Sensitivity to thyroid hormones is associated with sleep duration in the euthyroid population with depression degree lower than moderate

We collected thyroid-related hormone index levels, sleep duration, and other basic characteristics of the population with depression from the NHANES 2009–2012 cycles and evaluated the association of Thyroid-Stimulating Hormone Index (TSHI) with sleep duration in the euthyroid population with depression via different analysis methods. We found that the association between TSHI and sleep duration was only found in patients with depression degree < Moderate (score: 1–14) rather than > Moderate group. Among the populations with degree < Moderate (N = 1918), only 4 indexes (parametric Thyroid Feedback Quantile Index, PTFQI, Thyrotroph Thyroxine Resistance Index, TT4RI, Thyroid-Stimulating Hormone TSH, and TSHI) reflecting the sensitivity to thyroid hormones were related to the sleep duration, with a significant non-linear relationship after adjusting for potential confounders (all P < 0.05). Trend analysis indicated that with the level increase of these 4 indexes, the sleep duration increased (all P for trend < 0.001). Further, we found that TSHI was relatively more important among the 4 indexes. Sum up, sensitivity to thyroid hormones is associated with sleep duration in the euthyroid population with depression degree lower than Moderate. Poor sensitivity referred to a longer sleep duration.

FT3 levels in the adult the US population with sleep duration less than 7 h 14 .Kessler et al. indicated the decrease in serum TSH and free T4 during human sleep restriction 15 .Those studies only focus on the relationship between thyroid-related hormones and sleep, but they did not distinct the thyroid function (normal or abnormal).
In addition, it should be noted that FT3 or TSH alone may not be sufficient to reflect the regulation of thyroid hormone homeostasis 16 .Therefore, it is necessary to explore more thyroid hormone indices to comprehensively explain thyroid status.Recent years, the sensitivity to thyroid hormones have been widely paid more attention.Sensitivity to thyroid hormones refers to the body's response to these hormones, which can be influenced by various factors, including sleep durations and disturbances 17 .The sensitivity to thyroid hormones indicators includes Thyrotroph Thyroxine Resistance Index (TT4RI), Thyroid-Stimulating Hormone Index (TSHI), Thyroid Feedback Quantile-Based Index (TFQI), parametric Thyroid Feedback Quantile Index (PTFQI) and FT3 to FT4 ratio (FT3/FT4) 18 .The first three are quantitative markers for pituitary thyrotropic function 19 .FT3/FT4 for estimating the conversion efficiency of FT4 to FT3 can be used to indirectly reflect the peripheral sensitivity of thyroid hormones.At present, few studies reported the correlation of sensitivity to thyroid hormones with sleep or depression, but there is no distinction between people with euthyroid populations and people with thyroid function disorder.Furthermore, the potential interaction between sleep duration and sensitivity to thyroid hormones has not been extensively explored in the literature on the euthyroid population with depression grade.
To enrich such research, we explored the association of the sensitivity to thyroid hormones with sleep duration in euthyroid populations with depression in this study.

Data source and study participants
The data used in this research were collected from the National Health and Nutrition Examination Survey (NHANES), a publicly available cross-sectional, multistage survey database, we synthesized the 2009-2010 and 2011-2012 continuous cycles data (https:// www.cdc.gov/ nchs/ nhanes/ index.htm).A total of 20,293 participants were obtained for the next screening.Then 13,184 participants whose thyroid laboratory data was incomplete, 560 participants who had dysthyreosis, 2512 participants who lacked the covariate including marital status, smoking, alcohol use, and race; and 1974 participants who did not answer the PHQ9 questionnaire were excluded.Finally, the 2063 participants were included in this study.Flow of the inclusion and exclusion of participants was showed in Fig. 1.

Laboratory methodology and computing formula for testing thyroid function
The thyroid blood specimens was obtained form the participants who attended to in blood collection of NHANES.Thyroid blood specimens were processed, stored, and shipped to the University of Washington, Seattle, WA.Detailed specimen collection and processing instructions are discussed in the NHANES Laboratory/Medical Technologists Procedures Manual (LPM).The Access HYPERsensitive human TSH assay is a 3 rd generation, two-site immunoenzymatic ("sandwich") assay.The total T3, free T3, and total T4 assay were competitive binding immunoenzymatic assays.The free T4 assay is a two-step enzyme immunoassay.The access thyroglobulin (Tg) assay is a simultaneous one-step "sandwich" assay.The thyroglobulin antibody and thyroid peroxidase antibody (TPOAb) assay is a sequential two-step immunoenzymatic "sandwich" assay.The detailed steps of the experiment are displayed on the official website (https:// wwwn.cdc.gov/ Nchs/Nhanes/ 2009-2010/ THYROD_ F. htm) TSHI, TT4RI, TFQI, PTFQI, andT3/T4were calculated using the following formula, respectively.CDF: Cumulative distribution function.The euthyroid individuals were screened via normality reference ranges 7.74-20.64pmol/L for FT4 and 0.34-5.60mIU/L for TSH 20 .

Definitions of depression grade and sleep duration
The depression grades were defined via the answer score of the patient health questionnaire 9 (PHQ9).The PHQ9 is validated as a depressive symptom severity measure.Total score 1-4 was defined as minimal depression, 5-9 was defined as mild depression, 10-14 was defined as moderate depression, 15-19 was defined as moderately severe depression, and 20-27 was defined as severe depression 21 .
Sleep duration was obtained from the questionnaire as the follows: "Number of hours usually sleep on weekdays or workdays".Sleep duration was categorized as < 5 h, 5-8 h, > 8 h referring to reported study [22][23][24] .

Covariates
The sex, race, marital status, educational level, smoking, alcohol use, hypertension, diabetes, and stroke were collected from in-person interviews and considered as covariates in this study.The race included Mexican American, non-Hispanic White, non-Hispanic Black, other Hispanic, and other Race-Including Multi-Racial.The marital status included never marital, married /living with partner, widowed/divorced/separated.The education level included less than high school, high school or equivalent, and high school above.Smoker was defined as adults who smoked > 100 cigarettes in life and smoked some days or every day.Smokers who do not currently smoke

Statistical analyses
All the statistical analyses used the appropriate sample weights, 1/2 of two-year subsample weights according to the NHANES analysis guidelines via 4.2.2 version R software.The chi-square test was used to test for the association between categorical variables and one-way ANOVA for the association between categorical and continuous variables.The generalized additive model (GAM) and polynomial regression were used to analyze the sleep duration and depression grade.The generalized linear analysis was used to explore the relationship between the thyroid function index and sleep duration in different depression grades.Then, we found there was a relationship between sleep duration and thyroid function index in the minimal depression, mild depression, and moderate depression groups.Hence, the GAM analyses were performed to further explore the relationship between thyroid indexes and sleep duration.The generalized linear analysis was performed to evaluate the relationship between the thyroid hormone indexes with sleep duration.Model 1 was unadjusted.Model 2 was adjusted for race, education level, and BMI.Model 3 was adjusted for race, education level, BMI smoking, alcohol use, diabetes, and hypertension.Next, the trend test between the thyroid hormone indexes and sleep duration www.nature.com/scientificreports/ was analyzed.In addition, the four methods including Xgboost, AdaBoost, Random Forest, and light GAM were performed to evaluate the importance of thyroid hormone index.

Ethics approval
The Ethics Committee of the 904th Hospital of Joint Logistic Support Force of PLA deemed that this research is based on open-source data, so the need for ethics approval was waived.

Determination of study participants
We first explored the sleep duration of populations with depression.The GAM and Polynomial regression analyses were performed to explore the correlation between sleep duration and depression grade.The results showed that there was a non-linear relationship with an inverted U-shape (Fig. 2A, P < 0.001, β = − 0.336 [− 0.417, − 0.256]).In addition, we found a key inflection point (depression grade = Moderate).The sleep duration decreased with the depression grade increasing when the depression grade lower than Moderate.When the depression grade larger than Moderate, the sleep duration increased with the depression grade increasing.Consistent results are obtained by polynomial regression analysis (Fig. 2B).Hence, all participants were grouped into two clusters (lower than Moderate and larger than Moderate).Next, the generalized linear model was performed to explore the relationship between the thyroid function indexes and sleep duration in two groups.The results showed that among all indicators, TSH, PTFQI, TT4RI, and TSHI correlated with the sleep duration in participants with depression grade less than Moderate (Table 1, all P < 0.001).However, no correlation was observed between sleep duration and any of indicators in participants with depression grade larger than Moderate.Therefore, the participants with depression grade less than Moderate were selected to further explore.
The trend test was also performed to verify the relationship degree between hormone indexes and sleep duration (Table 4).The results showed that the association between 4 indicators and sleep duration gradually increased with the increase of their levels (all crude P for trend < 0.05), and their association remained significant even after adjusting for potential confounders including race, education level, BMI smoking, alcohol use, diabetes, and hypertension (all adjusted P for trend < 0.05).www.nature.com/scientificreports/

Verifying the importance of thyroid hormone indices affecting sleep duration
The above results have demonstrated the significant association between 4 indices and sleep duration.Finally, we deeply explored their importance on the depression score via different methods including Xgboost, AdaBoost, Random Forest, and light Gradient Boosting Machine (GBM).We found that the importance of TSHI ranked first in Xgboost and light GBM (Fig. 4A,D).TSHI was second only to age in AdaBoost (Fig. 4B).In the Random forest, TSHI ranked the three (Fig. 4C).Taken together, the importance degree of TSHI was relatively higher among 4 indices associated with sensitivity to thyroid hormones.

Discussion
In patients with depression, severe sleep problems are often manifested.Southmayd et al. indicated that major depression patients were able to alleviate sleep deprivation by receiving thyroxine 27 .The study revealed the influence of thyroid hormones on sleep in patients with severe depression.However, those studies were limited to the relationship between sleep duration and thyroid function, and did not reveal the association of thyroidrelated hormones with sleep duration in low-grade depression populations.Currently, there is limited research revealing the correlation between thyroid-related hormones and sleep duration in this specific group, which warrants further investigation.
In this study, we aimed to explore the association between thyroid-related hormone indexes and sleep duration in populations with depression.We found that there was a positive correlation between thyroid-related hormone indexes and sleep duration just in populations with depression degrees lower than Moderate.No significant association was observed in the subgroup with a greater than moderate degree.It followed that the findings from these studies may not be applicable for people experiencing depression with different degrees.More importantly, we found that only indicators reflecting the sensitivity to thyroid hormones (PTFQI, TSH, TT4RI, and TSHI) but not common T3 and T4, were significantly related to the sleep duration in this subgroup populations, suggesting that thyroid hormone sensitivity indexes may be more suitable for evaluating the sleep duration in depression, compared with these common thyroid-related hormone indexes.These results further highlighted the importance of sensitivity to thyroid hormones.To our knowledge, this is the first study to explore the association between the sensitivity to thyroid hormones and sleep duration in particular population.
Further, we sufficiently justified the association between these 4 indicators and sleep duration in the subgroup populations.With the increase of these 4 indicators, the sleep duration increased.These results suggested the negative association between longer sleep duration and poor sensitivity to thyroid hormones.Especially, TSHI was relatively more important among these indicators, which reflect the sensitivity of FT4 to pituitary 16,28 .These results highlighted the importance of thyroid hormone homeostasis on the sleep.
The importance of thyroid-related hormone on the sleep has been widely reported.Wang et al. indicated that there was a negative relationship between FT3 levels and sleep duration 14 .The T3, T4, and TSH levels were positively correlated with the severity of insomnia symptoms 29 .Song et al. 30 have shown that decreased thyroid hormone levels can lead to difficulty in falling asleep and reduced sleep duration.Tian et al. showed that increased T4 levels can stimulate symptoms similar to restless legs syndrome, leading to insomnia and reduced sleep duration 31 .The potential mechanism of thyroid hormone involved in the sleep needs to be revealed.It is well known that thyroid function is closely related to depression, and sleep disorder is a common complication in patients with depression.The hypothalamic-pituitary-adrenal (HPA) and hypothalamic-pituitary-thyroid (HPT) axis changes in patients with depression, both of which are closely related to sleep.The HPT axis is involved in the occurrence and development of sleep.For example, patients with thyroid dysfunction usually will www.nature.com/scientificreports/have insomnia, hypersomnia, etc. Intravenous injection of thyrotropin-releasing hormone (TRH) can change sleep parameters 28,32 .The activity of the HPA axis is related to the excessive awakening of insomnia 33,34 .Studies have shown that in patients with depression, the hippocampus, the negative feedback center of the HPA axis, was impaired in the negative feedback regulation of glucocorticoids through the HPA axis, resulting in increased cortisol.Then, glucocorticoids activated TRH neurons, resulting in increased TRH secretion, thereby down-regulating TRH receptors on thyrotrophs.Therefore, in depressive patients with normal thyroid function, thyroid function has changed, including increased T4 levels, decreased TSH response to TRH, and no increase in nocturnal TSH 35 .Conversely, sleep duration can also affect the levels of thyroid hormones.Kim et al. confirmed that long sleep duration and short sleep duration were risk factors for increased and decreased TSH, respectively, compared to optimal sleep duration, based on data from KNHANES in Korea 36 .The reason is that the rhythm of TSH secretion Table 2. Baseline characteristics of populations with depression grades lower than moderate.PTFQI parametric thyroid feedback quantile index, TT4RI thyrotroph thyroxine resistance index, TSHI thyroidstimulating hormone index, TFQI thyroid feedback quantile-based index, TSH thyroid-stimulating hormone, T4 thyroxine.may be related to melatonin which is the main hormone secreted by pineal stromal cells.There is a large number of melatonin receptors expressed in the anterior pituitary, and the change in light duration will affect the effect of melatonin on the anterior pituitary.
The study had some limitations.First, it is possible that the result was not an exact causal relationship due to cross-sectional analysis.Continuity research should be carried out in the future.
Second, we did not explore the level of adrenaline and other hormones.Third, the effect of drugs on thyroid hormones was not ruled out due to the lack of medication information.

Conclusion
To sum up, sensitivity to thyroid hormones was significantly related to the sleep duration in the euthyroid populations with depression degree lower than moderate.Among the subgroup populations, the poor sensitivity to thyroid hormones correlated with longer sleep duration.This study highlighted the importance of thyroid hormone homeostasis on the sleep in depression (Supplementary Information S1).

Figure 1 .
Figure 1.Flow charts of the inclusion and exclusion of participants.

Figure 2 .
Figure 2. The relationship between the depression grade and sleep duration.(A) Generalized additive model (GAM).(B) Polynomial regression.

Figure 4 .
Figure 4.The importance ranking bar chart of the top five factors via four methods.(A) Xgboost (B) AdaBoost (C) Random forest (D) light GBM.

Table 1 .
The association between the thyroid function index and sleep duration in different depression grade group by generalized linear model.PTFQI parametric thyroid feedback quantile index, TT4RI thyrotroph thyroxine resistance index, TSHI thyroid-stimulating hormone index, TFQI thyroid feedback quantile-based index, TSH thyroid-stimulating hormone, T4 thyroxine, T3 triiodothyronine, T3/T4 free T3 to free T4 ratio.

Table 4 .
The trend test between the thyroid hormone indexes and sleep duration (n = 1918).Crude model: no variables adjusted.Adjusted model: adjusted the race, education level, BMI smoking, alcohol use, diabetes, hypertension.TSHI thyroid-stimulating hormone index, TT4RI thyrotroph thyroxine resistance index, PTFQI parametric thyroid feedback quantile index, TSH thyroid-stimulating hormone.