Associations among thyroid hormone levels and mean corpuscular volume in adults in the US: A cross-sectional examination of the NHANES 2007–2012 dataset

Abstract

Mean corpuscular volume (MCV) is an important indicator used to determine the etiology of anemia and is associated with a variety of diseases. However, the link between thyroid function and MCV has yet to be clarified. This study was thus developed to assess relationships between thyroid function and MCV in a population of adults in the US. Results from the National Health and Nutrition Examination Survey study performed from 2007 to 2012 were used to conduct a cross-sectional analysis. Key thyroid-related variables included in this analysis were thyroid-stimulating hormone, total thyroxine (TT4), free triiodothyronine (FT3), total triiodothyronine (TT3), free thyroxine (FT4), antithyroglobulin, thyroglobulin, and antithyroid peroxidase levels. Generalized linear regression models were employed when estimating associations between MCV quartiles and thyroid parameters in 8104 adults 18 + years of age. In these participants, the weighted mean (SD) MCV was 89.36 (0.16) fL, with thyroid-stimulating hormone levels of 1.86 (0.03) mIU/mL, FT3 levels of 3.20 (0.01) pg/mL, FT4 levels of 0.80 (0.01) ng/dL, TT3 levels of 115.09 (0.64) ng/dL, and TT4 levels of 7.81 (0.04) μg/dL. When analyses were not adjusted, higher MCV values were related to reduced serum levels of FT3, TT3, or TT4. Following adjustment for possible confounding variables, this significant negative correlation between MCV and levels of FT3, TT3, and TT4 remained, and subgroup analysis revealed that this negative correlation was present in the male group and in the age group >50 years, but not in the female group and in the age group less than or equal to 50 years. These results suggest a significant negative correlation between MCV and FT3, TT3, and TT4, and this negative correlation originated more from the male population and those older than 50 years of age. The underlying mechanisms warrant additional investigation.

1. Introduction

Mean corpuscular volume (MCV) is a measurement of the average volume of red blood cells, and is a crucial routine complete blood count parameter. It is helpful for identifying the root cause of anemia, which is categorized as microcytic, orthocytic, or macrocytic anemia depending on the MCV value. Moreover, the red blood cell distribution width (RDW) can be calculated using MCV.^[1] However, the clinical utility of MCV is also believed to extend beyond the etiological differential diagnosis of anemia. It has become clear through a study evaluating the clinical significance of elevated MCV values that these values are frequently linked to alcohol consumption, liver illness, cancer, and the side effects of chemotherapy.^[2] MCV has recently been recognized as a predictive marker for ischemic stroke,^[3] colorectal or esophageal malignancy,^[4,5] chronic renal disease,^[6] or coronary intervention.^[7]

Thyroid hormones including 3,3′,5-l-triiodothyronine (T3) and l-thyroxine (T4) regulate virtually all cells in the body with respect to metabolism, differentiation, and proliferation.^[8,9] Thyroid-stimulating hormone (TSH) controls the production of these hormones by triggering a negative feedback loop in the hypothalamic–pituitary axis.^[10] The intranuclear thyroid receptor (TR), which heterodimerizes with the retinoid X receptor (RXR), is the principal target of thyroid hormone action. The thyroid response element (TRE) found in the DNA associated with certain genes can specifically interact with this heterodimer to control gene expression.^[11]

Hematopoietic stem cells express thyroid hormone receptors, and serum thyroid hormone levels can modulate the extent of blood cell production.^[12] However, the relationship between MCV and thyroid hormones remains unclear. Few studies to date have examined the association between MCV and thyroid function, with further analysis being critical to understanding whether and how MCV and thyroid function are related. This study was thus performed with the goal of investigating the link between MCV and blood thyroid parameters levels (TSH, free triiodothyronine [FT3], free thyroxine [FT4], total triiodothyronine [TT3], total thyroxine [TT4], thyroid peroxidase antibody [TPOAb], thyroglobulin antibody [TgAb], and thyroglobulin [Tg]) in a large population of adults with no history of thyroid disease.

2. Materials and methods

2.1. Study population

Relationships among thyroid hormone levels and MCV were examined with data from the ongoing National Health and Nutrition Examination Survey (NHANES) that monitors the health and nutritional status of the US population every 2 years. This survey makes use of a sophisticated multistage cluster design and incorporates an array of demographics, clinical interview, physical exam, and laboratory component sections. The NHANES study received Institutional Review Board (IRB) of the National Center for Health Statistics approval prior to data collection, and signed informed consent was gathered from all study participants.

The present analysis was performed using data from the 2-year survey periods (2007–2008, 2009–2010, and 2011–2012) for which thyroid hormone status-related data were available. This study included participants aged 18 years or older, with no clinical thyroid problems (excluding participants who answered yes to “Have you ever been told you had a thyroid problem/disease?” or taking the thyroxine drugs), were not pregnant, and had data on MCV and thyroid hormone levels (Fig. 1).

Figure 1.

Study flowchart. MCV = mean corpuscular volume, NHANES = National Health and Nutrition Examination Survey.

2.2. MCV measurement

The MCV (fL) is measured with a Coulter analyzer at Mobile Test Centers (MECs). This approach enables the accurate counting of cells and their sizes through the detection and measurement of changes in particle resistance within conductive liquids as they pass through small pore sizes. MCV represents the average volume of a single red blood cell (RBC) derived from a histogram of RBC results. The number of RBCs per channel was multiplied by the size of RBCs in that channel, with the product of each channel being added between 36 and 360 fL, followed by the division of this sum by the total number of RBCs from 36 to 360 fL. This was then multiplied by the calibration constant. The normal range for MCV values is 80 to 100 fL.^[1] In this study, MCV was assessed as a continuous and categorical variable using weighted quartiles (Q1 = 56.3–86.6 fL, Q2 = 86.7–89.7 fL, Q3 = 89.8–92.6 fL, Q4 = 92.7–118.1 fL).

2.3. Thyroid outcomes

The NHANES Laboratory/Medical Technologists Procedures Manual contains comprehensive instructions on specimen collection and processing (LPM). Analyzed thyroid parameters included FT3, FT4, TSH, TT3, TT4, Tg, TgAb, and TPOAb. Immunoenzymatic assays were used to detect all thyroid-related biomarker levels. The TSH reference range was 0.34 to 5.6 IU/mL. The reference range for the FT3 assay was 2.5 to 3.9 pg/mL for a competitive binding immunoenzymatic assay, while for FT4 the reference range was 0.6 to 1.6 ng/dL. TT3 and TT4 were also detected using competitive binding immunoenzymatic assays, with respective 80 to 220 ng/dL and 5.0 to 12.0 μg/dL reference ranges. A sequential 2-step immunoenzymatic sandwich assay was used to calculate the titers of TPOAb and TgAb, with respective reference ranges of 0 to 9.0 IU/mL and 0 to 4.0 IU/mL. The Access Tg assay is a simultaneous 1-step sandwich assay. We chose to exclude participants who have thyroid issues or are pregnant to lessen the impact of these factors on thyroid function.

2.4. Demographic characteristics

Participant age, ethnicity, education levels, poverty-to-income ratio (PIR), smoking history, gender, and history of alcohol intake were assessed with a standard questionnaire. Body mass index (BMI), serum creatinine (Cre), urinary iodine concentration (UIC), blood urea nitrogen (BUN), serum iron (Fe), and serum folate (SF) levels were measured by trained technicians in MECs.

The ethnicity of participants was separated into 4 categories (Non-Hispanic Black, Non-Hispanic White, Mexican American, or other), while education was separated into 5 categories (9th grade, 9th–11th grade [as well as 12th grade without graduation], graduated high school/GED, some college/AA degree, and college or greater). Participant socioeconomic status was assessed based on PIR values. BMI values were calculated by dividing participant weight (kg) by height squared (m²). Smoking status was classified as never (<100 lifetime cigarettes), former (>100 lifetime cigarettes but not a current smoker), or current. Alcohol intake history was also categorized as never (<12 lifetime drinks), former (≥12 drinks in 1 year but no drinks in the last year or ≥12 lifetime drinks but no drinks in the last year), mild (≤2 and ≤1 drinks per day for males and females, respectively, on average for the past 12 months) moderate (3 or 2 drinks per day for males and females, respectively, on average for the past 12 months), or heavy (≥4 or ≥3 drinks per day for males and females, respectively, on average for the past 12 months).^[13]

The laboratory indicators for this study included BUN (mg/dL), serum Fe (μg/dL), SF (nmol/L), UIC (μg/L), and Cre (mg/dL), and these were analyzed via certified technicians with standardized laboratory methods.

2.5. Statistical analyses

Analyses were performed by adapting the complex survey sampling approach for these NHANES data. The primary sampling unit in this study corresponds to the unit of variance (the sampling unit used to estimate sampling error). Individuals were assigned sampling and subsample weights as appropriate to adjust for the odds of nonequal selection, nonresponders, and independent population controls. Sample weight values were established through the rescaling of weight values such that the sums of these weights were consistent with the survey population at the midpoint for every 2-year study interval. In the present analysis, these sample weights were established for the combined 6-year survey dataset. R v 4.2.2 was used to conduct all analyses with the “nhanesR” package (v 0.9.4.1) and the “survey” package (v 4.1-1).

Data are reported as means with standard deviation (SD) values or medians with interquartile ranges (IQRs) or percentages, as appropriate. Relationships independent of distributional assumptions were examined by assessing associations between MCV quartiles and thyroid hormone levels. Nonadjusted and adjusted generalized linear regression models were constructed when exploring these relationships between MCV and thyroid hormone levels. Age, gender, race, BMI, alcohol intake, smoking status, SF, UIC, serum Cre, serum Fe, and BUN levels were used to adjust regression models.

3. Results

3.1. Baseline participant characteristics

In total, 8104 participants from the NHANES 2007 to 2012 dataset were incorporated into this study (Fig. 1). Of these participants, 4320 (weighted percentage: 53.31%) were males and 3784 (weighted percentage: 46.69%) were females, 3558 (weighted percentage: 43.9%) were Non-Hispanic White, 1678 (weighted percentage: 20.71%) were Non-Hispanic Black, and 1349 (weighted percentage: 16.62%) were Mexican American. The weighted mean (SD) age of participants was 44.99 (0.45) years. Individuals with elevated MCV values were more likely to be older, male, White, have a lower BMI, be smokers, and be alcohol users. Serum Fe, SF, and BUN levels rose with increasing MCV quartiles (see Table S1, Supplemental Content, http://links.lww.com/MD/L838 which shows baseline characteristics of participants according to MCV quartiles).

As shown in Table 1, the median MCV value was 89.7 fL, while the lower (Q1) and upper (Q3) quartile values were 86.6 and 92.6 fL, respectively, and a mean value of 89.36 (0.16) fL. The mean TSH levels of participants was 1.86 (0.03) mIU/mL, with serum free T3 levels of 3.20 (0.01) pg/mL, serum free T4 levels of 0.80 (0.01) ng/dL, serum total T3 levels of 115.09 (0.64) ng/dL, and serum total T4 levels of 7.81 (0.04) μg/dL. Study participants were free of clinical thyroid problems and not pregnant. Significant negative correlations were observed for FT3, TT3, and TT4 levels across MCV quartiles. However, a linear relationship was not observed when assessing the association between MCV quartiles and the levels of TSH, FT4, TPOAb, TgAb, and Tg.

Table 1.

Thyroid function of NHANES (2007–2012) study population in MCV quartiles.

Characteristics	MCV quartiles					P value
	Total	Q1	Q2	Q3	Q4
N	8104	2439	2005	1861	1799
MCV (fL)	89.7 (56.3–118.1)	56.3–86.6	86.7–89.7	89.8–92.6	92.7–118.1
MCV (fL)	89.36 (0.16)a	82.79 (0.17)	88.25 (0.03)	91.14 (0.03)	95.45 (0.10)	<.001
TSH (mIU/L)	1.86 (0.03)	1.85 (0.04)	1.88 (0.05)	1.86 (0.05)	1.85 (0.05)	.958
FT3 (pg/mL)	3.20 (0.01)	3.26 (0.01)	3.24 (0.02)	3.19 (0.02)	3.12 (0.02)	<.001
FT4 (ng/dL)	0.80 (0.01)	0.80 (0.01)	0.80 (0.01)	0.80 (0.01)	0.79 (0.00)	.723
TT3 (ng/dL)	115.09 (0.64)	118.99 (1.05)	117.14 (0.87)	113.37 (0.70)	110.75 (1.18)	<.001
TT4 (μg/dL)	7.81 (0.04)	8.03 (0.06)	7.92 (0.06)	7.68 (0.06)	7.59 (0.07)	<.001
TPOAb (IU/mL)	16.16 (1.52)	13.48 (1.41)	17.03 (2.42)	15.12 (2.57)	19.14 (4.17)	.368
TgAb (IU/mL)	5.75 (0.73)	5.71 (1.09)	6.52 (1.44)	5.85 (1.59)	4.88 (0.77)	.693
Tg (ng/mL)	15.13 (0.47)	16.04 (0.57)	14.89 (1.33)	13.93 (0.80)	15.66 (0.91)	.092

FT3 = free triiodothyronine, FT4 = free thyroxine, MCV = mean corpuscular volume, SD = standard deviation, Tg = thyroglobulin, TPOAb = thyroid peroxidase antibody, TSH = thyroid-stimulating hormone, TT3 = total T3, TT4 = total T4.

^aContinuous variables: weighted mean (SD).

3.2. Association between thyroid function and MCV

As demonstrated in Table 2 and Figure 2, MCV was negatively correlated with FT3 (β = −0.007, 95% CI: −0.010 to −0.004, P < .001), TT3 (β = −0.485, 95% CI: −0.662 to −0.307, P < .001), and TT4 (β = −0.031, 95% CI: −0.041 to −0.020, P < .001) when using an unadjusted model, and this remained true under Model 2 (FT3: β = −0.003, 95% CI: −0.006 to −0.0002, P = .035; TT3: β = −0.295, 95% CI: −0.458 to −0.132, P < .001; TT4: β = −0.025, 95% CI: −0.036 to −0.015, P < .001) and Model 3 (FT3: β = −0.005, 95% CI: −0.009 to −0.002, P = .01; TT3: β = −0.361, 95% CI: −0.588 to −0.135, P = .004; TT4: β = −0.029, 95% CI: −0.042 to −0.015, P < .001) following adjustment for confounding factors. Following MCV categorization into quartiles, this negative association was still present (FT3, P for trend = .030; TT3, P for trend < .001; TT4, P for trend = .004). Participants in the highest MCV quartile exhibited 0.081 pg/mL lower FT3, 5.632 ng/dL lower TT3, and 0.380 μg/dL lower TT4 levels relative to those individuals in the lowest MCV quartile.

Table 2.

The association between MCV (fL) and thyroid function.

	Model 1a β (95% CI)	P value	Model 2b β (95% CI)	P value	Model 3c β (95% CI)	P value
TSH (mIU/L)
Total MCV	0.002 (−0.005, 0.009)	.539	−0.009 (−0.015, −0.002)	.011	−0.001 (−0.007, 0.009)	.767
MCV (fL) categories
Q1	Reference		Reference		Reference
Q2	0.027 (−0.082, 0.137)	.616	−0.053 (−0.157, 0.051)	.311	−0.006 (−0.143, 0.132)	.929
Q3	0.012 (−0.113, 0.138)	.846	−0.095 (−0.216, 0.027)	.123	−0.026 (−0.167, 0.116)	.698
Q4	−0.003 (−0.130, 0.124)	.958	−0.155 (−0.279, −0.032)	.015	−0.006 (−0.161, 0.149)	.930
P for trend		.805		.182		.763
FT3 (pg/mL)
Total MCV	−0.007 (−0.010, −0.004)	<0.001	−0.003 (−0.006, −0.0002)	.035	−0.005 (−0.009, −0.002)	.01
MCV (fL) categories
Q1	Reference		Reference		Reference
Q2	−0.014 (−0.054, 0.026)	.481	−0.015 (−0.049, 0.019)	.375	−0.022 (−0.070, 0.026)	.331
Q3	−0.065 (−0.100, −0.031)	<.001	−0.039 (−0.071, −0.007)	.018	−0.075 (−0.119, −0.030)	.004
Q4	−0.131 (−0.177, −0.085)	<.001	−0.066 (−0.109, −0.0240)	.003	−0.081 (−0.126, −0.036)	.002
P for trend		<.001		.033		.030
FT4 (ng/dL)
Total MCV	−0.0003 (−0.0011, 0.0005)	.509	−0.0003 (−0.0011, 0.0006)	.540	−0.001 (−0.002, 0.001)	.387
MCV (fL) categories
Q1	Reference		Reference		Reference
Q2	−0.001 (−0.011, 0.008)	.806	−0.002 (−0.012, 0.008)	.708	0.001 (−0.013, 0.015)	.865
Q3	0.001 (−0.012, 0.013)	.887	0.001 (−0.012, 0.014)	.872	−0.001 (−0.015, 0.014)	.935
Q4	−0.007 (−0.020, 0.007)	.325	−0.007 (−0.022, 0.007)	.318	−0.010 (−0.036, 0.016)	.407
P for trend		.987		.919		.543
TT3 (ng/dL)
Total MCV	−0.485 (−0.662, −0.307)	<.001	−0.295 (−0.458, −0.132)	<.001	−0.361 (−0.588, −0.135)	.004
MCV (fL) categories
Q1	Reference		Reference		Reference
Q2	−1.857 (−4.419, 0.704)	.151	−1.448 (−3.866, 0.970)	.233	−1.317 (−4.504, 1.870)	.383
Q3	−5.626 (−7.495, −3.758)	<.001	−4.328 (−6.057, −2.600)	<.001	−5.894 (−8.079, −3.709)	<.001
Q4	−8.247 (−11.081, −5.412)	<.001	−5.206 (−7.850, −2.563)	<.001	−5.632 (−9.272, −1.992)	.006
P for trend		<.001		<.001		<.001
TT4 (μg/dL)
Total MCV	−0.030 (−0.041, −0.020)	<.001	−0.025 (−0.036, −0.015)	<.001	−0.029 (−0.042, −0.015)	<.001
MCV (fL) categories
Q1	Reference		Reference		Reference
Q2	−0.106 (−0.266, 0.054)	.190	−0.047 (−0.201, 0.108)	.545	−0.006 (−0.212, 0.200)	.947
Q3	−0.345 (−0.481, −0.209)	<.001	−0.295 (−0.431, −0.159)	<.001	−0.381 (−0.576, −0.186)	.001
Q4	−0.439 (−0.613, −0.265)	<0.001	−0.394 (−0.567, −0.221)	<.001	−0.380 (−0.596, −0.164)	.003
P for trend		<.001		<.001		.004

FT3 = free triiodothyronine, FT4 = free thyroxine, MCV = mean corpuscular volume, TSH = thyroid−stimulating hormone, TT3 = total T3, TT4 = total T4.

^aModel 1: no covariates were adjusted.

^bModel 2: age, gender, and race/ethnicity were adjusted.

^cModel 3: age, gender, race/ethnicity, education, poverty-to-income ratio, body mass index, urinary iodine concentration, serum folate, serum iron, creatinine, blood urea nitrogen, alcohol use, smoke was adjusted.

Figure 2.

The association between MCV and thyroid function. Solid line represents the linear fit between MCV and FT3 (A), as well as TT3 (B) and TT4 (C). Gray range represent the 95% of confidence interval from the fit. Age, gender, race/ethnicity, education, poverty-to-income ratio, body mass index, urinary iodine concentration, serum folate, serum iron, creatinine, blood urea nitrogen, alcohol use, and smoke was adjusted. FT3 = free triiodothyronine; TT3 = total T3, TT4 = total T4, MCV = mean corpuscular volume.

3.3. Subgroup analyses

As shown in Table 3, in further subgroup analysis, a negative correlation between FT3, TT3, TT4, and quartiles of MCV was found in the male group and in the age group >50 years (P for trend < .05), but in the female group and in the age group ≤50 years, there was no statistically significant difference in the trend of decreasing FT3, TT3, and TT4 with increasing MCV (P for trend > .05). In addition, in the White group, there was a negative correlation between TT4 and the quartiles of MCV (P for trend < .05), but no statistical difference in FT3 and TT3, and no statistical difference in the correlation between FT3, TT3, TT4, and quartiles of MCV in the Black and Mexican groups.

Table 3.

The results of subgroup analyses.

Model 3a	Q1	Q2	Q3	Q4	P for trend
FT3
Sex
Male	Reference	−0.05 (−0.100, 0.001)	−0.09 (−0.154, −0.025)	−0.092 (−0.141, −0.042)	.021
Female	Reference	0.006 (−0.097, 0.108)	−0.057 (−0.124, 0.009)	−0.063 (−0.149, 0.023)	.436
Age
>50	Reference	−0.03 (−0.091, 0.031)	−0.071 (−0.129, −0.012)	−0.08 (−0.138, −0.021)	.019
≤50	Reference	−0.03 (−0.094, 0.034)	−0.097 (−0.150, −0.044)	−0.134 (−0.187, −0.080)	.924
Eth
White	Reference	−0.033 (−0.097, 0.030)	−0.08 (−0.146, −0.014)	−0.076 (−0.135, −0.018)	.701
Black	Reference	−0.007 (−0.096, 0.081)	−0.014 (−0.075, 0.047)	−0.029 (−0.106, 0.048)	.398
Mexican	Reference	0.167 (−0.275, 0.609)	−0.051 (−0.152, 0.050)	−0.147 (−0.299, 0.005)	.216
Other	Reference	−0.102 (−0.231, 0.027)	−0.095 (−0.277, 0.088)	−0.163 (−0.314, −0.012)	.028
TT3
Sex
Male	Reference	0.051 (−3.373, 3.474)	−4.947 (−8.841, −1.053)	−5.219 (−9.377, −1.061)	<.001
Female	Reference	−3.153 (−9.518, 3.212)	−7.141 (−11.023, −3.258)	−5.781 (−10.705, −0.857)	.798
Age
>50	Reference	−0.131 (−4.093, 3.831)	−3.892 (−8.020, 0.236)	−5.792 (−11.125, −0.459)	.01
≤50	Reference	−2.381 (−6.148, 1.386)	−7.714 (−10.411, −5.017)	−7.16 (−11.210, −3.109)	.111
Eth
White	Reference	−0.652 (−4.087, 2.783)	−6.566 (−9.868, −3.263)	−5.47 (−10.526, −0.415)	.158
Black	Reference	0.26 (−5.533, 6.054)	−0.413 (−7.501, 6.676)	−3.88 (−9.658, 1.898)	.23
Mexican	Reference	−1.308 (−5.192, 2.576)	−3.592 (−7.492, 0.308)	−1.028 (−6.279, 4.223)	.35
Other	Reference	−8.249 (−19.013, 2.516)	−8.714 (−21.957, 4.530)	−10.413 (−21.273, 0.447)	.047
TT4
Sex
Male	Reference	0.011 (−0.175, 0.197)	−0.363 (−0.594, −0.132)	−0.353 (−0.586, −0.120)	<.001
Female	Reference	−0.043 (−0.385, 0.299)	−0.429 (−0.699, −0.159)	−0.463 (−0.794, −0.132)	.319
Age
>50	Reference	0.065 (−0.370, 0.501)	−0.33 (−0.660, 0.001)	−0.464 (−0.782, −0.145)	.022
≤50	Reference	−0.032 (−0.218, 0.155)	−0.421 (−0.623, −0.219)	−0.346 (−0.600, −0.092)	.131
Eth
White	Reference	0.045 (−0.188, 0.278)	−0.418 (−0.658, −0.179)	−0.46 (−0.706, −0.214)	.021
Black	Reference	0.028 (−0.342, 0.399)	−0.192 (−0.664, 0.279)	−0.214 (−0.610, 0.181)	.242
Mexican	Reference	−0.16 (−0.446, 0.126)	−0.446 (−0.787, −0.105)	−0.35 (−0.773, 0.073)	.015
Other	Reference	−0.338 (−0.816, 0.140)	−0.365 (−0.963, 0.234)	−0.008 (−0.745, 0.730)	.754

The significance of bold is the P-value of < 0.05 for the relationship between the corresponding thyroid parameter and MCV in this subgroup of the population.

FT3 = free triiodothyronine, MCV = mean corpuscular volume, TT3 = total T3, TT4 = total T4.

^aModel 3: age, gender, race/ethnicity, education, poverty-to-income ratio, body mass index, urinary iodine concentration, serum folate, serum iron, creatinine, blood urea nitrogen, alcohol use, smoke was adjusted.

4. Discussion

The present analysis of a large cohort of representative community-dwelling adults in the US that were not pregnant and had no history of thyroid problems revealed a linear association between MCV values and serum levels of FT3, TT3, and TT4. Specifically, higher MCV values were linked to reductions in FT3, TT3, and TT4 values. No independent association was detected between MCV values and levels of TSH, FT4, TPOAb, TgAb, or Tg. Therefore, elevated MCV levels are more likely to be associated with lower thyroid hormone levels, but are not associated with thyroid autoimmune diseases.

MCV is used clinically as a laboratory marker for the differential diagnosis of anemia.^[1] Thyroid hormones can affect erythropoiesis, promoting an increase in basal cellular metabolism and a corresponding increase in tissue oxygen consumption, leading to an increase in the synthesis and release of erythropoietin (EPO).^[14] Thyroid hormones can also stimulate iron incorporation into erythrocytes^[15] and increase iron uptake.^[16] These processes are likely the most important mechanisms through which thyroid hormones affect erythropoiesis.^[17,18] Therefore, in patients with thyroid dysfunction, anemia is often observed.^[19–21] A study by Horton et al^[22] showed that the prevalence of anemia in hypothyroid patients was surprisingly high at 47%. Some reports suggest that thyroid dysfunction may be associated with some forms of anemia, but the incidence of this association varies widely in adults.^[23,24] Earlier studies have demonstrated a substantial relationship between thyroid function hemoglobin, hematocrit, and RBC counts.^[17,25] However, investigations focused on the relationship between MCV and thyroid function are currently scarce and have yielded inconsistent results, even though the above studies support an underlying link between MCV and thyroid status.

Some types of anemia can be brought on by hypothyroidism, but it can also drive the excessive accumulation of immature progenitor cells. In cases of an elevated MCV and moderately severe hypothyroidism, anemia is typically macrocytic hypochromic and/or normocytic in etiology.^[26] The link between anemia and hypothyroidism has been explained as a physiological adaptation to tissue oxygen requirements. This notion is supported by the low plasma erythropoietin levels seen in cases of hypothyroid anemia and the overall increase in erythrocyte size in patients with simple primary hypothyroidism after thyroid surgery. These data align well with the present results, which revealed that the values of FT3, TT3, and TT4 decreased with increasing MCV values, suggesting that MCV is relatively at higher levels when thyroid hormones are at lower levels. Elevated MCV may be associated with rapidly progressive hypothyroidism. When treated with thyroxine replacement therapy, a progressive decrease in MCV was found, even when the initial values were within the normal range.^[24] Although some current studies have shown changes in MCV findings in patients with thyroid dysfunction,^[27,28] the exact mechanism remains unclear. Davidson et al^[29] hypothesized that it could be either an alteration in the number of mitotic divisions during erythrocyte maturation–proliferation or an alteration in the lipid composition of the erythrocyte membrane, and they examined the latter, but their study did not demonstrate that an alteration in the structure of the erythrocyte membrane lipids to explain the changes in MCV in patients with thyroid dysfunction.

Thyroid function is influenced by many factors, particularly sex, age, and race.^[30–32] In our subgroup analysis of these factors, our results showed that FT3, TT3, and TT4 were negatively correlated with MCV quartiles in the male population and in those older than 50 years, but this negative correlation was not significant in the female population and in those younger than 50 years. Thyroid dysfunction is known to be more common in older patients and in women,^[33] and in a meta-analysis of thyroid-associated traits, Porcu et al^[34] identified novel loci and sex differences in the regulation of thyroid function, suggesting that some of the loci were genome-wide significant only in women or men. With increasing age, the elderly population had higher and increased TSH levels and slightly lower T4 levels.^[35,36]

This study employed a large representative adult cohort from the US, but these findings are nonetheless subject to certain limitations. While our study observed a correlation between MCV and thyroid function during the time of testing, it is important to note that the lack of longitudinal follow-up and the cross-sectional design of the study limit our ability to establish a causal relationship. Furthermore, relying on a single time point for measuring thyroid parameters in cross-sectional data may not offer a comprehensive understanding of participants’ thyroid status over an extended period. This approach may not capture any potential changes in participants’ thyroid function over time.

5. Conclusion

In summary, the present results confirmed the strong link between MCV and thyroid function in a large US adult participant cohort. Specifically, serum FT3, TT3, and TT4 levels were independently associated with MCV, whereas serum TSH, FT4, TPOAb, TgAb, and Tg levels were not, and this association originated mainly in the male population and in people older than 50 years. This study did not examine potential mechanistic factors underlying this relationship, underscoring the need for further studies focused on elucidating the biological mechanisms underlying the effects of thyroid hormones on MCV.

Acknowledgments

We would like to thank everyone who participated in NHANES and the NHANES team for their time and work. Also, we thanks to Zhang Jing (Shanghai Tongren Hospital) for his work on the NHANES database. His outstanding work, the nhanesR package and webpage, makes it easier for us to explore the NHANES database.

Author contributions

Conceptualization: Mingzheng Wang, Junru Liu.

Data curation: Mingzheng Wang, Xiaofeng Lu.

Funding acquisition: Mingzheng Wang, Junru Liu.

Investigation: Xiaogang Zheng.

Methodology: Xiaotao Zhu.

Project administration: Mingzheng Wang, Junru Liu.

Software: Xiaofeng Lu.

Validation: Mingzheng Wang, Xiaotao Zhu.

Writing – original draft: Mingzheng Wang, Junru Liu.

Writing – review & editing: Mingzheng Wang, Xiaofeng Lu, Xiaogang Zheng, Xiaotao Zhu, Junru Liu.