Testosterone Prescribing in the United States, 2002-2016

Abstract

This study uses commercial claims database data to characterize trends in testosterone prescribing in the United States from 2002 to 2016.

Testosterone use in the United States tripled from 2001 through 2011, mostly in men without a clear indication.^1,2 In late 2013 and early 2014, 2 studies reported increased myocardial infarction and stroke associated with testosterone use.^3,4 The US Food and Drug Administration (FDA) issued a safety bulletin on January 31, 2014. We assessed recent trends in testosterone prescribing.

Methods

We used data from Clinformatics Data Mart (CDM), one of the nation’s largest commercial health insurance databases with more than 18 000 000 enrollees. Persons in the south and those aged 21 to 64 years are overrepresented in the data set. Separate cohorts were identified each year from 2002 through 2016, consisting of men 30 years and older with continuous enrollment during the index and prior year. For each year, we reported the percentage of men who were prescribed testosterone therapy and new testosterone users (no use in the prior year). We examined use over time with an interrupted time series analysis using joinpoint regression. Because testosterone use differs by age and region,^1,2 we stratified by these factors. All analyses were performed using SAS (SAS Institute), version 9.4, and the Joinpoint Regression Program (National Cancer Institute), version 4.5.0.1. Statistical tests were 2-tailed and significant at an α level of .05. This study was approved by the University of Texas Medical Branch institutional review board, which waived informed consent.

Results

We examined the records of 9 962 538 men 30 years and older from 2002 through 2016, with a minimum of 1 823 000 men in any year. Over time, the median age increased (46 years in 2002 to 53 years in 2016) as did the percentage living in the south (39.5% in 2002 to 42.1% in 2016). Total testosterone use increased among men from 0.52% (95% CI, 0.51% to 0.53%) in 2002 to 3.20% (95% CI, 3.18% to 3.22%) in 2013, then decreased to 1.67% (95% CI, 1.66% to 1.69%) in 2016 (Figure 1). For new users, the rate increased from 0.28% (95% CI, 0.27% to 0.29%) in 2002 to 1.26% (95% CI, 1.25% to 1.28%) in 2013, then decreased to 0.48% (95% CI, 0.48% to 0.49%) in 2016. The relative decrease between 2013 and 2016 was 48% (95% CI, 47%-48%) in established users and 62% (95% CI, 61%-62%) in new users. Joinpoint analysis showed significant changes in the annual percentage change of total users per year over time, from 0.10% (P = .007) for 2002-2007 to 0.34% (P = .001) for 2007-2013 and then −0.50% (P = .001) for 2013-2016. The inset in Figure 1 shows a 22% (95% CI, 19% to 26%) relative decrease in new testosterone users from October 2013 to December 2013, following a publication linking testosterone to cardiovascular adverse events³ and an additional 50% (95% CI, 47% to 53%) relative decrease over the next 8 months following a similar study⁴ and an FDA safety communication.

Figure 1. Total and New Testosterone Use Among Men 30 Years or Older in the United States, 2002-2016.

Denominators were calculated for each calendar year. Each denominator included all men who were ≥30 years at the start of the calendar year with continuous benefits for the entire study year and prior year. The denominators range from 1 823 677 in 2002 to 2 856 954 in 2016. Error bars represent 95% CIs. Interrupted time series analysis with joinpoint regression was used to assess time-related trends in testosterone use. The analysis allowed for a maximum of 5 joinpoints (indicated by red arrows). For total testosterone users, joinpoints were located at 2007 (95% CI, 2005-2010) and 2013 (95% CI, 2012-2014). For new testosterone users, joinpoints were located at 2007 (95% CI, 2004-2012) and 2012 (95% CI, 2010-2014). The inset presents new testosterone prescription rates by month, from January 2013 through December 2016. Denominators for monthly rates included all men ≥ 30 years at the start of the month with continuous benefits for the entire month and the 12 previous months. The listed numbers indicate the following specific dates: (1) article by Vigen et al³ released online November 6, 2013; (2) article by Finkle et al⁴ released online January 29, 2014; (3) US Food and Drug Administration (FDA) safety communication on testosterone therapy, January 31, 2014; (4) FDA advisory committee meeting on possible cardiovascular risks associated with testosterone therapy, September 17, 2014; (5) FDA requires testosterone label change indicating possible increased risk of myocardial infarction and stroke, March 3, 2015.

The decline in new testosterone users occurred in all age groups, ranging from 0.88% (95% CI, 0.84% to 0.91%) in 2013 to 0.41% (95% CI, 0.39% to 0.43%) in 2016 among men aged 30 to 39 years (relative decrease, 53% [95% CI, 51% to 56%]) to 0.86% (95% CI, 0.84% to 0.88%) in 2013 to 0.26% (95% CI, 0.25% to 0.27%) in 2016 among men 65 years and older (relative decrease, 69% [95% CI, 68% to 71%]) (Figure 2). The percentage of new testosterone users differed by region, but the relative decreases in the 4 regions were similar.

Figure 2. Annual Rates of New Testosterone Use Among Men in the United States by Age and Region, 2002-2016.

Denominators were calculated for each calendar year. Each denominator included men with continuous benefits for the entire study year and prior year. Regions were based on US census regions. Error bars reflect the 95% CIs. A, The denominators for any age category in any year ranged from 284 602 to 1 109 387. B, The denominators for any region category in any year ranged from 11 999 to 1 106 148.

Discussion

After a decade of growth, the percentage of US men receiving testosterone prescriptions decreased from 2013 through 2016. The steepest decrease coincided with 2 published reports of testosterone-associated adverse cardiovascular events^3,4 and an FDA communication.

This study has limitations. First, commercial insurance data selected for employed males; the results for men 65 years and older are not generalizable to most older men who are retired and on Medicare. Second, men may obtain testosterone from clinicians not reimbursed by their insurance. Third, indications for testosterone use could not be determined. Fourth, the demographic composition of the CDM population changed over time. Fifth, the reasons for the decrease in testosterone prescriptions cannot be determined. A recent study within the Veterans Administration system reported a 40% decrease in testosterone prescriptions from 2013 to 2016.⁵ Given the debate that has surrounded this issue,⁶ continued monitoring of testosterone prescribing trends will be important.

Section Editor: Jody W. Zylke, MD, Deputy Editor.