A Novel Approach to the Levothyroxine Absorption Test Using Only Two Free T4 Measurements

Study design and patients In this single-center, cross-sectional study, we extracted data from the patient’s electronic medical records (EMR) between 2017 and 2021. First, we identified for eligibility the records of adults (aged ≥ 18 years) with primary hypothyroidism who underwent an LT4 absorption test (LT4AT). LT4AT is usually performed in individuals with persistently high TSH levels despite LT4 doses of ≥ 1.9 mcg/kg. The records of pregnant women, patients with central hypothyroidism, and records with incomplete data were excluded. LT4 absorption test protocol LT4AT was conducted under the supervision of laboratory staff. Patients were asked to come after an overnight fast in the morning. Ingestion of medications that could interfere with LT4 absorption was inquired about before the test. All patients received a 1,000 mcg dose of LT4 (Eutirox®, Merck Healthcare KGaA) supplied as ten 100 mcg pills with two full glasses of water (400 mL) in the morning and were asked to remain seated for at least one hour. Swallowing of the tablets was supervised. Variables The following data were extracted from the EMR: demographic variables (age and sex), weight, height, body mass index (BMI), FT4 concentration at the beginning of the test and two and four hours after the LT4 dose, and medical prescription of medications before the test that could interfere with levothyroxine absorption, such as calcium and iron supplements, proton pump inhibitors, tyrosine kinase inhibitors, antiseizure medications, antidepressants, cholestyramine, and orlistat.15 In addition, relevant medical conditions were extracted, including a history of gastric bypass surgery, inflammatory bowel disease, atrophic gastritis, depression, anxiety disorder, lactose intolerance, and Helicobacter pylori infection. From the FT4 concentrations, the two and four-hour increases were computed. The qualitative absorption of levothyroxine was calculated using the following formula, adapted from Ghosh et al.:13 % Levothyroxine absortion = FT4∆ × 10 × VD LT4 dose in mg The FT4∆ is the difference between the basal and four-hour measurement of the FT4 in ng/dL. The original formula in Ghosh et al. used the total serum T4; however, a strong correlation has been reported between the total serum T4 and the FT4 increment.9 VD is the volume of distribution, which is calculated in dL using the following formula:

Serum TSH and FT4 were estimated using the chemiluminescence technique with commercially available kits from Abbot (Alinity I Free T and Alinity I TSH) (Abbott Laboratories, North Chicago, IL).8 Analytical sensitivity and total precision for TSH are 0.0083 mIU/L and 2.0%, respectively, and for FT4, are 0.42 ng/dL and 2.6%, respectively. The laboratory reference range for TSH is 0.35 mIU/L to 4.94 mIU/L, for FT4, it is 0.70 ng/dL to 1.48 ng/dL, and the inter-assay coefficient of variation are 1.5% and 2.2%, respectively.

Levothyroxine malabsorption was defined as an absorption percentage of less than 60%9 using the lower limit (0.70 ng/dL) of the expected LT4 absorption as the threshold. Hashimoto’s thyroiditis was confirmed by measuring antithyroid autoantibodies or typical ultrasonographic features. Atrophic gastritis was diagnosed using a gastric biopsy and an anti-parietal cell antibody test. Helicobacter pylori infection was confirmed by gastric biopsy or breath tests. A high levothyroxine dose was defined as a daily LT4 dose beyond 1.9 mcg/kg.16 Statistics methods Continuous variables are summarized using median and interquartile range (IQR), and categorical variables using absolute and relative frequencies. We used the Mann-Whitney test to compare basal FT4 with the two and four-hour measurements and the two and four-hour delta FT4. Comparison between samples was made using the Wilcoxon signed ranks test, and a 95% confidence interval (95%CI) was computed for the differences. Finally, we calculated an optimal threshold value for the four-hour FT4 delta that maximizes the absolute values of sensitivity and specificity after locally weighted scatterplot smoothing (LOEES), as it produces less bias estimation in small-sample studies and tests the out-of-sample performance of the cutoff point by bootstrapping using 500 samples.17 Finally, given that our definition of malabsorption is based on the qualitative LT4 absorption, we planned a sensitivity analysis to test the found cutoff point in two scenarios where the absorption percentage to define malabsorption was changed from less than 60% (original definition) to less than 80%9 and less than 98%.13 All studies were conducted using the R statistical package.18 As the levothyroxine absorption between the two-hour and four-hour marks was similar, both measurement points were compared. Ethics approval This study complied with the amendment made in 2013 during the general assembly in Fortaleza, Brazil, of the World Medical Association to the Declaration of Helsinki. This study was approved by the Ethics and Research from the sponsor institution. The committee did not require informed consent due to the study design. Only one of the researchers had access to patient identification data, whereas the remaining authors could only access a de-identified database.

Data were retrieved from the electronic medical records of 76 patients who underwent an FT4AT, (Figure 1). Most of the patients were female (N= 69), with a median age of 39 (IQR 16.5) years. Most patients (N=43) had autoimmune thyroiditis as the cause of hypothyroidism. Several patients were also treated with medications that could alter levothyroxine absorption or clearance. Patients’ characteristics are presented on Table 1.

Data presented as median with interquartile range or proportions. EMR: Electronic medical records, IQR: interquartile range.

The FT4 concentration increased significantly following LT4 loading. The median increase at two hours was 0.8 ng/dL (95%CI 0.75 to 0.95) and 1.0 ng/dL (95%CI 0.92 to 1.07) after four hours (Figure 2). The difference between the two and four-hour delta was 0.12 ng/dL (95%CI 0.05 to 0.19) (Table 2).

Presented as median with interquartile range. FT4: Free T4. The red line represents the lower reference limit of the FT4 test, 0.7 ng/ dL.6

FT4 qualitative absorption Using the four-hour increment of the FT4, the median LT4 qualitative absorption was 117% (IQR 52.6) (Table 2). Three subjects had a 60% or less qualitative absorption, labeled as true malabsorption. In the malabsorption group, the median qualitative absorption was 40.6 % (IQR 4.92), while in the pseudomalabsorption group, it was 118% (IQR 51.2). Only 1 patient diagnosed with true malabsorption was on treatment with a medication that potentially could interfere with levothyroxine absorption. Table 3 summarises the medications reported on the EMR.

True malabsortion (N=3) 0 (0%) 0 (0%) 0 (0%) 1 (33.3%) 0 (0%) 0 (0%)

FT4 four-hour delta cutoff point We found that 0.62 ng/dL increment of the FT4 after four hours of receiving 1,000 mcg of LT4 has 93.1% sensitivity and 100% specificity to differentiate pseudo-malabsorption from true malabsorption with an area under the curve of 0.98 (Table 4, Figure 3). Furthermore, the sensitivity analysis revealed that changing the malabsorption definition does not affect the cutoff point sensitivity and the AUC but reduces its specificity (Table 4).

AUC: area under the curve, PPV: Positive predictive value, NPV: Negative predictive value. Originally planned analysis, malabsorption was defined as less than 60% of levothyroxine absorption. Sensitivity analysis 1, malabsorption was defined as less than 80% of levothyroxine absorption. Sensitivity analysis 2, malabsorption was defined as less than 98% of levothyroxine absorption.

Figure 3. ROC Curve for FT4 four-hour delta cutoff point 1.00 0.75

Two and four-hour absorption comparison The two-hour absorption was 16% lower (95%CI -26.03 to -8.00) compared to the four-hour absorption. The optimal increment for diagnosis at the two-hour mark was a 0.58 ng/dL increase of the FT4, after four hours of receiving 1,000 mcg of LT4 has 92% sensitivity and 92% specificity to differentiate pseudo-malabsorption from true malabsorption with an area under the curve of 0.98.

We presented an alternative to traditional LT4AT. In a group of patients with high LT4 requirements, our results indicate that the duration of the LT4AT could be reasonably reduced to four hours, and a delta greater than ≥ 0.62 ng/dL from baseline suggests pseudomalabsorption. In our study, the thyroid hormone delta measurement and absorption percentage calculation were performed using FT4 values. Furthermore, although the original LT4 absorption studies were validated using total thyroxine TT4 measurements, the use of FT4 in LT4AT was validated by Sun et al. in 2014. The correlation between FT4 and TT4 was 0.88 (95% confidence interval, 0.56 to 0.97;

P<0.001). Therefore, the authors concluded that FT4 may be used interchangeably with TT4 for quantitative assessment of suspected malabsorption using oral LT4AT.6,9 LT4AT is an essential tool that can distinguish between non-adherence and true intestinal malabsorption of LT4. There is no gold standard method for LT4AT with various protocols advocated in the literature.12, 19, 20 The time commitment for implementation ranged from two to five days. Marques et al., in a study published in 2016, indicated that a 2.5 increase in FT4 from the baseline is suggestive of pseudomalabsorption. However, in contrast to our findings, a previous study did not determine the sensitivity and specificity of the FT4 cutoff point using ROC curve analysis, which prevented the determination of a good cutoff point. In addition, the sample size was small.21 Similar to our results, Gonzales et al., in a study performed with 16 patients diagnosed with refractory hypothyroidism who had completed the LT4AT, showed that most patients who completed the LT4AT had normal intestinal absorption at four hours. Nevertheless, this study characterized normal absorption by calculated absorptions of ≥60%, corresponding to an increase in TT4 toward peak levels.10 Previous studies used the computed percentage of LT4 absorption to define normal absorption.22, 23 However, the use of this formula is controversial, and its calculation is complex for routine use by clinicians, thus limiting its use. Our findings support a simplified interpretation of FT4AT, using only the delta of FT4 at four hours. Most subjects in our study had basal FT4 levels within the normal range. Similarly, Van Wilder et al.24 described cases in which a subject presented FT4 levels close to the reference range, which could indicate intentional ingestion before testing. However, during LT4AT with supervised intake, an increase in FT4 within the first 120 min, with the maximum level observed at four hours, strongly contradicts the true malabsorption.25 The possibility of true malabsorption should be evaluated when the absorption percentage is less than 60% or when there is no increase in FT4.20, 26, 27 Also, although there is only a minor difference in levothyroxine absorption between the two-hour and four-hour marks, the two-hour delta has lower specificity compared to the four-hour. However, the expected time interval for the absorption of LT4 by the small intestine is within the first 120 min.28 Nevertheless, we also described the FT4 four-hour delta cutoff point because our cohort reached maximum levels at four hours. Therefore, adopting this simplification may facilitate and increase the use of LT4AT and reduce costs by measuring FT4 only at four hours. Following the diagnosis of pseudomalabsortion, a mental health evaluation should be performed. Psychiatric disorders such as depression and anxiety are associated with autoimmune hypothyroidism.29, 30 In addition, some patients exhibit neuropsychiatric symptoms even after successful treatment of hypothyroidism.31 In contrast, others persist in low compliance with levothyroxine and are at risk of serious complications that lead to unnecessary medical and surgical procedures.32 Therefore, supervised weekly doses of LT4 have been proposed to improve treatment adherence in this subgroup of patients.33, 34, 35

Strengths, limitations, and final remarks Our study constitutes the largest cohort of patients published on this topic. Furthermore, we established an FT4 four-hour delta cutoff point. This could make it easier to perform the LT4AT and facilitate its use in routine care. The limitations of our study include its retrospective nature, the absence of a comparison with wellcontrolled hypothyroid individuals, and the small number of patients with true malabsorption, which restricted our findings. In addition, the lack of follow-up data made it difficult to establish the optimal dose for achieving euthyroidism and to determine whether patients should undergo psychiatric evaluation after being diagnosed with pseudomalabsorption. We had also planned to conduct several subgroup analyses for more prevalent conditions and medications that affect levothyroxine absorption, but the sample size was not large enough for this. Therefore, our results need to be confirmed in further research with a larger sample size. Ultimately, although the results may not be generalizable, they could serve as a starting point for future larger prospective cohorts.

Here, we present a simplified protocol for diagnosing pseudomalabsorption. The increase of FT4 to 0.62 ng/dL at four hours in people with a baseline FT4 below the reference range suggests pseudomalabsorption without needing to calculate the absorption percentage. We emphasize the importance of early LT4AT application in cases of refractory hypothyroidism, given the benefits of the test in distinguishing between pseudomalabsorption and malabsorption and, as a result, guiding clinical decisions in this context. Conflicts of interest The authors certify that they have no affiliation or are involved in any organization or entity with any financial interest (such as fees, financial aid for education, shares, employment contracts, work as consultants, or any other type of interest) or non-financial interest (such as personal or professional relationships, affiliations, or beliefs) on the topic of interest or any material discussed in this manuscript. Carlos A. Builes-Barrera received consulting and speaker fees from Merck, Abbott, Sanofi, Novo Nordisk, Novartis, Bayer, Merck Sharp and Dhome, Lilly, Amgen, Boehringer Ingelheim, Siegfried and Lafrancol. Carlos E. Builes-Montaño received consulting and speaker fees from Sanofi, Novo Nordisk, Novartis, and Boehringer Ingelheim. María Carolina Fragozo-Ramos received financial support from Novo Nordisk and Ipsen for the medical events.