment TSH level categories and dose used

TSH Category	5–9.99	10–29.99	30–99.99	>100	Total
Age	32.5±15.8	32.5±11.3	31.6±14.2	34.5±14.8	32.8±14.0 (P=0.90)
Dose of levothyroxine used for initial treatment (mg/day)	25	50	75	100	NA
Number of participants in each category	34 (23.9%)	46 (32.3%)	28 (19.7%)	34 (23.9%)	142
Euthyroidism Achieved	31 (91.1%)	31 (67.4%)	20 (71%)	29 (85.2%)	111 (78.2%)
Euthyroidism Not Achieved	3 (8.9%)	15 (32.6%)	8 (29%)	5 (14.8%)	31 (21.7%)

Table 3.

Comparison of basal characteristics among patients who achieved euthyroidism with the initial levothyroxine treatment over 6 weeks with those who could not

Parameter	TSH normalized	TSH not normalized	P
Age (years)	32.8±14.0	32.3±15.0	0.853
Weight (kg)	58.9±14.6	58.2±19.3	0.832
Dose/Weight (μg/kg/day)	1.05±0.59	1.29±0.59	0.09
BMI (kg/m2)	23.4±5.6	22.8±5.3	0.655
Log TSH (μg/day)	1.41±0.4	1.43±0.4	0.844
Dose per log TSH (μg//day/μIU/ml)	42.3±11.5	39.8±11.0	0.292
T4 (μg/dl)	5.4±3.6	6.3±2.3	0.316

When the data from 111 patients who achieved euthyroid state at six–eight weeks of therapy were analyzed, Log TSH, and TSH was found to have the best correlation with dose requirement (r = 0.81) [see Figure 2]. The other factors which had significant (P < 0.05) correlation were, T4 Levels (r = -0.61)) and body weight (r = 0.19) respectively. Age, lean body weight, height and BMI were found to have no significant correlation with the levothyroxine dose requirement [Table 4]. l Logistic regression confirmed the independent association of TSH, T4, and body weight with the initial dose requirement of the patient. Further based on these 3 factors a linear equation to predict the initial levothyroxine dose requirement was derived by linear regression [Table 5a-c]:

Figure 2.

Correlation between Log TSH (pretreatment) and dose requirement per kg body weight among participants with normalized TSH

Table 4.

Correlation between levothyroxine dose which controlled TSH levels with the various patient factors

	Pearson correlation coefficient (r)	P
Age	0.044	0.67
Body weight	0.193	0.04
Lean Body Mass	0.17	0.15
Body Mass Index	0.074	0.53
Height	0.153	0.19
Pre-treatment TSH	0.808	<0.001
Log Pre-treatment TSH	0.897	<0.001
T4	-0.599	<0.001

Table 5a.

Linear regression-based estimation of adequate replacement dose of levothyroxine using Log₁₀TSH and Body weight

	Model	Unstandardized coefficients		Standardized coefficients	t	Sig.
		B	Std. error	Beta
1	(Constant)	-31.969	6.444		-4.961	<.001
	Log TSH	58.532	2.791	0.886	20.970	<.001
	weight	0.201	0.091	0.093	2.204	0.030

Table 5c.

Linear regression-based estimation of adequate replacement dose of levothyroxine using TSH T4 and Body weight

	Model	Unstandardized coefficients		Standardized coefficients	t	Sig.
		B	Std. error	Beta
1	(Constant)	17.792	12.875		1.382	0.173
	weight	0.542	0.170	0.254	3.183	0.002
	TSH	0.469	0.076	0.646	6.159	<.001
	T4	-1.406	0.940	-0.157	-1.496	0.141

Table 5b.

Linear regression-based estimation of adequate replacement dose of levothyroxine using TSH and Body weight

	Model	Unstandardized coefficients		Standardized coefficients	t	Sig.
		B	Std. error	Beta
1	(Constant)	21.275	7.421		2.867	0.005
	weight	0.276	0.121	0.128	2.279	0.025
	TSH	0.553	0.039	0.798	14.244	<.001

Estimated adequate dose (μg/day) =0.54 (Body Weight) +0.47 (TSH) – 1.4 (Total T4) +17.79

If T4 values are not included the following linear and logarithmic equations are obtained:

1. Estimated adequate dose (μg/day) =0.27 (Body Weight) +0.553 (TSH) + 21

2. Estimated adequate dose (μg/day) =0.2 (Body Weight) +58.5 (Log10 TSH) – 32

Although lean body mass was not significantly associated with thyroxine requirement when an analysis of obese participants and non-obese with regard to the dose of thyroxine per kg per μIU/ml l of TSH was done the difference was not significant. But, when TSH was log-transformed, the difference was found to be significant with non-obese persons requiring a higher dose per kg per log TSH. This could indirectly mean that the lean body mass may have an influence on thyroxine dose requirements.

Discussion

Levothyroxine therapy for primary hypothyroidism usually proceeds from an empirical weight-based dose at initiation and measurement of TSH follow-up at six–eight weeks to assess the response to that initial dose. Further, based on TSH at the follow-up visit levothyroxine dosage is titrated to achieve euthyroid state confirmed at further follow-up visits. Once primary hypothyroidism is diagnosed and the decision to treat is taken, the next primary question in the mind of the treating physician is the dose of levothyroxine at which replacement should be initiated. The guidelines for primary hypothyroidism do not specifically mention the dose of levothyroxine, which needs to be initiated according to the level of TSH in spite of previous studies showing linear relation of dose of levothyroxine required for normalization of TSH to be directly proportional to the log of pre-treatment TSH value. The initial levothyroxine dose should be such that a majority of patients achieve normalization of TSH levels, thus minimizing the clinic visits that may be required to render the patient euthyroid while ensuring that the patient doesn’t end up in iatrogenic thyrotoxicosis. This prospective observational study was done to find whether a pre-treatment TSH-based levothyroxine dosage would normalize TSH levels in a majority of patients with newly diagnosed primary hypothyroidism or in those with treatment interrupted for more than two months. The study showed that such a TSH-based dosage regimen would normalize TSH levels within 6 weeks in nearly 80% of all the patients presenting with hypothyroidism especially in those with TSH levels less than 10 μIU/ml and those with TSH levels more than 100 μIU/ml where the normalization was seen in upto 86% patients at 6 weeks of therapy. In patients with TSH levels intermediate between 10 and 100, the normalization rate was seen in approximately 70% of the participants. None of the factors analyzed (age, sex, body weight, lean body mass, BMI, T4 levels, presence of T4 levels below normal, family history, or anti-TPO antibody positivity) could be associated with non-normalization of TSH levels.

Previous studies have described the determinants of replacement levothyroxine dosage to be age, TSH, and lean body weight.[8,9,10,11] Some studies have found a higher levothyroxine requirement in patients with post-surgical hypothyroidism compared to those with autoimmune thyroiditis.[12]

In the current study, the patients with TSH >100 achieved a better rate of TSH normalization than those with TSH 10–100 μIU/ml. This odd finding may be explained as follows. Patients with TSH > 100 μIU/ml were given the highest dose (100 mcg/day), which would be sufficient for a total replacement for the large majority of patients without any residual thyroid function whereas TSH values although representing lower degree of thyroid dysfunction may be a spectrum where the residual gland function cannot be predicted very accurately and doses which could have been inadequate for some patients lowest in the spectrum of residual thyroid function.

The current study couldn’t find a correlation between replacement levothyroxine dosage with age or lean body weight. The current study confirmed the strong correlation of dose requirement with TSH levels, found a moderately strong negative correlation with total T4 levels and also a significant correlation with body weight. Obese participants were found to have a lower dose per kg per log TSH requirement hinting an importance of lean body mass in the dose requirement.

Roos et al. conducted a randomized trial to compare a full replacement dose of levothyroxine (1.6 μg/kg/day) with a lower dose (25 μg/day) as the starting dose in patient with primary hypothyroidism irrespective of their TSH levels. Median TSH levels in that study were 61 and 48 μIU/ml in the full dose and low dose groups respectively and the study concluded that full replacement dose achieved biochemical euthyroidism faster than the low-dose therapy without the development of any adverse cardiac symptoms.[13] They also observed that finally when euthyroidism was achieved the dose requirement in the low-dose group was lower compared to the full replacement dose (1.5 μg/kg vs 1.7 μg/kg, P = 0.04). Also, there was no difference in the symptom score-based clinical improvement between the groups. This finding suggests that some patients with primary hypothyroidism may not require a full replacement dose to achieve clinical and biochemical euthyroidism. That study has not evaluated a TSH-based initial levothyroxine dosage nor reported the success or failure rate of either approaches (full replacement vs low dose) in normalizing TSH levels within six weeks of treatment onset.

Relationship between levothyroxine dosage and TSH was initially reported by Kabadi et al. when they studied the dosage requirements of 192 patients and found a close correlation between pre-treatment TSH levels and levothyroxine dosage required for TSH normalization.[14] The same team of investigators further studied reliability of the predicted optimal daily levothyroxine (LT4) dose based on pre-treatment serum thyrotropin (thyroid-stimulating hormone or TSH) levels in primary hypothyroidism and found a close match between both TSH based or log TSH-based calculated levothyroxine dosage and the actual maintenance levothyroxine dosage required by the patients when the dose adjustments were done in the conventional manner.[15] This study too did not report on the success rate with the starting levothyroxine dose. A study from India showed that sex and pre-treatment TSH were the best predictors of thyroxine dosage and that a prediction model based on these could achieve 68% success (n = 50) in achieving euthyroidism at 6 weeks.[16]

In summary, the previous studies have hinted a role for TSH in the determination of the maintenance doses required at the euthyroid state.[17] But none of the previous studies have reported on the success rate of TSH-based initial levothyroxine dosage nor have they included factors other than pre-treatment TSH together (like body weight or T4 levels) for the determination of initial levothyroxine dosage.[18] The current study adds to literature by confirming the role of pre-treatment TSH levels, assessing the capacity of TSH based dosage regimen to achieve euthyroidism, finding a role for the pre-treatment T4 levels and body weight in determining the optimal initial dose of levothyroxine in untreated hypothyroid patients, and devising a linear regression based formula to determine initial levothyroxine dosage for achieving euthyroidism in nearly 80% patients or possibly more with untreated primary hypothyroidism within a span of 6 weeks. The simplest formula among these based on pre-treatment TSH (maximum value upto 100 μIU/ml) and body weight can be further simplified for easy usage as Initial Thyroxine Dose (μg/day) = TSH (μIU/ml)/2 + Body Weight (kg)/4 + 21. All of the above are strengths and novelties in this study. The current study suffered from some limitations also. First, serum T4 and anti-TPO antibody reports were not available for a few patients. Secondly, some patients did not report back for follow-up to our center and could not be included in the analysis. Further presence of pro-brain natriuretic peptide or other markers of cardiac stress would have provided objective evidence on the safety of the levothyroxine dosages. Nevertheless, the study is one of the larger studies on this topic as most other studies had a sample size lower than the current study.

Conclusion

Thyrotropin-based initial levothyroxine dose for treating primary hypothyroidism leads to euthyroidism at 6 weeks of treatment initiation in nearly 80% of patients. The dose required for adequate replacement of levothyroxine has a high correlation with pre-treatment TSH levels and a moderate correlation with serum thyroxine levels. Body weight also influences the adequate replacement dose. Linear as well as logarithmic dose calculator based on the above parameters is proposed, which needs to be validated in further studies.

Study highlights

1. What is the current knowledge on the topic?

   Higher level of pre-treatment TSH indicates lower residual function, but the variation in levothyroxine dose requirements according to the level of TSH are not clearly known.

2. What question did this study address?

   What proportion of patients will achieve euthyroidism if treated with a predetermined dose of levothyroxine based on TSH levels?

3. What does this study add to our knowledge?

   Thyrotropin-based initial levothyroxine dose for treating primary hypothyroidism leads to euthyroidism at 6 weeks of treatment initiation in nearly 80% of patients. Dosing regimens of thyroxine should take TSH levels into consideration apart from weight. A simple dose calculation formula, i.e., Initial Thyroxine Dose (μg/day) = TSH (μIU/ml)/2 + Body Weight (kg)/4 + 21 is proposed.

4. d.How might this change clinical pharmacology or translational science?

   Guidelines would consider individualized thyroxine dosage based on pre-treatment TSH level and body weight for the earlier achievement of euthyroidism and avoiding over-replacement.

Author contributions

Author Nair A conceived the study. Nair A, Jayakumari C, and Sarayu Soumya conducted the study. Nair A, Sujatha C, and Jabbar PK analyzed the data. Gopi A prepared the manuscript. All authors reviewed the manuscript.

Financial support and sponsorship

Study was conducted using departmental resources only and no specific funding was received.

Conflicts of interest

There are no conflicts of interest.