https://orcid.org/0009-0001-2842-3940
Figures
Abstract
Background
To compare the levels of allostatic load score (ALS) between thyroid cancer patients and patients with other types of cancer and explore whether ALS mediates the association between thyroid cancer and alterations in physiological function.
Methods
This cross-sectional study conducted a secondary analysis of 181 cancer patients using NHANES data from 2007 to 2020, including 91 individuals with thyroid cancer. Generalized linear regression, logistic regression, and sensitivity analysis were used to analyze the association between thyroid cancer and ALS in three different models. Receiver operating characteristic (ROC) curve analysis and feature importance analysis were utilized to assess the clinical predictive value of thyroid cancer. We also conducted a series of mediation analyses to examine the mediating role of ALS.
Results
The ALS in thyroid cancer patients was higher than that in other cancer types (P < 0.05). Thyroid cancer was significantly associated with ALS even after adjusting for demographic variables (β = 0.770, 95%CI: 0.315–1.480; OR=2.255, 95%CI: 1.111–4.575). This association remained robust to missing data (all P < 0.05) and was not confounded by drinking, diabetes, or thyroid disease (all P > 0.05). Although thyroid cancer had limited predictive value on ALS, it exerted strong explanatory power. The mediation analysis conducted with imputed data and adjusted for confounding variables revealed that ALS fully mediated the effect of thyroid cancer on red cell distribution width (RDW) (IE: β = 0.103, P = 0.008; DE: β = 0.389, P = 0.056), with a mediation proportion of 20.93%.
Citation: Wang W-r, Liu G-d, Ren D-x, Liu X-y (2026) Allostatic load in thyroid cancer is higher than that of other cancers: A secondary analysis using NHANES. PLoS One 21(1): e0341063. https://doi.org/10.1371/journal.pone.0341063
Editor: Diego A. Bonilla, Dynamical Business & Science Society - DBSS International SAS, COLOMBIA
Received: May 26, 2025; Accepted: December 30, 2025; Published: January 22, 2026
Copyright: © 2026 Wang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Allostatic load (AL), a concept first proposed by McEwen and Stellar in 1993 [1], emphasizes the cumulative physiological burden of multiple systems resulting from prolonged adaptation to stressors. Allostatic load score (ALS), developed from key biomarkers, quantifies the effects of stress on the cardiovascular, metabolic, and immune systems [2,3]. Studies have highlighted that an increased ALS was associated with cardiovascular disease, cognitive function, and increased all-cause mortality [4–7], and ALS outperformed individual biomarkers in stress assessment and the associated biological burden [8]. High AL exerts cumulative wear and tear on multiple physiological systems (e.g., metabolic and cardiovascular systems) [4]. This impact manifests as dysregulation of physiological functions, such as impaired glucose metabolism and dysregulated inflammatory response [8]. Identifying individuals’ AL levels facilitates the implementation of targeted health intervention strategies.
Over the past several decades, the incidence of thyroid cancer has increased more rapidly than that of any other malignancy [9,10]. This epidemiological trend has resulted in a marked rise in the number of cancer survivors, given the disease’s 98.5% five-year survival rate [11]. Projections indicate that over one million individuals will be living with a thyroid cancer diagnosis within the next decade [12]. Based on its prognostic characteristics, thyroid cancer may impose less physiological strain in response to external stressors compared to other cancers. However, previous studies indicated that thyroid cancer survivors reported significantly lower quality of life than breast cancer survivors, while exhibiting similar outcomes to colon cancer survivors [13,14]. Moreover, thyroid cancer was associated with higher bankruptcy rates compared to other cancer types [15,16]. These findings starkly contradict the prevailing perception that thyroid cancer is a generally mild form of cancer. Consequently, compared to patients with other types of cancer, individuals with thyroid cancer may exhibit elevated ALS. As previously discussed, high ALS has been shown to exert detrimental effects on multiple physiological systems. Nevertheless, it remains to be determined whether the observed alterations in physiological system functions among thyroid cancer patients are mediated by ALS.
Therefore, we conducted a large-scale cross-sectional study to compare the levels of ALS between thyroid cancer patients and patients with other types of cancer and explore whether ALS mediates the association between thyroid cancer and alterations in physiological function. We hypothesized that compared with patients with other cancers, patients with thyroid cancer are associated with elevated ALS, and ALS may mediate the association between thyroid cancer and alterations in physiological function.
2. Methods
2.1. Data source and study participants
Secondary data were obtained from the National Health and Nutrition Examination Survey (NHANES), conducted by the U.S. Centers for Disease Control and Prevention (CDC), which is a nationally representative cross-sectional survey assessing health and nutritional status in the U.S. population. It combines interviews, physical examinations, and laboratory tests to collect data on demographics, diet, chronic diseases, biomarkers, and environmental exposures. All protocols used were approved by the Health Statistics Ethics Review Board of the U.S. CDC. All participants signed written informed consent forms, allowing anonymous data to be used for scientific research. This study followed the STandardisierte BerichtsROutine für Sekundärdaten Analysen (STROSA), which provide standardized guidelines for reporting secondary data analysis.
The most recent seven cycles of NHANES data available from 2007 to 2020 were included in this analysis. 61,648 participants were involved. Of these participants, we selected 3,746 individuals diagnosed with tumors of identified types. We employed propensity score matching (PSM) at a ratio of 1:3 to select patients with thyroid cancer and other cancers. Matching was performed using covariates including age, gender, race, education level, marital status, and poverty income ratio (PIR). After PSM, 325 individuals remained, including 88 samples with thyroid cancer and 237 samples with other types of cancer. Eventually, 181 participants with full ALS data and tumor data were enrolled in the study (Fig 1).
2.2. Definition of allostatic load score
AL was determined as an aggregate measure reflecting multisystem dysregulation, encompassing biomarkers related to cardiovascular, metabolic, and immune system [17]. Specifically, for the cardiovascular domain, systolic and diastolic blood pressure (SBP and DBP, mmHg), as well as total serum cholesterol (TC, mmol/L) and high-density lipoprotein cholesterol (HDL_C, mg/dL) levels, were incorporated. In terms of the metabolic system, glycohemoglobin (HbA1c, g/dL), albumin concentrations (g/dL), and body mass index (BMI, kg/m2) were included. For the immune system, C-reactive protein (CRP, mg/dL) levels were incorporated [18]. The eight biomarkers have been widely used in recent NHANES studies for the assessment of AL [3,18–20].
Based on prior research, ALS was calculated by converting each biomarker into a binary variable according to its statistical distribution within our dataset (using quartiles). If a biomarker fell within the highest quartile, it was assigned 1; otherwise, it received 0 (i.e., the lowest three quartiles). Exceptions were made for albumin and HDL_C, where lower values indicated higher risk. For these two variables, being in the lowest quartile resulted in 1 point being assigned. The possible ALS score ranged from 0 to 8, with a higher score indicating a stronger association of stress with physiologic dysregulation [3].
2.3. Assessment of cancer
Data on the cancer diagnosis and types were gathered from self-reported cancer assessments. Individuals who answered “Yes” to the question “Have you ever been told by a doctor or other health professional that you had cancer or a malignancy of any kind?” and “Thyroid” to the question “What kind of cancer was it?” were defined as thyroid cancer patients. Those who answered “Yes” to the first question but answered other types of cancer to the second were defined as patients with other types of cancer. In NHANES, the health questionnaire, physical examination, and biological sample collection were all completed on the same day at the Mobile Examination Center (MEC). Therefore, we can infer that the cancer diagnosis time of the study subjects must have occurred before the biomarker measurement. Some cycles of the NHANES dataset provide the age at the first diagnosis of thyroid cancer. We have listed the age information of the first cancer diagnosis for 9 thyroid cancer patients in this study from the 2007–2008 cycle in S1 Table for reference. The specific cancer types of the 181 patients enrolled in the analysis are presented in S1 Fig.
2.4. Covariates
This study incorporated covariates that may potentially influence the relationship between thyroid cancer and ALS, guided by previous research and clinical expertise: age, gender (male, female), race (Mexican American, Others), marital status (married, others), education level (under high school, high school or equivalent, above high school), PIR, smoking (yes, no), drinking (never, former, current), central obesity (yes, no), obesity (yes, no), hypertension (yes, no), diabetes (yes, no), thyroid disease (yes, no), liver disease (yes, no), hyperuricemia (yes, no), arthritis (yes, no), hypoalbuminemia (yes, no), gout (yes, no), menopause (yes, no), mean corpuscular hemoglobin concentration (MCHC, g/dL), and red cell distribution width (RDW, %).
Smoking status was classified as no (cotinine ≤15 ng/mL) or yes (cotinine >15 ng/mL). Drinking status was categorized as never (<12 lifetime drinks), former (≥12 lifetime drinks but no drinks in the last year), or current (drinks in the last 12 months) [21]. Obesity was defined as BMI ≥ 29.9 kg/m2. Hypertension was defined as a self-reported prior diagnosis, the use of antihypertensive medication, or a blood pressure reading of ≥140/90 mmHg. Diabetes was defined as a self-reported prior diagnosis, a fasting plasma glucose level of ≥7 mmol/L, a 2-h post-load plasma glucose level of ≥11.1 mmol/L, or the use of insulin or oral hypoglycemic medications. Hyperuricemia was defined as uric acid >7.0 mg/dL (male) or >6.0 mg/dL (female). Hypoalbuminemia was defined as albumin ≤35 d/L. Both MCHC and RDW are measured and calculated during the complete blood count (CBC) using the Beckman Coulter MAXM instrument. Serum cotinine is analyzed using isotope dilution-high-performance liquid chromatography coupled with atmospheric pressure chemical ionization tandem mass spectrometry (ID HPLC-APCI MS/MS). Fasting glucose and 2-hour post-load plasma glucose were determined via the hexokinase method. The serum uric acid concentration was quantified using the timed-endpoint method. Blood pressure levels were calculated as the mean of three consecutive readings obtained using a calibrated mercury sphygmomanometer, following a minimum five-minute rest period in a quiet room with a controlled ambient temperature. HbA1c was analyzed using a glycated hemoglobin analyzer. TC and HDL_C were assessed using a Roche Modular P chemistry analyzer. Albumin was measured via the bichromatic digital endpoint method. CRP was quantified through latex-enhanced nephelometry. BMI was calculated based on anthropometric measurements.
2.5. Statistical analysis
Continuous variables were normally distributed and expressed as mean and standard deviation (SD), while those with a skewed distribution were expressed as median and interquartile range (IQR). Categorical variables were expressed as number and percentage (%). Differences between groups were evaluated using the T-test or Mann-Whitney U-test for continuous variables and the Chi-square test or Fisher’s exact test for categorical variables. All statistical analyses were performed using SPSS version 26.0 and R version 4.2.1, and P values <0.05 (two-sided) were considered statistically significant.
Three different models were established to explore the association between thyroid cancer and ALS: the crude model (without adjusting for any covariates), Model 1 (adjusted for the PSM factors), and Model 2 (adjusted for the significant variables at baseline). Generalized linear regression was used to analyze the association between thyroid cancer and ALS in three different models. Based on its quartile distribution, ALS was dichotomized into a binary variable, with participants in the highest quartile classified as the high ALS group and those in the lower three quartiles categorized as the non-high ALS group. Logistic regression examines the relationship between one or several independent variables and a binary dependent variable [22]. Therefore, the relationship between thyroid cancer and binary variable ALS was investigated by three distinct logistic regression models. Diagnostic checks were performed to verify the adherence of all regression models to their respective statistical assumptions. In addition, we performed several sensitivity analyses to further validate the robustness of our results. Specifically, we examined whether the relationship between thyroid cancer and ALS was affected by missing data using two imputation methods (random forest and interpolation). Additionally, subgroup analyses were conducted to investigate whether other factors (drinking, diabetes, thyroid disease) might modify the association between thyroid cancer and ALS. The false discovery rate (FDR) correction for multiple comparison was performed.
Moreover, the clinical predictive value of thyroid cancer was assessed using receiver operating characteristic (ROC) curve analysis. Additionally, we employed logistic regression, XG Boost models, and SHapley Additive exPlanations (SHAP) values to evaluate the feature importance of thyroid cancer on ALS. SHAP enhances the interpretability of tree-based models by using a game-theoretic method to aggregate local feature contributions and explain global model behavior. This approach outperforms other global approximation methods and provides insights into both overall feature importance and specific prediction roles [23]. To further examine the mediating role of ALS, we performed a series of mediation analyses.
3. Results
3.1. Descriptive results
This study included 50 patients with thyroid cancer and 131 patients with other types of cancer, with a median age of 58 years and a female proportion of 80.11% (Table 1). Compared with survivors of other cancers, thyroid cancer survivors were more likely to have diabetes and thyroid disease (all P < 0.05). We also found a lower proportion of current drinkers in thyroid cancer participants (P < 0.05). Moreover, thyroid cancer patients exhibited higher levels of RDW, hs-CRP, HbA1c, and ALS, along with lower levels of MCHC (all P < 0.05). There was no significant difference in age, gender, race, education, marital status, PIR, BMI, smoking, central obesity, obesity, hypertension, liver disease, hyperuricemia, arthritis, hypoalbuminemia, gout, menopause, albumin, hemoglobin, CRP, HDL_C, and TC (all P > 0.05).
3.2. Association between thyroid cancer and ALS compared with other cancers
We employed generalized linear regression models to examine the association between thyroid cancer and ALS. We found that thyroid cancer was significantly associated with ALS even after adjusting for various covariates (Table 2). The unadjusted model (Crude model) yielded a β coefficient of 0.770 (95% confidence interval [CI]: 0.199–1.342), which remained significant in the adjusted models, with β coefficients of 0.898 (95% CI: 0.315–1.480) in Model 1 (adjusted for age, gender, race, education, marital status and PIR) and 0.881 (95% CI: 0.044–1.718) in Model 2 (adjusted for MCHC, RDW, drinking, diabetes and thyroid disease).
After dichotomizing ALS using the highest quartile as the cutoff point, logistic regression models were established to investigate the association between thyroid cancer and ALS (Table 2). In the crude model, individuals with thyroid cancer exhibited a 2.255-fold increased risk of elevated ALS (odds ratio [OR]: 2.255, 95%confidence interval [CI]: 1.111–4.575). After adjusting for demographic confounders (Model 1), the association strengthened with an OR of 2.600 (95% CI: 1.245–5.431). Even the association did not exist in Model 2, with an OR of 3.414 (95%CI: 0.948–12.289), the P value was close to the threshold of 0.05, which suggested that the insignificant association may be caused by the relatively small sample size.
All assumptions for the linear regression models, including linearity, normality, and homoscedasticity, were met upon diagnostic testing. Similarly, the assumptions for the logistic regression models—including the absence of multicollinearity, linearity of independent variables and the log odds, and the lack of strongly influential outliers—were also satisfied. Detailed diagnostic plots and result of multicollinearity analysis are provided in the Supplementary Materials (S2-S3 Fig, S2 Table).
3.3. Sensitivity analysis and subgroup analysis
Random forest imputation captures complex patterns and handles non-linear relationships robustly, while interpolation offers simplicity and computational efficiency for structured data trends [24]. Thus, we employed these two imputation techniques to impute the missing data of 8 ALS-related variables in the remaining 325 individuals after PSM, to investigate whether the association between thyroid cancer and ALS was affected by missing data. The associations after data imputation remain consistent with the original results (Table 3). In the random forest approach, the crude model showed a significant positive association between thyroid cancer and ALS (β = 0.409, 95%CI: 0.014–0.804), which strengthened in Model 1 (β = 0.507, 95%CI: 0.106–0.909) and further increased in Model 2 (β = 0.690, 95%CI: 0.076–1.304). A similar pattern emerged using interpolation imputation, with borderline significance in the crude model (β = 0.375, P = 0.056) transitioning to statistically significant estimates in Model 1 (β = 0.486, P = 0.015) and Model 2 (β = 0.627, P = 0.038). Imputation of missing data increases the likelihood of non-significant findings for the crude model (P = 0.056).
Moreover, subgroup analysis was utilized to examine whether the association between thyroid cancer and ALS was modified by drinking, diabetes, and thyroid diseases (as these 3 indicators showed difference at baseline). As shown in Table 4, we found that the association between thyroid cancer and ALS was significant among current drinkers (OR=3.06, 95%CI: 1.21–7.69) and subgroups with comorbid diabetes (OR=6.00, 95%CI: 1.46–24.73) and thyroid diseases (OR=6.45, 95%CI: 1.34–30.95). Notably, this significant association remained statistically significant after FDR correction. There was no interaction effect among these three variables and thyroid cancer on ALS (all P for interaction >0.05), highlighting the stability of their association.
3.4. ROC curve analysis and Importance analysis
To explore the performance of thyroid cancer in predicting ALS levels, variables with significant differences in the baseline table were included. ROC curve analysis revealed that RDW demonstrated the highest area under the curve (AUC) 0.689, followed by MCHC (AUC: 0.617), and thyroid cancer (AUC: 0.596). Diabetes (AUC: 0.566), thyroid disorder (AUC: 0.555), and drinking (AUC: 0.550) showed relatively lower predictive values (Fig 2A).
(A) ROC plot; (B) Feature importance analysis by logistic regression model; (C) Feature importance analysis by XG Boost model; (D) SHAP feature importance plot; (E) SHAP summary plot. Abbreviations: MCHC, mean corpuscular hemoglobin concentration; RDW, red cell distribution width.
Feature importance analysis was conducted using different models. In the logistic regression model, thyroid cancer ranked first, followed by diabetes, RDW, MCHC, thyroid diseases, and drinking (Fig 2B). In the XG Boost model, thyroid cancer also ranked first, followed by RDW, MCHC, drinking, diabetes, and thyroid diseases (Fig 2C). Feature importance analysis based on SHAP values revealed that thyroid cancer ranked first, followed by RDW, diabetes, drinking, MCHC, and thyroid disease (Fig 2D, 2E). These results suggested that thyroid cancer exerted strong explanatory power but has limited discriminatory capacity on ALS.
3.5. Mediation effects of ALS on the relationship of thyroid cancer with MCHC, RDW, and diabetes
The main analysis revealed that patients with thyroid cancer exhibited higher ALS compared to individuals with other types of cancer. Given that elevated ALS has been shown to impair the functioning of multiple physiological systems, we further investigated whether the relationship between thyroid cancer and alterations in physiological functions is mediated by ALS. We selected variables that demonstrated significant intergroup differences in the baseline analysis and were pathophysiologically associated with various physiological systems—including MCHC, RDW, and diabetes—as biomarkers to assess changes in physiological function. Fig 3A and 3C showed thyroid cancer was associated with MCHC and diabetes (P for total effect [TE] <0.05), and ALS acted as a full mediator (indirect effect [IE]: β = 0.049, P = 0.016; direct effect [DE]: β = 0.102, P = 0.133) in thyroid cancer-diabetes pathway, accounting for 32.45% of the mediation effect. We observed that ALS exhibited a significant mediating effect in certain pathways, despite the total effect being non-significant (Fig 3B).
(A) MCHC; (B) RDW; (C) Diabetes; (D-F) Imputed data analyses (adjusted for age, gender, race, education, marital status, and PIR): (D) MCHC; (E) RDW; (F) Diabetes. Abbreviations: ALS, allostatic load score; MCHC, mean corpuscular hemoglobin concentration; RDW, red cell distribution width; PIR, poverty income ratio.
To enhance the robustness of our findings, we re-conducted the mediation analysis by increasing the sample size, utilizing imputed data, and adjusting for potential confounding factors (including age, gender, race, education level, marital status, and PIR). As illustrated in the participant selection flowchart (Fig 1), 144 individuals were excluded after PSM due to missing data on 8 ALS-related variables required for calculation. To maximize the sample size, we performed multiple imputation for the missing ALS-related variables, resulting in an imputed dataset comprising 325 participants. The results showed that ALS did not act as a mediator in the thyroid cancer-MCHC pathway (Fig 3D, P for IE > 0.05), but served as a complete mediator in the thyroid cancer-RDW pathway (Fig 3E; IE: β = 0.103, P = 0.008; DE: β = 0.389, P = 0.056). However, the analysis of imputed data failed to detect any significant influence of thyroid cancer on diabetes (Fig 3F, P for TE > 0.05).
4. Discussion
This study found that thyroid cancer was significantly associated with ALS even after adjusting for covariates. This association remained robust to missing data and was not confounded by drinking, diabetes, or thyroid disease. Although thyroid cancer did not demonstrate the best predictive value in the ROC curve analysis, it exerted strong explanatory power. Finally, the mediation analysis confirmed that ALS served as a complete mediator in the thyroid cancer-RDW pathway.
Compared with other types of cancer, thyroid cancer exhibited a robust association with elevated AL. Cancer and its treatment introduce numerous physiologic, psychological, social, and economic stressors to the individual with the diagnosis. Although thyroid cancer generally has a favorable prognosis, a systematic review indicated that the quality of life of thyroid cancer survivors was not higher than that of patients with other cancers; instead, it is lower than that of some malignant tumors, such as breast cancer [14]. Many patients feel deprived of support from their families and doctors because thyroid cancer is regarded by medical personnel as a “good kind of cancer”. This perception may underestimate the severity of the diagnosis and thereby increase AL [13]. Furthermore, patients receiving radioactive iodine (RAI) treatment or insufficient doses of levothyroxine may develop hypothyroidism. This condition can lead to a broad spectrum of symptoms, such as fatigue, altered appetite, disrupted sleep patterns, anxiety, and reduced therapeutic efficacy, all of which can significantly elevate AL [25]. The fear of cancer recurrence, concern over the risk of a second primary cancer, and anxiety regarding distant metastasis can impose significant psychological stress on thyroid cancer survivors, despite having received definitive treatment and potentially being in a disease-free state [26].
This study demonstrated that thyroid cancer could elevate RDW levels by increasing AL. Thyroid cancer elevates RDW by increasing AL through multiple pathways (inflammation, metabolic dysregulation, oxidative stress, etc.), which disrupt erythropoiesis and red cell metabolism. First, thyroid cancer may trigger chronic inflammation, releasing pro-inflammatory cytokines that suppress bone marrow erythropoiesis, leading to impaired red blood cell production and increased size heterogeneity, thereby elevating RDW [27–29]. Second, iron-deficiency anemia is a common cause of elevated RDW, which can be exacerbated by impaired nutrient absorption or drug side effects in thyroid cancer [30]. Third, thyroid cancer and its treatments may increase oxidative stress, damaging red blood cell membranes and shortening their lifespan, resulting in greater size variability and elevated RDW [31,32]. Four, abnormal thyroid hormone levels disrupt erythropoiesis by altering bone marrow microenvironments or erythropoietin signaling [33,34]. Lastly, thyroid cancer patients often have autoimmune comorbidities where autoantibodies may attack erythrocyte precursors or disrupt iron metabolism, contributing to elevated RDW [35]. These mechanisms are closely linked to the pathophysiology of the disease and treatment-related side effects.
The findings of this exploratory study highlight the potential clinical relevance of AL in thyroid cancer survivorship. These findings suggest that assessing AL might identify patients at higher risk of physiological dysregulation and could represent a promising target for future interventions aimed at improving long-term outcomes. However, before any clinical application can be considered, further research is essential to validate these findings and to establish standardized, feasible, and clinically relevant biomarker panels for measuring AL in routine practice.
Future longitudinal studies may help to understand whether AL is the cumulative result of the cancer burden. Such studies will also be helpful in examining the AL trajectory to enhance the understanding of susceptibility to cancer progression and mortality [36]. Comprehensive research is essential to elucidate how the interplay among genetics, psychology, physiology, behavior, and environment influences disease risk and treatment response [37]. Furthermore, future studies should aim to identify the optimal combination of biomarkers across various bodily systems for accurately quantifying AL, thereby enhancing comparability between studies [38].
This study has several limitations. First, the cross-sectional design precludes establishing temporal relationships between exposures and outcomes, thereby limiting causal inferences. Second, biomarker data from NHANES were available only at a single time point, which prevented the investigation of longitudinal changes or trajectories in ALS. Additionally, due to the relatively limited sample size in this study, certain subgroups within the subgroup analysis had particularly small sample sizes, leading to insufficient statistical power. As a result, the conclusions derived from these subgroup analyses should be interpreted cautiously. Future research should aim to validate our findings in larger and more representative cohorts. Moreover, cancer diagnosis and type were self-reported by participants, introducing potential recall and reporting biases. The study population comprised American adults, which introduces geographical and ethnic limitations, thereby restricting the generalizability of the findings to populations in other regions. Finally, residual confounding due to unmeasured variables cannot be entirely ruled out.
5. Conclusion
This study demonstrated a robust association between thyroid cancer and ALS. Despite thyroid cancer has limited discriminatory capacity on ALS level, it exerted strong explanatory power. Mediation analyses revealed ALS as a key mediator between thyroid cancer and RDW. Collectively, our findings position ALS as a significant risk marker and a potential biological pathway through which thyroid cancer may influence physiological dysregulation, offering a conceptual framework for understanding the long-term health burdens in this patient population.
Supporting information
S2 Fig. Diagnostic plots for the linear regression model.
https://doi.org/10.1371/journal.pone.0341063.s002
(TIFF)
S3 Fig. Diagnostic plots for the logistic regression model.
https://doi.org/10.1371/journal.pone.0341063.s003
(TIFF)
S1 Table. Time information for 9 patients with thyroid cancer in this study from the 2007–2008 cycle.
https://doi.org/10.1371/journal.pone.0341063.s004
(DOCX)
S2 Table. Multicollinearity analysis between variables.
https://doi.org/10.1371/journal.pone.0341063.s005
(DOCX)
References
- 1. McEwen BS, Stellar E. Stress and the individual. Mechanisms leading to disease. Arch Intern Med. 1993;153(18):2093–101.
- 2. Lai KY, Kumari S, Gallacher J, Webster CJ, Sarkar C. Association between Residential Greenness and Allostatic Load: A Cohort Study. Environ Sci Technol. 2024;58(11):4884–93. pmid:38437596
- 3. Dai Z, Zhou X. Associations between allostatic load and hepatic steatosis and liver fibrosis: evidence from NHANES 2017-2020. BMC Public Health. 2024;24(1):1602. pmid:38879469
- 4. Guidi J, Lucente M, Sonino N, Fava GA. Allostatic Load and Its Impact on Health: A Systematic Review. Psychother Psychosom. 2021;90(1):11–27. pmid:32799204
- 5. López-Cepero A, McClain AC, Rosal MC, Tucker KL, Mattei J. Examination of the Allostatic Load Construct and Its Longitudinal Association With Health Outcomes in the Boston Puerto Rican Health Study. Psychosom Med. 2022;84(1):104–15. pmid:34581702
- 6. Gillespie SL, Anderson CM, Zhao S, Tan Y, Kline D, Brock G, et al. Allostatic load in the association of depressive symptoms with incident coronary heart disease: The Jackson Heart Study. Psychoneuroendocrinology. 2019;109:104369. pmid:31307010
- 7. Zhang T, Yan LL, Chen H-S, Jin H-Y, Wu C. Association between allostatic load and mortality among Chinese older adults: the Chinese Longitudinal Health and Longevity Study. BMJ Open. 2021;11(8):e045369. pmid:34344673
- 8. Juster R-P, McEwen BS, Lupien SJ. Allostatic load biomarkers of chronic stress and impact on health and cognition. Neurosci Biobehav Rev. 2010;35(1):2–16. pmid:19822172
- 9. Kilfoy BA, Zheng T, Holford TR, Han X, Ward MH, Sjodin A. International patterns and trends in thyroid cancer incidence, 1973-2002. Cancer Causes Control. 2009;20(5):525–31.
- 10. Kilfoy BA, Devesa SS, Ward MH, Zhang Y, Rosenberg PS, Holford TR, et al. Gender is an age-specific effect modifier for papillary cancers of the thyroid gland. Cancer Epidemiol Biomarkers Prev. 2009;18(4):1092–100. pmid:19293311
- 11. Boucai L, Zafereo M, Cabanillas ME. Thyroid Cancer: A Review. JAMA. 2024;331(5):425–35.
- 12. Davies L, Welch HG. Thyroid cancer survival in the United States: observational data from 1973 to 2005. Arch Otolaryngol Head Neck Surg. 2010;136(5):440–4. pmid:20479371
- 13. Aschebrook-Kilfoy B, James B, Nagar S, Kaplan S, Seng V, Ahsan H, et al. Risk Factors for Decreased Quality of Life in Thyroid Cancer Survivors: Initial Findings from the North American Thyroid Cancer Survivorship Study. Thyroid. 2015;25(12):1313–21. pmid:26431811
- 14. Applewhite MK, James BC, Kaplan SP, Angelos P, Kaplan EL, Grogan RH, et al. Quality of Life in Thyroid Cancer is Similar to That of Other Cancers with Worse Survival. World J Surg. 2016;40(3):551–61. pmid:26546191
- 15. Lubitz CC, Kong CY, McMahon PM, Daniels GH, Chen Y, Economopoulos KP, et al. Annual financial impact of well-differentiated thyroid cancer care in the United States. Cancer. 2014;120(9):1345–52. pmid:24481684
- 16. Ramsey S, Blough D, Kirchhoff A, Kreizenbeck K, Fedorenko C, Snell K, et al. Washington State cancer patients found to be at greater risk for bankruptcy than people without a cancer diagnosis. Health Aff (Millwood). 2013;32(6):1143–52. pmid:23676531
- 17. Frei R, Haile SR, Mutsch M, Rohrmann S. Relationship of Serum Vitamin D Concentrations and Allostatic Load as a Measure of Cumulative Biological Risk among the US Population: A Cross-Sectional Study. PLoS One. 2015;10(10):e0139217. pmid:26451600
- 18. Du EY, Jiang K, Carlson MC, Reed NS, Deal JA. Hearing Impairment and Allostatic Load in Older Adults. JAMA Otolaryngol Head Neck Surg. 2023;149(7):597–606. pmid:37200015
- 19. Johnson NB, Jones EM, Ovbiagele B, Markovic D, Towfighi A. Effects of Allostatic Load on Long-Term Survival After Stroke. Stroke. 2025;56(1):87–94. pmid:39676664
- 20. Li C, Howard SP, Rogers CR, Andrzejak S, Gilbert KL, Watts KJ, et al. Allostatic Load, Educational Attainment, and Risk of Cancer Mortality Among US Men. JAMA Netw Open. 2024;7(12):e2449855. pmid:39656456
- 21. Wang M, Lu X, Zheng X, Xu C, Liu J. The relationship between sleep duration and thyroid function in the adult US population: NHANES 2007-2012. PLoS One. 2023;18(9):e0291799. pmid:37733750
- 22. Elkahwagy DMAS, Kiriacos CJ, Mansour M. Logistic regression and other statistical tools in diagnostic biomarker studies. Clin Transl Oncol. 2024;26(9):2172–80. pmid:38530558
- 23. Qi X, Wang S, Fang C, Jia J, Lin L, Yuan T. Machine learning and SHAP value interpretation for predicting comorbidity of cardiovascular disease and cancer with dietary antioxidants. Redox Biol. 2025;79:103470. pmid:39700695
- 24. Ou H, Yao Y, He Y. Missing data imputation method combining random forest and generative adversarial imputation network. Sensors (Basel). 2024;24(4).
- 25. Dow KH, Ferrell BR, Anello C. Quality-of-life changes in patients with thyroid cancer after withdrawal of thyroid hormone therapy. Thyroid. 1997;7(4):613–9. pmid:9292951
- 26. Lee JI, Kim SH, Tan AH, Kim HK, Jang HW, Hur KY, et al. Decreased health-related quality of life in disease-free survivors of differentiated thyroid cancer in Korea. Health Qual Life Outcomes. 2010;8:101. pmid:20840792
- 27. Guarino V, Castellone MD, Avilla E, Melillo RM. Thyroid cancer and inflammation. Mol Cell Endocrinol. 2010;321(1):94–102. pmid:19835928
- 28. Tafani M, De Santis E, Coppola L, Perrone GA, Carnevale I, Russo A, et al. Bridging hypoxia, inflammation and estrogen receptors in thyroid cancer progression. Biomed Pharmacother. 2014;68(1):1–5. pmid:24286852
- 29. Marzouk H, Mostafa N, Khalifa I, Badawi N, Mohamed Fathy Sabry NI. Red Cell Distribution Width (RDW) as a Marker of Subclinical Inflammation in Children with Familial Mediterranean Fever. Curr Rheumatol Rev. 2020;16(4):298–303.
- 30. Aulakh R, Sohi I, Singh T, Kakkar N. Red cell distribution width (RDW) in the diagnosis of iron deficiency with microcytic hypochromic anemia. Indian J Pediatr. 2009;76(3):265–8. pmid:19205647
- 31. Joosse H-J, van Oirschot BA, Kooijmans SAA, Hoefer IE, van Wijk RAH, Huisman A, et al. In-vitro and in-silico evidence for oxidative stress as drivers for RDW. Sci Rep. 2023;13(1):9223. pmid:37286717
- 32. Muzza M, Pogliaghi G, Colombo C, Carbone E, Cirello V, Palazzo S, et al. Oxidative Stress Correlates with More Aggressive Features in Thyroid Cancer. Cancers (Basel). 2022;14(23):5857. pmid:36497339
- 33. Baliram R, Latif R, Zaidi M, Davies TF. Expanding the Role of Thyroid-Stimulating Hormone in Skeletal Physiology. Front Endocrinol (Lausanne). 2017;8:252. pmid:29042858
- 34. Bebeshko VG, Bruslova KM, Boyarska OY, Lyashenko LO, Tsvyetkova NM, Gonchar LO, et al. Endocrine regulation of erythroid lineage of hematopoiesis in children living under a low-dose radiation exposure after the chornobyl npp accident. Probl Radiac Med Radiobiol. 2020;25:374–89. pmid:33361848
- 35. Antonelli A, Ferrari SM, Corrado A, Di Domenicantonio A, Fallahi P. Autoimmune thyroid disorders. Autoimmun Rev. 2015;14(2):174–80. pmid:25461470
- 36. Mathew A, Doorenbos AZ, Li H, Jang MK, Park CG, Bronas UG. Allostatic Load in Cancer: A Systematic Review and Mini Meta-Analysis. Biol Res Nurs. 2021;23(3):341–61. pmid:33138637
- 37. Bu S, Li Y. Physical activity is associated with allostatic load: Evidence from the National Health and Nutrition Examination Survey. Psychoneuroendocrinology. 2023;154:106294. pmid:37216739
- 38. Rosemberg M-AS, Granner J, Li Y, Seng JS. A scoping review of interventions targeting allostatic load. Stress. 2020;23(5):519–28. pmid:32602798