Thyroid axis adaptations to moderate short-term energy restriction in healthy, young women

Abstract

Objective

Short-term dieting has gained popularity in women. We studied thyroid hormone change after short-term, moderate energy restriction.

Methods

Nineteen, healthy women aged 23.36 ± 2.08 yr (mean ± SD) without obesity and thyroidal disease underwent a neutral (NEA, ±0%) and a deficient (DEA, −55%) diet for 5 days each in the early follicular phases of successive menstrual cycles. Blood was sampled every 10 min between 8 AM and 4 PM and analyzed for TSH, GH, and cortisol every 30 min, total T3 (TT3), reverse T3 (rT3), and total T4 (TT4) hourly and free T3 (fT3), free T4 (fT4), and TBG at the beginning and end of the studies. Liquid chromatography-tandem mass spectrometry (LCMS) assessed all thyroid hormones except TSH and TBG. Data shown as mean difference or least squared (ls) mean (±SEM).

Results

There was a small decrease in body mass index and body weight after DEA (0.4 ± 0.08 kg/m²; P < .001 and 1.1 ± 0.21 kg; P < .001, respectively), with unchanged fat mass. Compared to NEA, TT3 (ls mean ± SEM; 95.55 ± 2.89 ng/dL vs. 89.15 ± 2.89; P < .0001), and TSH (1.03 ± 0.07 vs. 0.92 ± 0.07 μIU/mL; P < .0001) declined while TT4 (6.06 ± 0.25 vs. 6.26 ± 0.25 μg/dL; P = .005), fT4 (1.71 ± 0.07 vs. 1.83 ± 0.07 ng/dL; P = .0052) and rT3 (9.02 ± 0.56 vs. 12.04 ± 0.56 ng/dL; P < .0001) increased after DEA with no change in TBG, GH and cortisol.

Conclusion

Adaptive central and peripheral changes in thyroid hormones occur after short-term, moderate dieting in young women without obesity.

ClinicalTrials.gov

NCT02858336

Significance statement.

Short term dieting is prevalent, occurring in 36% of normal and underweight women and may contribute to the population prevalence of functional hypothalamic amenorrhea (FHA). Thyroid adaptations after dieting have been challenging to study using immunoassays due to multiple confounders. After 5 days of a deficient (−55%) caloric intervention, 19 healthy euthyroid women without obesity experienced a reduction in TSH. There was also a reduction in T3, and increases in T4 and rT3 measured using advanced liquid chromatography-mass spectrometry (LCMS) assays. These findings suggest central and peripheral adaptations of the thyroidal axis to even short-term, modest caloric restriction directed at conserving energy expenditure in response to reduced energy availability. This first report of LCMS thyroidal changes in women without obesity has important implications for understanding the complex hormonal adaptations to dieting.

Introduction

Adoption of the modern lifestyle has led to a surge in the rates of obesity and being overweight and their associated complications,^1,2 along with an increasing number of potential weight loss strategies. Such strategies range from restricting some energy intake on a number of days to fasting on some days, typically over 1–4 week duration.³

In addition to improving the associated risk of diabetes and metabolic syndrome in overweight/obese individuals, energy restriction, whether due to dieting or increased exercise, results in a switch from carbohydrate to free fatty acid metabolism. These changes have been implicated in health advantages such as reduced risk of cancer, cardiovascular disease and increased longevity, even dissociated from weight loss.^4-6 Importantly, results from a national survey in the United States indicate that attempts at weight loss are not limited to obese individuals but are also seen in 36% of normal or underweight women.⁷ Among young, educated women, misconceptions about body image coupled with a lack of awareness of healthy exercise and eating behaviors may drive the prevailing restrictive eating tendencies.^8-10 Such restrictive eating accompanied by unchanged or increased energy expenditure may contribute to the pathogenesis of functional hypothalamic amenorrhea (FHA). The population prevalence of FHA is 4.6% in women of childbearing age¹¹ although it is dramatically higher in association with exercise training and even subclinical eating disorders.^12-14 Published data indicates an association between “low-T3 syndrome” and FHA^15,16 that occurs at a submaximal dose of decreased energy availability of ∼20 kcal/kg of lean body mass (LBM)/day.^15,17

Both chronic and intermittent energy restriction are associated with homeostatic adaptations that resist the effects of energy restriction.^18-20 The thyroid hormone axis plays a critical role in metabolism and is tightly controlled at the hypothalamic, pituitary, thyroid, and peripheral levels with a degree of integration that is not seen in other hormonal axes. Deiodinase activity is a unique feature of thyroid physiology that regulates peripheral bioactive T3 availability and feedback at the hypothalamic-pituitary levels.²¹ Additional contributors to the bioavailability of thyroid hormones are transport proteins such as thyroid-binding globulin, the chief transporter of T3 and T4, along with minor contributions from transthyretin and albumin.²² Previous studies have examined the thyroid hormone effects of prolonged energy restriction, complete fasting or other diet regimens which resulted in significant changes in body composition.^23-25 In obese and overweight individuals, these adaptive changes in thyroid hormone dynamics may well contribute to the well-known difficulties in losing weight and maintaining weight loss. However, despite the prevalence of dieting in women without obesity, there is much less information on how modest short-term energy restriction affects thyroid hormone dynamics within this demographic. In all previous studies, immunoassays, which are readily influenced by nonthyroid factors such as heterophilic antibodies and biotin supplements, were used to assess thyroid hormones, making them less reliable than liquid chromatography-mass spectrometry (LCMS) for measurement of T3, T4 and their free fractions.^26-28

In this study, we determined the effects of 5 days of neutral energy intake compared with a 55% reduction in energy availability of the same duration on thyroid hormone dynamics measured across the day using LCMS to measure T3 and T4 and their free fractions.

Methods

Study participants

Participants were healthy, nonsmoking, sedentary women with regular menstrual cycle lengths between 25 and 35 days. Participants were between the ages of 18–28 years, and all were ≤14 years post menarche. The body mass index (BMI) was 18.5–27 kg/m² at the time of screening. Women were excluded if they had an energy intake of <35 or > 55 kcal/kg LBM/day determined by 4-day food diaries, were dieting, had experienced a significant weight loss (>2 kg) over the preceding 3 months, or were nonsedentary [maximal oxygen consumption (VO_2max) > 40 mL/kg/min or rigorous daily exercise defined as >4 h of aerobic exercise per week]. Women were also excluded if they were taking prescribed or over-the-counter hormones (including hormonal birth control), sleep medications or supplements (eg, biotin) that might interfere with the results of the study. All subjects had normal hemoglobin, thyroid stimulating hormone, testosterone, and prolactin levels. The NIH Institutional Review Board approved this study (IRB ID: 16E0157/CR001873, Continuing Review Approval Dated: July 11, 2024) and all participants provided informed consent prior to study procedures. All study procedures complied with ethical principles outlined in the Declaration of Helsinki.

Protocol

The study was designed as a paired, nonrandomized study of 2 dietary interventions administered in the early follicular phase (EFP) of 2 separate menstrual cycles in each participant to evaluate the endocrine and reproductive effects of short-term energy restriction. Eligible women were exposed to a neutral energy availability (NEA) and a deficient energy availability diet (DEA, −55.55%) for the first 5 days of 2 successive menstrual cycles followed by blood sampling in EFP on day 5 of both dietary interventions (Figure 1).

Figure 1.

Study schema. Participants were recruited at the beginning of a menstrual cycle (Pre NEA). Baseline anthropometrics were performed, and diet prescriptions were administered at this visit. During a subsequent menstrual cycle, the NEA cycle, participants consumed a neutral diet for 5 days followed by an ad libitum diet and had an assessment of body composition, resting energy expenditure and hormones. During the first 5 days of a subsequent menstrual cycle, the DEA cycle, participants received a decreased energy diet (−55%) compared to neutral energy diet with unchanged exercise and all previous measurements were reperformed. Abbreviations: NEA, neutral energy availability; DEA, decreased energy availability; LBM, lean body mass; EFP, early follicular phase of a menstrual cycle; GH, growth hormone.

After preliminary screening, participants underwent an exercise test to determine maximum oxygen intake (VO_2max) and indirect calorimetry to determine resting energy expenditure (REE) using open-circuit spirometry and a calibrated metabolic cart (TrueOne 2400 Metabolic Measurement System, Parvo Medics Inc., Sandy, UT). LBM was measured using dual-energy x-ray absorptiometry (DXA; Lunar iDXA, GE Medical Systems, Madison, WI). These procedures were performed at the Department of Exercise and Sport Science, University of North Carolina, Chapel Hill (UNC) during the EFP.

Continuous actigraph monitoring of activity level, energy expenditure, and heart rate were measured using wrist band actigraphy (ActiGraph GT3X-BT wrist band monitor devices, ActiGraph, LLC., Pensacola, FL) beginning before ovulation in the cycle before the NEA intervention continuing to the end of the cycle of the DEA intervention.

For eligible participants, REE and LBM were used to calculate individualized intervention diets and exercise regimens. Interventions were based on energy availability, defined as dietary intake minus energy expenditure. Two energy availability prescriptions were determined for each subject using baseline diet logs, LBM, and VO_2max. NEA was determined as 45 kcal/kg LBM/day and DEA as 20 kcal/kg LBM/day. Change in diet was the sole source of reduced energy availability in DEA as exercise regimens in the form of treadmill, exercise bike, or run/walk were kept constant between NEA and DEA. Individualized exercise regimens were designed to produce a calorie expenditure of ∼150-300 kcal and lasting ≤1 hour per day. Exercise compliance measured by actigraphy was not different either during or before the 2 interventions and did not differ between the 2 interventions.

Energy intake was achieved by study subjects solely consuming standard meal replacements for each intervention (5 days) provided as cookies (Lenny & Larry®), bars (OhYeah®, Quest® and One Bar®) and/or shakes (GNC Total Lean Shakes®), keeping macronutrient concentrations as close as possible and comparable to a general diet. Macronutrient composition for each intervention for NEA vs DEA was as follows: carbohydrate 36.34% vs 29.07%; fat 28.82% vs 32% and protein 35.03% vs 39.27%. Compliance was determined by diet logs and wrapper counts. To avoid a potential carry-over effect of reduced energy availability on the following cycle, the NEA intervention occurred first, followed by the DEA intervention in a subsequent cycle.

Transvaginal or abdominal pelvic ultrasounds that measured the size of the dominant follicle and endometrial thickness (Samsung H60 system, Seoul, Republic of Korea) and urine LH kits (First Response^TM Ovulation Plus Pregnancy Test, Church & Dwight Co. Inc., Ewing, NJ) were performed in the cycle prior to each intervention to determine the day of expected menses and the timing of the diet interventions. On the fifth day of each intervention and always within the EFP, participants reported to the Clinical Research Unit at the National Institute of Environmental Health Sciences (NIEHS) following an overnight fast. DXA was used to assess body composition and weight at enrollment and for diet prescriptions. On day 5 of the NEA and DEA interventions, body composition, BMI, body weight, body fat percentage, LBM and REE were assessed using BOD POD® (air displacement plethysmography).

An intravenous plastic cannula was inserted, and blood was sampled for 8 h starting at ∼0800 h. Fasting and course of day samples were assayed for TSH and growth hormone (GH) collected every 10 min, cortisol every 30 min, total T3 (TT3), reverse T3 (rT3), and total T4 (TT4) every 60 min, free T3 (fT3), and free T4 (fT4) and TBG at the beginning and end of the sampling day. Participants were fed a standardized breakfast and lunch 08:00 and 12:00 h according to their energy availability prescription. A snack was consumed at 15:00 h, equivalent to the participant's NEA intake on both sampling days.

Assays

ELISA was used to measure TSH and GH (Myriad RBM, Austin, TX), and TBG and cortisol were performed by the Department of Laboratory Medicine, Clinical Center, NIH (Simens Immulite 2000 XPi, Malvern, PA). TSH had reference and recovery ranges of 0.55-4.3 mcIU/mL and 96%-109%, respectively with interassay and intra-assay CVs of 5%-13% and 0%-12%.²⁹ GH had reference and recovery ranges of 0-12 ng/mL and 100%-111%, respectively with inter and intra-assay CVs 5%-10% and 0%-8%.²⁹ Cortisol was analyzed using Architect i2000 (Abbott Laboratories®, IL) and had reference and linear ranges of 5-25 mcg/dL and 1-150 mcg/dL, respectively, with interassay and intra-assay CVs 6.4%-13.5% and 4.6%-8.4%.

FT4 and fT3 were analyzed using the SCIEX Triple-Quad-6500 system, as previously reported.^26,30 FT4 had a reference and recovery range of 1.3-2.4 ng/dL and 93%-107%, respectively and interassay and intra-assay CVs of 11%-12% and 5.3%-10.8%, respectively. FT3 had a reference and recovery range of 2.0-6.3 pg/mL and 92%-108%, respectively and interassay and intra-assay CVs of 9.4%-12.6% and 6.9%-11.5%, respectively.

TT4, TT3, and rT3 were analyzed using the Agilent 6460-Triple-Quad system, as previously reported.^30,31 TT3 and TT4 had a recovery range of 65-193 ng/dL and 4.6-12.9 μg/dL, respectively and rT3 had a recovery range of 4.5-30.0 ng/dL. Interassay CVs for TT3, TT4, and rT3 were 5.2%-6.2%, 3.6%-4.8%, and 10.2%-11.9%, respectively. Intra-assay CVs for TT3, TT4, and rT3 were 5.0%-5.2%, 3.4%, and 7.6%-8.7%, respectively.

Fasting estradiol was analyzed using Abbott ARCHITECT® Immunoassays. Interassay and intra-assay CVs for estradiol were 2.1%-7.4% and 1.5%-6.4% (manufacturer reported), respectively, with an analytical sensitivity of ≤10 pg/mL.

Data analysis

Nineteen participants completed both NEA and DEA interventions and actigraph data acquisition. Inspection of the data revealed the expected decline in a number of hormones of interest across the day from the early morning peak. The data for hormone levels were log- or square-root transformed as needed, and differences in across-the-day values were analyzed using linear mixed models with repeated measures to assess the effect of the intervention. Additional comparisons of fasting differences were performed. The mixed models included random intercepts for subjects with an unstructured covariance matrix. Results are presented as least squared (ls) mean ± SEM unless otherwise indicated. P values < .05 were considered statistically significant.

Results

Nineteen healthy women, with regular, ovulatory menstrual cycles and a gynecological age of ≤14 years completed this study (Table 1). Participants had a relatively modest weight loss and decrease in BMI between NEA and DEA interventions without changes in body fat, LBM, or REE (Table 1). All subjects tested negative for thyroid autoimmunity when assessed for the presence of anti-TPO antibodies. E2 was not different between the 2 interventions on the day of sampling (NEA vs DEA: 28.00 vs 31.67 pg/mL, P = .29).

Table 1.

Participant characteristics.

Demographics (n = 19)
Age (y)	23.4 ± 2.1
Age of menarche (y)	12.6 ± 1.0
Gynecological age (y)	10.8 ± 2.2
Race
White, n (%)	8 (42.1)
Black, n (%)	6 (31.6)
Asian, n (%)	2 (10.5)
Ethnicity (n, % Hispanic)	3 (15.8)

Physiologic characteristics	NEA (n = 19)	DEA (n = 19)
Body weight (kg)	60.8 ± 7.1	59.7 ± 6.8a
Body mass index (kg/m2)	23.1 ± 2.3	22.7 ± 2.2b
Body fat (%)	28.6 ± 6.9	28.2 ± 5.8
Lean body mass (kg)	43.4 ± 6.1	43.0 ± 5.6
Resting energy expenditure (kcal/d)	1189.2 ± 155.8	1176.8 ± 142.8

Mean ± SD values (unless otherwise indicated) are reported for paired NEA and DEA cycles for 19 participants. Physiological characteristics were measured on day 5 of each dietary intervention (NEA and DEA).

^a P < .001. ^bP < .01.

Abbreviations: DEA, decreased energy availability; NEA, neutral energy availability.

TSH decreased from NEA to DEA when all values across the day were included in the analysis (NEA vs. DEA, ls mean ± SEM; 1.03 ± 0.07 vs. 0.92 ± 0.07 mcIU/mL, P < .0001), although fasting results did not quite reach significance (1.58 ± 0.11 vs. 1.43 ± 0.10 mcIU/mL, P = .11). The normal circadian decrease in TSH from morning to late afternoon was preserved in both interventions (∼53% decline in both NEA and DEA) (Figure 2).

Figure 2.

TSH values across the day during a neutral or a decreased energy availability diet (NEA, blue; DEA, red). TSH values were higher in NEA compared to DEA. There was a decline in TSH across the day through the late afternoon that was present both during NEA and DEA.

TT4 was not different between NEA and DEA measured in fasting samples (6.54 ± 0.28 vs 6.86 ± 0.28 mcg/dL, P = .10) but was higher in DEA compared with NEA in samples measured across the day (6.06 ± 0.25 vs 6.26 ± 0.25 mcg/dL, P = .005). TT4 declined from morning to late afternoon in both interventions (P < .0001) (Figure 3A). fT4 increased from NEA to DEA (1.71 ± 0.07 vs. 1.83 ± 0.07 ng/dL, P = .0052), driven by differences in fasting samples (1.74 ± 0.07 vs. 1.87 ± 0.08 ng/dL, P = .03) but was not different between the 2 interventions at the end of the sampling period (1.71 ± 0.08 vs. 1.80 = 0.08 ng/dL, P = .13) (Figure 3B). However, fT4 did not differ significantly between the beginning and end of the sampling period across both interventions (P = .313).

Figure 3.

(A) TT4 values across the day during a neutral or a decreased energy availability diet (NEA, blue; DEA, red). TT4 was higher during DEA compared to NEA and followed the usual pattern of decline across the day in both conditions. (B) fT4 was higher during DEA overall with the change being significant at fasting but not at the end of the sampling period.

TT3 declined from NEA to DEA at fasting (110.33 ± 3.51 vs. 101.22 ± 3.51 ng/dL; P = .006) and in across-the-day samples (95.55 ± 2.89 vs. 89.15 ± 2.89 ng/dL; P < .0001). Moreover, TT3 declined across the day (P < .0001) in both conditions (Figure 4A). Fasting fT3 was somewhat lower after DEA but did not reach statistical significance (4.65 ± 0.42 vs 3.90 ± 0.44 pg/mL, P = .08). FT3 declined across the day during NEA (P < .0001) but not DEA (P = .11), and samples drawn at the end of the sampling period were not different between the 2 interventions (3.34 ± 0.42 vs. 3.23 ± 0.44 pg/mL; P = .77) (Figure 4B). This may indicate a blunting of the usual changes related to the time of day in fT3 after caloric restriction.

Figure 4.

(A) T3 values across the day during a neutral or a decreased energy availability diet (NEA, blue; DEA, red). T3 values decreased from NEA to DEA and followed the usual pattern of decline across the day during both interventions. (B). Free T3 values were not different between NEA and DEA overall. (C). Reverse T3 was higher at DEA compared to NEA but did not exhibit a change across the day during either intervention.

Reverse T3 was higher with DEA than NEA in fasting samples (9.90 ± 0.86 vs 12.40 ± 0.87 ng/dL; P = .02) and in samples drawn across the day (9.02 ± 0.56 vs. 12.04 ± 0.56 ng/dL; P < .001) (Figure 4C). Unlike the other thyroid hormones, rT3 did not change across the day.

Finally, cortisol and GH followed the usual diurnal and postprandial patterns, but neither cortisol (P = .69) nor GH (P = .21) was different between NEA and DEA.

In summary, 5 days of moderate caloric deficiency was associated with lower TSH levels, higher TT4 and fT4, lower TT3 and an increase in rT3. These changes are consistent with energy conservation through the overall effect of lower levels of the most biologically active T3 on metabolic rate.

Discussion

In the present study, we investigated the impact of short-term energy restriction in young women and found that thyroid hormone dynamics adapt to even modest reductions in energy availability to maintain homeostasis. Specifically, the decline in TSH and T3 and the increase in rT3 are consistent with a reduction in energy expenditure in the face of reduced energy intake. These thyroid hormone changes occurred without changes in thyroid binding globulin, cortisol and GH, or major changes in body composition and REE.

Previous diet studies used older immunoassays focused on more extreme caloric restriction, primarily in men. Energy restriction in these studies ranged from total caloric restriction for 2.5 to 8 days,³² 83% reduction for 14 days,²⁴ and 70% for 7-18 days.²⁵ In these studies, TT4 did not change,^24,25,33 TT3 increased in some^15,24 and decreased in others^25,32,33 while rT3 was unchanged,^15,25 decreased,²⁴ or increased.^32,33 TBG was reduced in one study²⁴ with concomitant changes in free hormones but was unchanged in a second study.³³ In the current study, moderate dietary restriction resulted in decreased TSH, TT3, and morning fT3 with an increase in TT4, fT4, and rT3, supporting a causative role of such changes in the pathogenesis of FHA.^34-36 Differences from prior studies may stem from varying energy restriction levels, sampling methods, or population demographics.^32,33 Additionally, using LCMS assays, which are more sensitive in detecting changes in T3, T4, and rT3, may explain discrepancies with previous studies that relied on older immunoassays.³⁰

Our results implicate alterations in peripheral deiodinase activities (D1, D2, and D3), resulting in the decreased extrathyroidal conversion of T4 to T3, leading to reduced T3 and increased rT3 levels as a means of conserving cellular energy expenditure³⁷ in the face of even moderate reduction in energy availability. Morning TSH was lower with dietary restriction and declined across the day. However, TSH was not different between NEA and DEA at its nadir in the late afternoon, changes that are reflected in fasting and late afternoon fT3 and fT4. These results suggest that reduced energy availability affects nighttime energy conservation more than daytime. Further studies are needed to determine if daytime meals mitigate the impact of reduced caloric intake on TSH and thyroid hormones or if a true circadian difference exists. TT3 and TT4, with longer half-lives due to TBG binding, remain different across the entire day.

In relatively small studies of total^38,39 or moderate⁴⁰ caloric restriction in obese and overweight subjects lasting weeks to months, TSH decreased,³⁹ T4 was unchanged,^38,40 T3 declined,^38,40 and rT3 increased.³⁸ One study showed sex differences in thyroid hormone adaptation to diet, with TSH declining in women but not men.⁴⁰ In obese women, weight loss over 4 months reduced 24-h TSH secretion.⁴¹ Similarly, a cross-sectional study found lower TSH and fT3 in girls with anorexia nervosa and higher levels in obese girls, indicating a relationship between body weight and thyroidal differences.⁴²

There are even fewer longer-term studies of energy restriction and thyroid function. In 2 studies in which participants lost weight by restricting energy through decreased intake or increased exercise for months to years, decreased dietary intake but not increased exercise resulted in a significant reduction in T3 levels.^23,43 Neither group had changes in fT4, TSH,^23,43 or rT3.²³ These findings suggest that with chronic energy restriction, the balance between T4 and T3 is driven by decreased energy intake rather than changes in body composition.

The current study induced deficient energy availability by reducing caloric intake while maintaining exercise. TT3, fT3, and TSH decreased, aligning with rodent data showing T3, TRH mRNA, and TSH reductions under food restriction, unlike exercise, which only affected TRH mRNA and TSH.⁴⁴ Increases in cortisol and GH have been described in studies of more severe caloric restriction,^39,45 and both influence thyroid hormone dynamics.⁴⁶ Neither was affected by dietary manipulation, likely due to modest energy restriction. We closely monitored diet and exercise compliance, controlled for gonadal steroids, and used precise assays for blood hormone measurement.

Previous studies suggest LCMS-based thyroid assays outperform immunoassays. van Deventer et al. (2011) found LCMS-FT4 correlated better with TSH than immunoassay-FT4 (ρ: 0.84 vs. 0.45, P < .0001).²⁶ A study of 50 critically ill patients confirmed LCMS assays' superior sensitivity in detecting TT3, rT3, and TT4 changes, as they are less affected by altered binding globulins and hormonal fluctuations.³⁰ Although our participants denied taking over-the-counter supplements, it is also of note that biotin can interfere with thyroid hormone immunoassays, although this is not the case for LCMS assays⁴⁷ and that biotin levels may be elevated above the range shown to interfere with thyroid hormone immunoassays even in individuals who deny taking supplements.⁴⁸ Taken together, these data provide evidence for the use of LCMS assays in the current study and may account for some of the differences observed when compared with previous studies.

Thyroid hormone triiodothyronine (T3), and prohormone thyroxine (T4) are intricately involved in maintenance of energy expenditure and metabolic rate.^49,50 T3, and to a lesser extent T4, act on nearly all tissues through thyroid hormone receptors, TRα and TRβ.⁵¹ T3 also promotes energy expenditure by increasing expression of uncoupling proteins leading to increased mitochondrial density,⁵² orchestrating an elaborate process of adaptive thermogenesis.⁵³ It is widely hypothesized that decreased T3 and increased rT3 during energy restriction lower metabolic rate to preserve energy for vital functions. In this study, plethysmography assessed body composition and calculated REE on day 5 of each diet. Short-term energy restriction does not consistently alter REE via indirect calorimetry. Some fasting studies show increased REE due to a carbohydrate-to-fat metabolism shift, while others show no change.^54,55 Thus, the unchanged REE in our subjects may reflect altered substrate metabolism linked to reduced energy availability.⁵⁶ It is possible that we did not find changes in REE with energy restriction because of the use of calculated REE (during NEA and DEA visits) rather by indirect calorimetry. However, previous similar reports make our results more likely to reflect a true lack of difference in REE.⁵⁷

Strengths of this study include carefully selecting participants to exclude those adapted to dietary restriction or strenuous exercise. Diet and exercise were tailored to individual needs, with exercise designed for 150-300 kcal/day expenditure. Compliance was closely monitored through dietary tracking and actigraphy. Other strengths include the use of (LCMS) for thyroid hormone measurement, and standardizing sample collection to the EFP with measurement of gonadal hormones on the days of testing.⁵⁸ Furthermore, we assessed cortisol and GH values during both interventions due to their potential effect on thyroid physiology. For logistical reasons, we assessed body composition and REE during both interventions using BODPOD®, which may have been less sensitive in detecting changes compared with indirect calorimetry and DXA.

Collectively, our findings suggest that even short-term, moderate energy restriction in women without obesity triggers thyroid hormone adaptations, reducing TSH secretion and altering deiodinase activities within the thyroid and peripherally. These changes decrease biologically active T3 and increase inactive rT3, reflecting a physiological response to conserve energy and maintain cellular fuel availability.

Funding

ZID ES102465 and ZID ES103323 of the Intramural Research Program of the National Institute of Environmental Health Sciences, National Institutes of Health.

Authors’ contributions

Skand Shekhar (Conceptualization [supporting], Data curation [lead], Formal analysis [lead], Investigation [lead], Methodology [lead], Project administration [supporting], Resources [supporting], Validation [lead], Visualization [lead], Writing—original draft [lead]), Joselyne Tessa Tonleu (Formal analysis [supporting], Project administration [equal], Resources [supporting], Writing—review & editing [lead]), Chinelo C. Okigbo (Methodology [supporting], Project administration [supporting], Validation [supporting], Writing—review & editing [supporting]), Helen Leka (Investigation [supporting], Project administration [supporting], Resources [supporting], Validation [supporting], Writing—review & editing [supporting]), Anne E. Kim (Investigation [supporting], Methodology [supporting], Project administration [supporting], Validation [supporting], Writing—review & editing [supporting]), Bona P. Purse (Data curation [supporting], Investigation [supporting], Project administration [lead], Resources [supporting], Validation [supporting], Writing—review & editing [supporting]), Katie R. Hirsch (Data curation [equal], Formal analysis [equal], Investigation [equal], Project administration [equal], Validation [equal], Writing—review & editing [equal]), Brian R. Stolze (Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Project administration [equal], Resources [equal], Validation [equal], Writing—review & editing [equal]), John A. McGrath (Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Project administration [equal], Software [equal], Validation [equal], Visualization [equal], Writing—review & editing [equal]), Abbie E. Smith-Ryan (Conceptualization [supporting], Investigation [supporting], Project administration [supporting], Resources [supporting], Validation [equal], Writing—review & editing [equal]), Steven J. Soldin (Formal analysis [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal], Supervision [equal], Validation [equal], Visualization [equal], Writing—review & editing [equal]), and Janet E. Hall (Conceptualization [lead], Data curation [lead], Funding acquisition [lead], Investigation [equal], Methodology [lead], Project administration [supporting], Resources [lead], Software [lead], Supervision [lead], Validation [lead], Visualization [lead], Writing—review & editing [lead])

Data availability

Data are available upon reasonable request from the corresponding author.