Liquid or soft-gel levothyroxine vs tablet levothyroxine in athyreotic adults after total thyroidectomy: Systematic review and meta-analysis

Abstract

BACKGROUND

Achieving stable thyroid hormone replacement after total thyroidectomy is challenging owing to the variability in levothyroxine (LT4) absorption. Newer liquid and soft-gel formulations may offer improved bioavailability compared to conventional tablets, but their benefits in thyroidectomized patients remain uncertain.

AIM

To compare the effectiveness of liquid and soft-gel LT4 with tablet LT4 in athyreotic adults following total thyroidectomy.

METHODS

A systematic review and meta-analysis were conducted according to the PRISMA 2020 guidelines. The MEDLINE, EMBASE, CENTRAL, and Scopus databases were searched from inception to January 1, 2026. Randomized controlled trials and cohort studies enrolling adults after total thyroidectomy were included. The primary outcome was serum thyroid-stimulating hormone level. Secondary outcomes included circulating total thyroxine and triiodothyronine levels. Standardized mean differences (SMD) with 95% confidence intervals (CIs) were pooled using a random effects model. Subgroup analyses were performed based on the disease etiology, malabsorption status, and study design.

RESULTS

Nine studies comprising 1115 participants were included. Liquid or soft-gel LT4 was associated with significantly lower thyroid-stimulating hormone level compared with tablets (SMD: -0.91; 95%CI: -1.52 to -0.30; P = 0.003), with high heterogeneity (I² = 95%). Subgroup analysis showed a significant effect in patients treated for benign disease but not in those with differentiated thyroid cancer. A significant benefit was observed in patients without documented malabsorption, whereas no statistically significant difference was found in those with malabsorption. Liquid LT4 resulted in higher total thyroxine (SMD: 0.34; 95%CI: 0.15-0.52) and triiodothyronine levels (SMD: 0.25; 95%CI: 0.06-0.44), with no heterogeneity.

CONCLUSION

While observational data suggest improved control, this benefit was not replicated in randomized controlled trials, highlighting the need for further high-quality trials. Routine substitution for all patients was not supported, highlighting the need for individualized LT4 therapy.

Key Words: Levothyroxine formulations; Thyroidectomy; Thyroid-stimulating hormone level outcomes; Thyroid hormone replacement; Triiodothyronine

Core Tip: Levothyroxine (LT4) is one of the most widely prescribed medications for hypothyroidism globally; however, its narrow therapeutic index requires careful dose individualization to avoid the risks of iatrogenic hyperthyroidism or hypothyroidism. In this systematic review and meta-analysis of athyreotic adults following total thyroidectomy, liquid or soft-gel LT4 was associated with significantly lower thyroid-stimulating hormone levels and modest but consistent increases in circulating total thyroxine and triiodothyronine levels compared with tablet formulations. These findings support a personalized approach to LT4 therapy, rather than the routine substitution of formulations, in all thyroidectomized patients.

INTRODUCTION

Levothyroxine (LT4) is one of the most widely prescribed medications for hypothyroidism globally; however, its narrow therapeutic index requires careful dose individualization to avoid the risks of iatrogenic hyperthyroidism or hypothyroidism[1]. Serum thyroid-stimulating hormone (TSH) remains the most reliable marker of thyroid hormone homeostasis; however, achieving and maintaining target TSH levels in patients being treated for hypothyroidism is challenging[2]. Even with standard dosing of 1.6-1.8 μg/kg/day, nearly half of patients fail to reach their therapeutic goal, and many require dose increases of up to 30%-40%[3,4]. Multiple factors affect LT4 absorption. It is primarily absorbed in the small intestine[5], with minimal absorption occurring in the stomach. As a result, patients who undergo small intestinal resection often require higher doses[6]. In euthyroid individuals, the time to reach maximum concentration is approximately 2 hours, with a bioavailability of roughly 60%-80%, while in hypothyroid patients, it is delayed to approximately 3 hours with a slightly higher bioavailability[7]. Food intake also affects absorption, with bioavailability dropping from approximately 79% under fasting conditions to approximately 64% for a 100-μg dose[8,9]. Concomitant medications such as proton-pump inhibitors and calcium supplements can also affect absorption and further complicate therapy[10].

These challenges are particularly pronounced in patients who have undergone total or near-total thyroidectomy[11], where LT4 replacement is the sole source of thyroid hormones. In this population, precise dosing is critical not only to restore euthyroidism but also, in cases of differentiated thyroid cancer (DTC), to achieve long-term TSH suppression. While suppressive therapy may reduce the recurrence risk, it must be balanced against cardiovascular and skeletal complications. The Current American Thyroid Association guidelines recommend a graded, risk-adapted approach that underscores the importance of individualized therapy[12,13].

In addition to dose and timing, increasing attention has been paid to LT4 formulations. Newer presentations, including oral solutions and soft gel capsules, have demonstrated improved absorption compared to traditional tablets, particularly in patients with gastrointestinal comorbidities or concomitant drug use[14]. Large-population-based studies and smaller clinical trials suggest that switching to liquid formulations may reduce variability in TSH levels, which can be helpful in reducing the need for repeated dose adjustments and can decrease the need for repeated monitoring in these patients[15]. However, findings across individual studies remain heterogeneous, and the magnitude of the benefit in thyroidectomized patients has not been fully established. While a previous meta-analysis and some clinical trials reported a superior absorption profile of liquid LT4, others found little or no difference between the tablet and liquid formulations[16]. Moreover, no previous meta-analyses have analyzed outcomes in thyroidectomized patients.

To address these uncertainties, we conducted a meta-analysis to compare LT4 formulations in thyroidectomized patients, examine absorption-related influences, and provide evidence-based recommendations for individualized treatment of high-risk individuals.

MATERIALS AND METHODS

This systematic review and meta-analysis were conducted in accordance with the PRISMA 2020 statement[17], and the protocol was prospectively registered in the International Prospective Register of Systematic Reviews[18].

Study selection and inclusion/exclusion criteria

The authors included studies enrolling adults aged ≥ 18 years who were asymptomatic following total thyroidectomy for benign disease or differentiated thyroid carcinoma. Pregnant women, pediatric patients, and individuals with residual thyroid tissue were excluded. The intervention of interest was oral liquid LT4, either ethanol-containing, ethanol-free, or soft-gel LT4, administered at clinically used doses. The comparator used was an oral LT4 tablet. The primary outcome was the difference in TSH levels between the formulations. The secondary outcomes included changes in the triiodothyronine (T3) and total thyroxine (T4) levels. Eligible study designs included randomized controlled trials (RCTs) and prospective cohort studies.

Search strategy

We systematically searched MEDLINE (via PubMed), EMBASE, Cochrane Central Register of Controlled Trials, and Scopus from inception to January 1, 2026. A comprehensive search strategy was developed with input from a medical librarian, combining controlled vocabulary [Medical Subject Headings (MeSH)] and free-text keywords related to LT4 formulations (liquid, soft gel, capsule, tablet), thyroidectomy, and TSH outcomes. No language restrictions were applied. The reference lists of included studies were manually searched to identify additional reports. The complete search strategy for all databases is provided in search strategy in Supplementary material.

Data extraction

Two reviewers independently extracted data using a standardized form. Extracted variables included study characteristics, participant demographics, intervention details (ethanol content, dose, timing), and outcomes (mean and SD for TSH, free thyroxine, and free T3). For studies reporting medians and interquartile ranges, the sample mean was estimated using the formula x ≈ (Q1 + M + Q3)/3. The SD was derived from the interquartile ranges by dividing it by 1.35, in accordance with the validated methods described by Wan et al[19] and the Cochrane Handbook.

Statistical analysis

The difference in TSH levels was calculated as the standardized mean difference (SMD) with 95% confidence intervals (CIs). Given the inclusion of both randomized trials and observational studies, a random-effects model was applied to account for anticipated inter-study heterogeneity. Heterogeneity was assessed using I² statistic and Cochran’s Q test. Sensitivity analyses were performed using the leave-one-out method to assess the robustness of the pooled estimates by sequentially excluding one study at a time and recalculating the effect size[20]. Two primary subgroup analyses were pre-specified: (1) Presence of malabsorption of the population; and (2) Thyroid disease etiology (DTC vs benign disease). All analyses were conducted using the RevMan software. Small-study effects and publication bias will be assessed through the funnel plots to observe the trends. However, if the included studies are less than ten, it limits the power and interpretability of such visual assessments.

Risk of bias and certainty of evidence

Two reviewers independently assessed risk of bias. The Cochrane RoB 2 tool was applied[21]. The remaining six observational and cohort studies were evaluated using the ROBINS-I tool[22]. Specific attention was paid to confounding domains and participant selection, given the inherent limitations of nonrandomized designs. Finally, the certainty of evidence for each outcome was graded using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach, considering the risk of bias, inconsistency, indirectness, imprecision, and publication bias[23].

RESULTS

Study selection

The systematic search strategy identified 636 records, which were subsequently reduced to 434 after duplicate entries were removed. Screening of titles and abstracts led to the retrieval of 16 full-text reports for an intensive eligibility assessment. Of these, seven reports were excluded, four due to inappropriate intervention, and three due to an ineligible study population. Ultimately, 9 studies met the rigorous inclusion criteria and were included in this meta-analysis. The complete screening and selection process was delineated in the PRISMA flow diagram (Figure 1).

Figure 1  PRISMA flow diagram illustrating the study selection process.

Study characteristics

This systematic review synthesized data from nine studies published from 2014 to 2025, comprising a total of 1115 participants. The included studies consisted of three RCTs and six cohort studies, with a mix of prospective and retrospective designs. All nine studies were conducted in Italy. The study population varied significantly in clinical presentation, including patients with DTC, benign goiter, and autoimmune thyroiditis. The follow-up durations across the studies ranged from 10 weeks to 54 weeks, providing both short- and medium-term evidence for the comparative efficacy of LT4 formulations, as shown in Table 1[24-32].

Table 1 Characteristics of the included studies.

Ref.	Study design	Follow-up (weeks)	n (total)	Female (%)	Age (years)	BMI	Thyroid-stimulating hormone baseline	Malabsorption	Disease
Brancato et al[24], 2014	Retrospective cohort	14	53	94.3	45	26.8	3.04	35.9%	Thyroiditis (66%), thyroidectomy (26%), goiter (8%)
Negro et al[25], 2014	Prospective cohort	54	200	81	53	-	1.71	None	Surgical (29%) or autoimmune (71%)
Vita et al[26], 2014	Prospective cohort	27	24	75	56.2	-	4.1	100%	Primary hypothyroidism (58%) or DTC (42%)
Cappelli et al[27], 2017	RCT	54	102	78.4	49.4	24.1	0.28	None	DTC
Fallahi et al[28], 2016	Prospective cohort	54	21	81	51	24.4	7.6	None	Autoimmune thyroiditis (62%), nodular goiter (38%)
Lombardi et al[29], 2017	RCT	10	155	84.5	51.5	26.4	8	None	Benign diseases only
Fallahi et al[30], 2018	Prospective cohort	15	105	77	48	24.5	2.55	None	Thyroid cancer only (papillary/follicular)
Baratta et al[31], 2025	RCT	10	130	100	51.6	26.4	8	None	Thyroidectomized patients
Gatta et al[32], 2025	Retrospective cohort	13	275	78.2	53	25.7	1.2	27.6%	DTC (28%) vs benign (72%)

BMI: Body mass index; RCT: Randomized controlled trial; DTC: Differentiated thyroid cancer.

Primary outcome: TSH levels

The pooled analysis of all nine included studies, encompassing a total of 1115 participants, demonstrated that liquid LT4 was associated with significantly lower TSH levels than the tablet formulation (SMD: -0.91; 95%CI: -1.52 to -0.30; P = 0.003) (Figure 2). Despite the statistical significance of the overall effect, a high degree of interstudy heterogeneity was observed (I² = 95%). A leave-one-out sensitivity analysis confirmed the robustness of the primary findings; notably, the exclusion of the study by Fallahi et al[28] reduced the magnitude of the effect, but heterogeneity remained high and the results remained statistically significant (SMD: -0.54; 95%CI: -1.08 to -0.01; P = 0.05) (Supplementary Figures 1 and 2).

Figure 2 Forest plot comparing thyroid-stimulating hormone levels between liquid or soft-gel and tablet levothyroxine in all included studies. CI: Confidence interval.

To investigate the sources of this variance, prespecified subgroup analyses were performed. Analysis based on the etiology of thyroid disease indicated a significant TSH reduction in the benign and hypothyroidism subgroups (SMD: -1.30; P = 0.003), whereas the difference in the DTC subgroup did not reach statistical significance (SMD: -0.26; P = 0.55) (Supplementary Figure 3). Further stratification by clinical status revealed that while the liquid formulation significantly reduced TSH levels in patients without malabsorption (SMD: -1.25; P = 0.003), the effect was not significant in those with confirmed malabsorption (SMD: -0.31; P = 0.47) (Figure 3A). Additionally, subgrouping by study design showed a significant effect in cohort studies (SMD: -0.99; P = 0.01), which was not replicated in the RCT subgroup (SMD: -0.77; P = 0.19) (Figure 3B). Visual inspection of the funnel plot showed no evidence of publication bias, but power of this finding is limited as less than ten studies reported TSH outcome (Supplementary Figures 1 and 2).

Figure 3 Subgroup analysis of thyroid-stimulating hormone levels. A: Subgroup analysis of thyroid-stimulating hormone levels according to malabsorption status; B: Subgroup analysis of thyroid-stimulating hormone levels according to study design (randomized controlled trials vs cohort studies). CI: Confidence interval.

Secondary outcome: T4 levels

Evaluation of circulating thyroxine levels across five studies indicated that liquid LT4 administration resulted in significantly higher T4 concentrations than tablet LT4 (SMD: 0.34; 95%CI: 0.15-0.52; P = 0.0003) (Figure 4). In contrast to the primary outcome, this analysis demonstrated excellent consistency across the included reports with no evidence of statistical heterogeneity (I² = 0%).

Figure 4 Forest plot comparing circulating total thyroxine levels between liquid or soft-gel and tablet levothyroxine. CI: Confidence interval.

Secondary outcomes: T3 levels

In alignment with the T4 results, the meta-analysis of four studies reporting T3 levels revealed a statistically significant increase in favor of the liquid formulation (SMD: 0.25; 95%CI: 0.06-0.44; P = 0.010) (Figure 5). Similar to the T4 analysis, the T3 outcome showed a high degree of homogeneity (I² = 0%).

Figure 5 Forest plot comparing circulating triiodothyronine levels between liquid or soft-gel and tablet levothyroxine. CI: Confidence interval.

Risk of bias assessment

The methodological quality of evidence was assessed using two distinct frameworks. For the three RCTs, the Cochrane RoB 2 tool was utilized, with all studies (Cappelli et al[27], Lombardi et al[29], Baratta et al[31]) being judged as having a low risk of bias across all assessed domains (Figure 6). Six nonrandomized studies were evaluated using the ROBINS-I tool. One study (Negro et al[25]) was classified as having a low overall risk of bias, whereas the remaining five (Brancato et al[24], Vita et al[26], Fallahi et al[30], Fallahi et al[28], Gatta et al[32]) were determined to have a moderate risk of bias (Figure 7). The moderate risk ratings were predominantly driven by potential confounding factors in participant selection and the open-label nature of the interventions, which are inherent limitations in observational clinical research.

Figure 6  Risk of bias assessment for randomized controlled trials using the Cochrane RoB 2 tool.

Figure 7  Risk of bias assessment for non-randomized studies using the ROBINS-I tool.

Certainty of evidence

Using the GRADE framework, the certainty of evidence was rated as low for the primary outcome, the TSH level. This was mainly because of the serious risk of bias and heterogeneity among the included studies. For the secondary outcomes, T3 and T4 levels, certainty of evidence was rated as moderate due to the moderate risk of bias found in the included RCTs. The full GRADE assessment results are presented in Table 2.

Table 2 Certainty of evidence using the Grades of Recommendation, Assessment, Development and Evaluation approach.

Outcomes	Anticipated absolute effects (95%CI)	n of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Liquid levothyroxine
Thyroid-stimulating hormone	SMD 0.91 lower (1.52 lower to 0.3 lower)	1115 (9 studies)	Lowa,b	Liquid levothyroxine may lead to a greater reduction in thyroid-stimulating hormone levels compared to tablet levothyroxine, but the evidence is uncertain
T4 levels	SMD 0.34 higher (0.15 higher to 0.52 higher)	480 (5 studies)	Moderatea	Liquid levothyroxine probably results in slightly higher circulating T4 levels compared to tablet levothyroxine
T3 levels	SMD 0.25 higher (0.06 higher to 0.44 higher)	432 (4 studies)	Moderatea	Liquid levothyroxine probably results in slightly higher T3 levels compared to tablet levothyroxine

^aMost included studies had a moderate risk of bias.

^bHigh heterogeneity found in analysis.

CI: Confidence interval; GRADE: Grades of Recommendation, Assessment, Development and Evaluation; SMD: Standardized mean differences; T4: Total thyroxine; T3: Triiodothyronine.

DISCUSSION

This systematic review and meta-analysis evaluated the comparative effectiveness of liquid or soft gel vs LT4 tablets in athyreotic adults following total thyroidectomy. Our pooled analysis of nine studies, including 1115 patients, demonstrated that liquid formulations were associated with significantly lower serum TSH levels and modest but consistent increases in circulating T4 and T3 concentrations. These findings support the hypothesis that non-tablet formulations provide a more efficient and predictable absorption. The pharmacokinetic advantages of liquid and soft-gel LT4, including rapid dissolution and reduced dependence on gastric acidity, have been well-described in both experimental and clinical settings[33-35]. A prior systematic review and meta-analysis reported significant TSH reductions following conversion from tablet to liquid LT4, particularly in patients with impaired absorption or unstable biochemical control[36]. Our study extends this evidence to thyroidectomized populations, a group in which exogenous LT4 is the sole source of thyroid hormones and in whom stable biochemical control is essential for both replacement and suppressive therapy.

Despite the overall favorable effect, substantial heterogeneity was observed, prompting subgroup analyses to provide clinically meaningful insights. A significant reduction in TSH levels was observed in patients treated for benign thyroid disease, whereas no significant difference was detected in patients with DTC. This probably reflects the different therapeutic targets among the populations. In DTC, LT4 is often administered at suppressive doses to maintain subnormal TSH levels, thereby limiting the ability of formulation changes to reduce TSH[13,37] further. In contrast, patients treated for benign diseases aim for euthyroidism, where improvements in absorption consistency may translate into more stable biochemical control. Observational studies have reported reduced intra-individual TSH variability and fewer dose adjustments after switching to liquid LT4 in non-cancer populations[15,27,38]. These findings suggest that the formulation choice may be more impactful in patients receiving replacement therapy than in those undergoing suppressive treatment for malignancy.

Unexpectedly, the subgroup of patients with documented malabsorption did not demonstrate a statistically significant benefit from liquid formulations, whereas a clear advantage was observed in patients without malabsorption. Several factors may explain this finding, including limited sample size, heterogeneous definitions of malabsorption, and the inclusion of patients with diverse etiologies, such as autoimmune gastritis, medication interference, and postsurgical gastrointestinal disorders, each of which may affect LT4 bioavailability differently[9,39]. Furthermore, some observational studies preferentially enrolled patients who had already failed tablet therapy, potentially exaggerating the apparent benefit of switching formulations. Nonetheless, multiple mechanistic and clinical studies support the biological plausibility that liquid and soft-gel LT4 formulations are less susceptible to pH-dependent dissolution and interactions with food, calcium, iron, and proton-pump inhibitors[40-42]. Improved absorption has been demonstrated in patients with lactose intolerance, Helicobacter pylori infection, autoimmune gastritis, and acid-suppressive therapy[43,44]. While our pooled estimate in the malabsorption subgroup did not reach significance, the consistency of external evidence suggests that liquid formulations remain a rational option for carefully selected patients.

Analysis according to study design revealed that cohort studies demonstrated a significant benefit of liquid LT4, whereas RCTs did not. The lack of statistical significance in the RCT subgroup suggests that the overall positive effect may be partially driven by observational data. In cohort studies, clinicians often switch patients to liquid formulations precisely because they are struggling with tablet absorption, leading to ‘selection bias’. Conversely, RCTs utilize standardized protocols that may minimize the real-world absorption challenges liquid formulations are designed to solve. Observational studies frequently include patients with unstable TSH levels or inadequate response to tablets, which may overestimate treatment effects through selection bias and regression to the mean. In contrast, RCTs typically enroll more homogeneous populations with standardized dosing, timing, and monitoring, thereby reducing the likelihood of detecting modest formulation-related differences. The very high heterogeneity observed for TSH outcomes mirrors that reported in previous meta-analyses of LT4 formulations and reflects variations in baseline TSH, follow-up duration, ethanol content of liquid preparations, dosing strategies, and inclusion of both replacement and suppressive therapy populations[36,45]. Although sensitivity analyses confirmed the robustness of the primary findings, the magnitude of the effect was dependent on individual studies, underscoring the need for larger, multicenter RCTs with standardized protocols.

Beyond TSH, our analysis demonstrated significantly higher T4 and T3 levels with liquid LT4, with no observed heterogeneity, supporting the concept of improved systemic availability of thyroid hormones. Importantly, the modest increase in T3 suggests enhanced peripheral conversion rather than supraphysiological exposure. Previous pharmacokinetic studies have similarly demonstrated higher post-dose serum T4, earlier time to peak concentration, and more consistent absorption with liquid LT4 than with tablets[46,47].

Despite its strengths and the observed benefits, the study has some notable limitations. The substantial heterogeneity (I² = 95%) likely stems from variations in therapeutic targets (suppressive vs replacement), differing follow-up durations (10 weeks to 54 weeks), and the inclusion of both prospective and retrospective designs. The conceptual heterogeneity introduced by pooling patients with benign disease and DTC is also a notable limitation. In DTC, the narrow window for TSH suppression may mask the potential absorption benefits of liquid formulations, whereas the broader targets in benign disease allow these pharmacokinetic advantages to manifest as significant TSH reductions. Another important limitation of this meta-analysis is the geographical homogeneity of the data; all included studies were conducted in Italy. This limits the generalizability of our findings to other populations with different dietary habits (e.g., varying iodine or soy intake), differing healthcare infrastructures, and different generic tablet bioequivalence standards. Furthermore, this review primarily evaluates biochemical markers (TSH, T4, T3) rather than clinical endpoints. While biochemical stability is crucial, the impact of liquid LT4 on patient-reported outcomes, such as quality of life, symptom resolution, or the frequency of long-term dose adjustments, remains under-researched in thyroidectomized populations.

Clinically, this may translate into improved biochemical stability, without increasing the risk of overtreatment. Taken together, these findings support a targeted, individualized approach for LT4 formulation selection in thyroidectomized patients. Liquid or soft-gel LT4 appears to be most beneficial in patients with unstable TSH, difficulty achieving euthyroidism on tablets, or those receiving replacement therapy for benign disease, while routine substitution in all thyroidectomized patients is not supported. Future adequately powered, multicenter RCTs stratified by disease etiology, absorption status, and therapeutic targets are required to confirm these findings, evaluate long-term outcomes, and provide evidence-based recommendations for personalized LT4 therapy.

CONCLUSION

In this systematic review and meta-analysis of athyreotic adults following total thyroidectomy, liquid or soft-gel LT4 was associated with significantly lower TSH levels and modest but consistent increases in circulating T4 and T3 levels compared with tablet formulations. However, the clinical benefits were not uniform across all the patient groups. The advantage of liquid formulations was most evident in patients treated for benign thyroid disease and those receiving replacement therapy. No significant difference was observed in patients with DTC undergoing TSH-suppressive treatment. Despite biological plausibility and pharmacokinetic profile, this meta-analysis did not find statistical evidence supporting switching based solely on malabsorption status. These findings support a personalized approach to LT4 therapy, rather than the routine substitution of formulations, in all thyroidectomized patients. Large, well-designed, multicenter randomized trials are required to confirm these results, clarify subgroup effects, and determine whether improved biochemical stability with liquid formulations translates into meaningful long-term clinical outcomes.