Understanding worry about risks associated with thyroid hormone therapy: A national survey of endocrinologists, family physicians and geriatricians

Abstract

Objective:

Thyroid hormone use is widespread and prior studies have shown that over- and undertreatment are common. Our objective was to understand physician worry regarding risks associated with thyroid hormone therapy, specifically overtreatment or undertreatment.

Methods:

A nationwide survey was administered to physician members of the Endocrine Society, the American Academy of Family Practice, and the American Geriatrics Society. Participants were asked how often they were worried about various risks that may be associated with thyroid hormone over- or undertreatment, i.e., cardiovascular complications, bone complications, and poor quality of life due to overtreatment or undertreatment with thyroid hormone. Multivariable regression analyses were conducted to determine physician characteristics associated with each worry.

Results:

Response rate was 63% (359/566); 36% were primary care physicians, 32% were endocrinologists and 32% were geriatricians. Overall, 74 (21%) physicians reported they frequently or always worried about cardiovascular complications, 74 (21%) about bone complications, 111 (31%) about poor quality of life due to symptoms from undertreatment with thyroid hormone, and 87 (24%) about poor quality of life due to symptoms from overtreatment with thyroid hormone. Endocrinologists were more likely to frequently/always worry about patients’ poor quality of life due to symptoms from overtreatment [odds ratio (OR), 2.05; 95% confidence interval (CI), 1.09–3.93], compared to primary care physicians.

Conclusion:

Up to one-third of physicians frequently or always worried about risks resulting from thyroid hormone overtreatment or undertreatment. More research is needed across specialties to understand physician perceptions of how thyroid hormone therapy impacts patients’ quality of life.

INTRODUCTION

Thyroid hormone use is common, with levothyroxine being one of the top three most frequently prescribed medications in the United States since 2012.^1–3 Levothyroxine is considered the standard of care for hypothyroidism and nearly 90% of patients remain on lifelong levothyroxine treatment after diagnosis.⁴ The role of managing thyroid hormone therapy is assumed by providers across multiple specialties, including primary care physicians and endocrinologists. However, maintaining euthyroid levels can sometimes be challenging and often result in either over- or undertreatment despite frequent monitoring and dosage adjustments.^5–7 This can be particularly problematic in vulnerable populations, such as older adults and those with various comorbidities, including preexisting cardiovascular disease and osteoporosis.

Prior studies have shown that both over- and undertreatment with thyroid hormone can result in deleterious effects and cause patient harm.^8–17 Both over- and undertreatment with thyroid hormone have been shown to lead to an increased risk of cardiovascular disease and mortality.^14–17 Undertreatment has been found to be associated with decreased quality of life due to increased anxiety and depression.^{12, 13} Additionally, overtreatment with thyroid hormone carries a greater risk of adverse skeletal effects, such as increased levels of bone loss and fracture risk, especially in post-menopausal women.^{8–11, 18} Despite widespread use of thyroid hormone and significant potential risks associated with thyroid hormone over- and undertreatment, physician worry regarding risks associated with thyroid hormone therapy remains underexplored.

In order to better understand physician worry regarding risks of thyroid hormone therapy as they relate to either over- or undertreatment, we conducted a nationwide survey among a diverse group of physicians, including primary care physicians, endocrinologists, and geriatricians.

METHODS

Study Population

Physicians were randomly selected from active membership lists of the Endocrine Society, the American Academy of Family Practice, and the American Geriatrics Society (N = 600). The random samples were generated from those physicians with complete information using the functions in Excel by a research assistant, while the primary investigators were blinded to the process. Of these, 566 physicians who manage thyroid hormone therapy were eligible to be surveyed. Of the 34 physicians that were ineligible to be surveyed, 9 were unreachable, 1 was deceased, 2 were retired, and 22 reported not treating patients with thyroid disease. Response rate was 63% (359/566).

The modified Dillman method of survey administration was used to enhance the response rate.¹⁹ We initially mailed an introductory letter, the survey instrument, a postage-paid return envelope, and a small monetary gift to eligible physicians. Three weeks later, we mailed a postcard reminder. Following another three weeks, we mailed a second survey with a postage-paid return envelope to non-respondents. Follow-up telephone calls were also conducted for non-respondents. The survey was completed by physicians in 2018. The data were de-identified and electronically entered into a database using a double entry method in order to ensure <1% error.

The University of Michigan Institutional Review Board granted an exemption for this study.

Survey Design and Measures

A team of endocrinologists, primary care physicians, geriatricians, and survey methodologists developed the survey instrument based on literature review and clinical expertise and reviewed it for content validity. Prior to administration, the survey was piloted among a diverse group of physicians at the University of Michigan and the survey instrument was refined based on feedback.

Covariates

Physicians were asked to indicate their specialty, practice setting and years in practice. Physician specialty was categorized as primary care, endocrinology, or geriatrics. Physicians who reported their specialty as family practice or internal medicine were categorized as primary care physicians. Practice setting was categorized as academic tertiary care center, community-based academic affiliate, private practice, and other. Years in practice were categorized as 0–5 years, 6–10 years, 11–20 years, and >20 years. In order to create groups of appropriate sample size for further analyses, the number of years practiced were re-categorized as 0–10 years, 10–20 years, and > 20 years.

Physician reported worry about risks when treating a patient with thyroid hormone

Physicians were asked to rate how often they were worried about the following: a) cardiovascular complications (e.g. cardiac arrhythmias and heart failure) b) bone complications (e.g. osteoporosis), c) poor quality of life due to symptoms from undertreatment with thyroid hormone, and d) poor quality of life due to symptoms from overtreatment with thyroid hormone. A 5-point Likert scale was used as follows: always, frequently, occasionally, rarely, and never. For the purpose of the analyses, physician worry was dichotomized as frequently to always worried versus never to occasionally worried.

Statistical Analyses

First, summary descriptive statistics were generated using frequencies and percentages. Univariate analyses were performed using chi-square tests to determine the association between the covariates and the dichotomized outcomes of physician reported worry of cardiovascular complications, bone complications, and poor quality of life due to symptoms from undertreatment or overtreatment with thyroid hormone.

Multivariable logistic regression analyses, with separate models for each physician worry related to thyroid hormone treatment, were performed to determine the independent effect of the covariates on physician worry. The covariates used in the models were physician specialty, practice setting, and years in practice.

A p-value of <0.05 was considered to be statistically significant. A 95% confidence interval (CI) that did not include the null value was used to determine statistical significance. All statistical analyses were carried out using R version 3.5.2.

RESULTS

Table 1 summarizes physician characteristics of the respondents (missing data not included). Of the respondents who reported their specialty, 128 (36%) were primary care physicians, 114 (32%) were endocrinologists, and 113 (32%) were geriatricians. A total of 186 (52%) respondents were in practice for more than 20 years, 102 (29%) for 11–20 years, and 67 (19%) for 10 or fewer years. Over half of the respondents practiced in private practice (51%), whereas 31% practiced in community-based academic affiliates and 18% in academic tertiary care centers. Of note, 316 (89%) of respondents stated that they almost always use T4 monotherapy when treating patients with hypothyroidism, with endocrinologists being more likely to use preparations other than T4 monotherapy (i.e., desiccated thyroid extracts, T4/T3 combination therapy, over-the-counter nutraceuticals) compared to geriatricians and primary care physicians (p<0.001).

Table 1.

Physician Demographics (N=359)^*

Physician Demographics	N (%)
Specialty
Primary Care	128 (36)
Endocrinology	114 (32)
Geriatrics	113 (32)
Years in Practice
0–10	67 (19)
11–20	102 (29)
>20	186 (52)
Practice Setting
Private Practice	173 (51)
Community-Based Academic Affiliate	106 (31)
Academic Tertiary Care Center	63 (18)

^*Missing data not included

Figure 1 shows the frequency of physician worry about cardiac complications (Figure 1a), bone complications (Figure 1b), quality of life due to symptoms from undertreatment (Figure 1c) and quality of life due to symptoms from overtreatment with thyroid hormone (Figure 1d) stratified by physician specialty. Overall, 74 (21%) physicians reported they frequently or always worried about cardiovascular complications, 74 (21%) about bone complications, 111 (31%) about poor quality of life due to symptoms from undertreatment with thyroid hormone, and 87 (24%) about poor quality of life due to symptoms from overtreatment with thyroid hormone. Additionally, 120 (33%) physicians reported they occasionally worried about cardiovascular complications, 122 (34%) about bone complications, 109 (30%) about poor quality of life due to symptoms from undertreatment with thyroid hormone, and 115 (32%) about poor quality of life due to symptoms from overtreatment with thyroid hormone.

Figure 1.

Frequency of physician worry about cardiac complications (Figure 1a), bone complications (Figure 1b), quality of life due to symptoms from undertreatment (Figure 1c) and quality of life due to symptoms from overtreatment with thyroid hormone (Figure 1d) stratified by physician specialty.

Table 2 demonstrates the results of the multivariable logistic regression analysis of physician characteristics associated with worry of poor quality of life due to symptoms from overtreatment with thyroid hormone. Endocrinologists (odds ratio [OR] = 2.05 [CI 1.09–3.93], compared to primary care physicians) were more likely to frequently to always worry about poor quality of life due to symptoms from overtreatment after controlling for years in practice and practice setting. Multivariable analyses for the other three worry outcomes (worry about cardiovascular complications, worry about bone complications, and worry about poor quality of life due to symptoms from undertreatment with thyroid hormone) did not yield any statistically significant associations.

Table 2.

Physician worry about patient poor quality of life due to symptoms from thyroid hormone overtreatment

Physician Characteristics	N (%)**	OR	95% CI
Specialty
Primary care	23 (27)	Ref
Geriatrics	27 (31)	1.41	0.70–2.86
Endocrinology	36 (42)	2.05	1.09–3.93
Years in practice
0–10	21 (24)	Ref
11–20	31 (36)	1.16	0.57–2.38
>20	35 (40)	0.58	0.29–1.18
Practice setting
Private practice	43 (52)	Ref
Community-based academic affiliate	24 (29)	0.78	0.41–1.43
Academic tertiary care center	16 (19)	0.76	0.36–1.54

^**Physicians who always or frequently worry

Abbreviations: OR, Odds Ratio; CI, Confidence Interval

DISCUSSION

In this nationwide survey of a diverse group of physicians we found that up to one-third of physicians treating patients on thyroid hormone therapy always or frequently worried and an additional third occasionally worried about risks resulting from thyroid hormone over- or undertreatment. Endocrinologists were more likely than primary care physicians and geriatricians to worry about poor quality of life due to overtreatment after controlling for practice setting and years of in practice.

Several prior studies have shown that both over- and undertreatment with thyroid hormone are common^{20, 21} and can be associated with risks, including adverse cardiac and skeletal effects, as well as increased mortality.^8–17 In a large retrospective cohort study of patients treated for hypothyroidism (N=162,369), the authors found evidence of adverse health outcomes, including increased risk of ischemic heart disease, heart failure, mortality and fragility fractures, when the TSH concentration was outside the euthyroid range.¹⁶ A registry-based case-control study nested within a population-based cohort of 275,467 individuals between 1995–2011 found that, among thyroid hormone users, cardiovascular risk increased as a function of the cumulative duration of both decreased TSH (overtreatment, p<0.001) and elevated TSH (undertreatment, p=0.001).¹⁵ Similarly, in another study by the same investigators (N=235,168), it was shown that both over- and undertreatment of hypothyroidism was significantly associated with excess mortality compared to euthyroid controls.¹⁴

In view of significant risks that can be associated with both over- and undertreatment with thyroid hormone, understanding physician worry about these risks and factors associated with worry is important as it may guide decision-making. We found that approximately one-third of physicians frequently or always worried about risks associated with thyroid hormone over- or undertreatment, with another third occasionally being worried. We also found that endocrinologists, but not geriatricians, were more likely to worry about poor quality of life from thyroid hormone overtreatment compared to primary care physicians. The reasons for these findings are unclear and may be relevant to variation in practice patterns, physician knowledge, experiences and perceptions, patient preferences or all of the above. A case-based mail survey of American Thyroid Association members and primary care providers regarding hypothyroidism management, showed that primary care physicians were more likely to choose a broader TSH goal overall when treating patients with hypothyroidism, whereas endocrinologists were more likely to choose a narrower range for TSH in younger patients and broadening that range in the older adults. For younger patients, primary care physicians were more likely to start a lower dose of thyroid hormone and titrate upward, whereas endocrinologists were more likely to initiate a full replacement dose.²² Possible reasons to explain why endocrinologists were more likely to frequently or always worry about patient poor quality of life due to symptoms from overtreatment in our study may stem from an attempt to target a narrower TSH range or the higher use of thyroid hormone preparations other than T4 monotherapy, thus possibly increasing likelihood of overtreatment. Furthermore, geriatricians and primary care physicians have to prioritize issues to address when caring for complex, multimorbid patients who typically have a long problem list that can influence quality of life, and thyroid hormone therapy may be low on their priority list.

Our study has several strengths, including nationwide sampling of a diverse group of physicians who manage patients on thyroid hormone therapy, including endocrinologists, geriatricians, and primary care physicians, a high response rate and a novel research question. Our study also has several limitations. Inherent to survey studies, there is a risk of non-response bias, which is partly mitigated by our high response rate. Additionally, other factors that could influence physician worry, such as patient volume and type of thyroid hormone preparation used, were not included in our multivariable model. Finally, this study focused on physician perspectives on risks of thyroid hormone over- or undertreatment. We acknowledge that worry about risks of thyroid hormone therapy as it relates to over- and undertreatment may also depend on patient and system characteristics.^{23, 24}

Our study has implications for physicians managing thyroid hormone therapy and their patients. As thyroid hormone treatment for most patients can be lifelong, it is important to gain insight on physician worry regarding risks of thyroid hormone over- and undertreatment and how this may impact long-term thyroid hormone management. Future studies should be conducted to further understand physicians’ worry about risk of over- and undertreatment with thyroid hormone in specific patient scenarios and elicit patients’ perspectives pertaining to potential risks from thyroid hormone therapy and impact on long-term management. Additionally, more research is needed across specialties to understand physician perceptions of how thyroid hormone therapy impacts patients’ quality of life.

HIGHLIGHTS.

• Up to one-third of physicians frequently or always worried about risks resulting from thyroid hormone overtreatment or undertreatment, including cardiovascular and bone complications and poor quality of life due to symptoms from under- and overtreatment.

• Endocrinologists were more likely to frequently/always worry about patients’ poor quality of life due to symptoms from overtreatment, compared to primary care physicians.

• More research is needed across specialties to understand physician perceptions of how thyroid hormone therapy impacts patients’ quality of life.