Concurrent Use of Thyroid Hormone Therapy and Interfering Medications in Older US Veterans

Abstract

Context

Thyroid hormone management in older adults is complicated by comorbidities and polypharmacy.

Objective

Determine the prevalence of concurrent use of thyroid hormone and medications that can interfere with thyroid hormone metabolism (amiodarone, prednisone, prednisolone, carbamazepine, phenytoin, phenobarbital, tamoxifen), and patient characteristics associated with this practice.

Design

Retrospective cohort study between 2004 and 2017 (median follow-up, 56 months).

Setting

Veterans Health Administration Corporate Data Warehouse.

Participants

A total of 538 137 adults ≥ 65 years prescribed thyroid hormone therapy during the study period.

Main Outcome Measure

Concurrent use of thyroid hormone and medications interfering with thyroid hormone metabolism.

Results

Overall, 168 878 (31.4%) patients were on at least 1 interfering medication while on thyroid hormone during the study period. In multivariable analyses, Black/African-American race (odds ratio [OR], 1.25; 95% CI, 1.21-1.28, compared with White), Hispanic ethnicity (OR, 1.12; 95% CI, 1.09-1.15, compared with non-Hispanic), female (OR, 1.11; 95% CI, 1.08-1.15, compared with male), and presence of comorbidities (eg, Charlson/Deyo Comorbidity Score ≥ 2; OR, 2.50; 95% CI, 2.45-2.54, compared with 0) were more likely to be associated with concurrent use of thyroid hormone and interfering medications. Older age (eg, ≥ 85 years; OR, 0.48; 95% CI, 0.47-0.48, compared with age 65-74 years) was less likely to be associated with this practice.

Conclusions and Relevance

Almost one-third of older adults on thyroid hormone were on medications known to interfere with thyroid hormone metabolism. Our findings highlight the complexity of thyroid hormone management in older adults, especially in women and minorities.

Thyroid hormone use is widespread, with levothyroxine being consistently in the top 3 most commonly prescribed medications in the United States (1). Population-based studies have shown that, independent of sex and race, older age is a significant predictor of thyroid hormone initiation (2, 3). Several factors have probably contributed to this disproportionate thyroid hormone use in older adults, including increased screening for hypothyroidism, as well as a falling threshold of serum TSH for thyroid hormone initiation (4, 5). Once initiated, thyroid hormone therapy continues lifelong for nearly 90% of patients, contributing to polypharmacy (6).

Effective treatment of hypothyroidism in older adults relies on consideration of potential drug interactions given the rates of polypharmacy seen in this population. Although the exact definition of polypharmacy remains controversial, the consensus of its negative implications is clear (7). Attention to the drug–drug interactions that may potentiate or reduce the effects of levothyroxine is crucial because many common medications are known to affect thyroid hormone via varying mechanisms (8, 9). In the case of thyroid hormone therapy, both overtreatment and undertreatment have been shown to have adverse effects, particularly in the cardiovascular and skeletal systems (10-18). Close monitoring of thyroid function tests, in addition to clinical symptoms, can help mitigate these risks to ensure maintenance of euthyroidism.

To better understand the prevalence of and patient characteristics associated with concurrent use of thyroid hormone and several commonly used interfering medications in older adults, we conducted a retrospective cohort study using a national cohort of older adults aged 65 years and older on thyroid hormone.

Methods

Data Source and Study Population

This was a retrospective cohort study of adults aged 65 years or older on thyroid hormone therapy. The study used deidentified national data from the Veterans Health Administration (VHA) Corporate Data Warehouse (CDW) between January 1, 2004, and December 31, 2017. The VHA is the largest integrated health care system in the United States, providing service to more than 9 million veterans across more than 1200 facilities (19). The CDW is a central collection of databases that contains patient-level data from across the VHA’s national health system (20).

Our study included 538,137 thyroid hormone users aged 65 years or older who were on thyroid hormone therapy (identified as taking either T4 alone or a T4 plus T3 combination preparation) at any point during the study period. This was determined via data obtained from the Veterans Administration pharmacy files regarding prescription dispensing and fulfillment. This study was exempt from the University of Michigan institutional review board and approved by the Ann Arbor Veteran Affairs institutional review board.

Measures

Study Outcome

Our study outcome was concurrent use of thyroid hormone and at least 1 of the following medications known to interfere with thyroid hormone metabolism during the study period: amiodarone, prednisone, prednisolone, carbamazepine, phenytoin, phenobarbital, tamoxifen. Several medications or medication classes, such as amiodarone, glucocorticoids, and antiepileptic agents, interfere with thyroid hormone on multiple levels (21). In particular, we chose to focus on medications that may affect thyroid hormone transport and metabolism through a variety of mechanisms. For example, antiepileptics such as phenobarbital, carbamazepine, and phenytoin increase hepatic enzymes that may reduce the half-lives of T4 and T3. Amiodarone can impair the peripheral de-iodination of T4 and thus conversion of T4 to T3. Similarly, short-term administration of large doses of glucocorticoids, such as prednisone or prednisolone, can decrease the peripheral conversion of T4 to T3, whereas long-term glucocorticoid therapy may result in slightly decreased T3 and T4 levels because of decreased thyroxine-binding globulin production. Last, tamoxifen increases thyroxine-binding globulin levels, thus reducing availability of free T4.

The median follow-up was 59 months. Data were censored at death or last follow-up.

Covariates

Data collected included information on patient sex, age, race, ethnicity, and number of comorbidities. Patient sex was recorded at study entry as described in the VHA CDW database. We categorized patient age into the following clinically meaningful categories: 65 to 74, 75 to 84, and ≥ 85 years. Race was self-reported as White, Black/African-American, Asian, Native Hawaiian or Pacific Islander, American Indian or Alaskan Native, multiracial, or unknown. Because of small numbers, Asian, Native Hawaiian or Pacific Islander, American Indian or Alaskan Native, and multiracial were combined into 1 category and reclassified into an “Other” category for the analyses. Ethnicity was self-reported as Hispanic, non-Hispanic, or unknown. As done in our previous studies, comorbidities were determined by the Charlson/Deyo Comorbidity Score, calculated by using the Charlson/Deyo Comorbidity Score mapping table (22-25). The Charlson/Deyo Comorbidity Score is an age-independent score that incorporates 17 weighted comorbidity conditions (eg, myocardial infarction, dementia, diabetes with chronic complications, metastatic solid tumor). Each comorbidity category has an associated weight (from 1 to 6) based on the adjusted risk of mortality or resource use, and the sum of all the weights results in a single comorbidity score for a patient (22, 23). For example, myocardial infarction has an associated weight of 1, diabetes with chronic complications has an associated weight of 2, metastatic solid tumor has an associated weight of 6, and so on. A score of 0 indicates that no comorbidities were found. This score was categorized as 0, 1, and ≥ 2.

Statistical Analysis

First, we generated descriptive statistics. Next, we described the prevalence of concurrent use of thyroid hormone and medications that commonly interfere with the thyroid hormone metabolism as outlined previously. Then, we performed a multivariable logistic regression analysis to determine patient characteristics associated with concurrent use of thyroid hormone and at least 1 interfering medication during the study period. Variables included in the regression model were patient sex, age, race, ethnicity, and number of comorbidities as defined by the Charlson/Deyo Comorbidity Score. No observations were excluded from the statistical analyses because of missing information. Analyses were conducted using SAS Enterprise Guide Version 7.15 HF3 (SAS Institute Inc, Cary, NC, USA). A 95% CI was used to determine statistical significance.

Results

The characteristics of the patients on thyroid hormone therapy included in the study are shown in Table 1. Majority were male (96.5%), White (77.1%), non-Hispanic (82.3%), and had 2 or more comorbidities (62.6%). Median patient age was 77 years (range, 65-110).

Overall, 168 878 (31.4%) patients were on at least 1 interfering medication while on thyroid hormone during the study period, as demonstrated in Table 2. The most frequent interfering medication was corticosteroids with prednisone and prednisolone coadministration with thyroid hormone occurring in 15.0% and 13.4% of the patients, respectively. Details on the prevalence of the other studied medications interfering with thyroid hormone metabolism are shown in Table 2.

Table 3 shows that Black/African-American race (odds ratio [OR], 1.25; 95% CI, 1.21-1.28, compared with White), Hispanic ethnicity (OR, 1.12; 95% CI, 1.09-1.15, compared with non-Hispanic), female (OR, 1.11; 95% CI, 1.08-1.15, compared with male), and presence of comorbidities (eg, Charlson/Deyo Comorbidity Score ≥ 2; OR, 2.50; 95% CI, 2.45-2.54, compared with 0) were more likely to be associated with concurrent use of thyroid hormone and interfering medications. Older age (eg, ≥85 years; OR, 0.48; 95% CI, 0.47-0.48, compared with age 65-74 years) was less likely to be associated with this practice.

Discussion

In this large retrospective cohort study, we found that nearly one-third of thyroid hormone users aged 65 years and older were on at least 1 medication that has been shown to interfere with thyroid hormone. This practice was more likely in women, Black race, Hispanics, and patients with comorbidities. Increasing age was less likely to be associated with concurrent use of thyroid hormone and interfering medications.

To our knowledge, there are no prior studies that have examined the prevalence of concurrent use of thyroid hormone therapy and possible interfering medications in older adults. However, there have been a number of previous studies that have shown a high prevalence of polypharmacy leading to drug–drug interactions among older adults (26, 27). A previous cross-sectional, population-based study evaluated the potential drug–drug interactions between medications used by older adults (N = 934). They found that the prevalence of these interactions was 36.9%, with medications with a narrow therapeutic index, such as levothyroxine, being more prone to serious interactions (26). Another study of more than 2 billion patient visits using data from the Centers for Disease Control and Prevention’s National Ambulatory Medical Care Survey from 2009 to 2016 found that 65.1% of patients aged 65 years and older had experienced some degree of polypharmacy (27). They also found patients with major polypharmacy (> 5 medications) to be older compared with those with moderate (4-5 medications) or minor polypharmacy (2-3 medications) (27), whereas our study found that concurrent use of thyroid hormone and interfering medications decreased with older age. As expected, patients with more comorbidities were more likely to be on medications that interfere with thyroid hormone because it is well-known that age-related comorbidities predispose older adults to polypharmacy.

We found that women were more likely to be on concurrent medications that interfere with thyroid hormone metabolism. Similarly, a retrospective study among patients aged 65 years and older found that polypharmacy was more prevalent among women relative to men, with an average of 6.28 medications for women vs an average of 5.69 medications for men (28). In addition, several prior studies have shown that older women are at increased odds of receiving potentially inappropriate prescriptions (29-31). Several explanations behind this observed trend have been proposed including varying biological and social influences. Biological considerations include the differences in the prevalence of different medical conditions between sexes for which a given treatment may be prescribed. Other plausible factors include the varying use of health services and the number of care providers that may in turn increase the risk of polypharmacy. Furthermore, implicit bias may affect how health professionals diagnose and treat women and men who present with similar illnesses.

With regard to race/ethnicity, prior work has demonstrated that minority groups have high rates of polypharmacy (32, 33). In our study, we observed an increased likelihood of concurrent use of interfering medications and thyroid hormone in patients of Black/African-American race and Hispanic ethnicity. The factors that predispose certain groups to these practices remain unclear. As noted previously, it is likely a combination of biologic and social influences including differences in the prevalence of certain conditions among minorities as well as implicit bias. A prior study suggested that race and socioeconomic status have intertwined effects in those vulnerable to polypharmacy, supporting that race/ethnicity and socioeconomic status have combined/interdependent rather than independent effects on polypharmacy (34).

Major strengths of our study include the large cohort of US patients with detailed descriptive and prescription data. Moreover, our research question is novel because the concurrent use of interfering medications with thyroid hormone was not previously explored as it pertains to older adults. Yet, several limitations exist. Our database was composed of a male-predominant population. However, given the large size of the cohort, women were still adequately represented, with more than 18,000 women in our study sample. Additionally, even though the VHA provides an incentive for eligible veterans to fill all medications within the Veterans Administration system, it is possible that some veterans used outside health care systems or obtained over-the-counter medications, which could not be ascertained. We did not include known interfering medications that are widely available over the counter in our study, such as iron supplements, proton-pump inhibitors, and antacids because of an inability to fully account for their usage accurately using the Veterans Administration’s electronic medical records. Thus, the actual prevalence of older adults on thyroid hormone and interfering medications is likely much greater, making the results of our study a conservative representation of the magnitude of the issue. In addition, data on dosage and duration of use of the interfering medications was not available for our study. Last, we acknowledge that socioeconomic factors such as education level and household income status may be confounding variables that we were unable to account for because of a lack of available data.

With little prior evidence available, our study underscores that concurrent use of thyroid hormone and interfering medications in older adults is fairly common, which may increase the risk of drug interactions and the potential for side effects in an already vulnerable patient population. Our findings highlight the complexity of thyroid hormone management in older adults in the context of polypharmacy and multimorbidity and draw attention to at risk groups including women and minorities. Because thyroid function tests are important to determine appropriate thyroid hormone dose and effectiveness of therapy, and because many older patients often use multiple medications for numerous chronic conditions that can interfere with biochemical results, prescribers should be vigilant to ensure safe medication use in their older patients. Additionally, providers should conduct a comprehensive medication review and reconciliation and discuss the appropriateness of medication use with their older patients. It is imperative to carefully follow thyroid function tests to avoid prolonged periods of over- or undertreatment, which could have significant clinical adverse effects for many older patients.

Abbreviations

Abbreviations
 
• CDW

  Corporate Data Warehouse

 
• OR

  odds ratio

 
• VHA

  Veterans Health Administration

Acknowledgments

We acknowledge Brittany Gay who assisted with literature review and manuscript formatting and submission and Jennifer Burns who assisted with statistical analyses.

Funding

This work was supported by the National Institute on Aging of the National Institutes of Health (NIH) under Award Number K08 AG049684 and by a pilot grant cofunded by the Claude D. Pepper Older Americans Independence Center (OAIC), the Michigan Institute for Clinical and Health Research (MICHR), and the Michigan Biology of Cardiovascular Aging (M-BoCA) to M.P. A.B.C. is supported by the National Institute on Aging of the National Institutes of Health (NIH) under Award Number K08 AG071856.

Disclosures

The authors have nothing to disclose.

Data Availability

Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

References