Comparing Outcomes Between Thiazide Diuretics and Other First-line Antihypertensive Drugs in Long-term Nursing Home Residents

Purpose: Hypertension occurs in >50% of US nursing home (NH) residents, but it is unclear which antihypertensive classes offer the best balance of benefits and risks in this population. The objectives of this study were to describe the patterns of antihypertensive medication treatment in this population, focusing on thiazide diuretics, and to determine the association between thiazide diuretics (DIURs) and outcomes important to NH patients. Methods: This observational cohort study was conducted in long-term NH residents treated for hypertension in the second quarter (Q2) of 2013, from all US NHs. The primary exposure was the frequency of use of antihypertensive treatment class (DIURs, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers [ARBs], calcium channel blockers, and β-blockers) according to Medicare Part D dispensing data. Because DIUR-related urinary symptoms were a focus, residents receiving nonthiazide diuretics were excluded. We ascertained continued medication use by class from Q2 to Q4 of 2013, and ascertained 6-month incontinence and hospitalization using data from Medicare claims and the Minimum Data Set. Findings: Of 152,902 NH residents treated for hypertension, 52.2% were treated with β-blockers (22% as a single agent), 39.7% with calcium channel blockers (14% as a single agent), 38.8% with angiotensin-converting enzyme inhibitors (14% as a single agent), 14.2% with DIURs (2% as a single agent), and 13.2% with ARBs (4% as a single agent). Overall, 55.1% were treated with 1 drug; 33.2%, with 2 drugs; and 11.8%, with 3 or more drugs. From Q2 to Q4, DIURs were more likely to have been discontinued than any other class (19.4% vs 14.1%–16.1% for each of the other 4 classes; all, p < 0.05) and less likely to have been started than any other class except ARBs (1.4% vs 3.8%–5.3% for each of the other 3 classes). Urinary incontinence occurred in 76.6% of the sample. In a multivariate logistic regression model, new DIUR use from Q2 to Q4 of 2013 was not significantly associated with urinary incontinence in Q4, and none of the antihypertensive drug classes were associated with 6-month hospitalization. Implications: In 2013, long-term NH residents treated for hypertension were least likely to receive, more likely to discontinue, and less likely to start a new DIUR than any other first-line antihypertensive medication. DIURs were not associated with increased incontinence or hospitalization, so in the absence of indications for other drugs, DIURs may be a reasonable first-line choice for hypertension treatment in this population.