Comparing Clinical Effectiveness and Drug Toxicity With Hydrochlorothiazide and Chlorthalidone Using Two Potency Ratios in a Managed Care Population

Abstract

This study compared the clinical effectiveness and drug toxicity of chlorthalidone and hydrochlorothiazide. Electronic health records and claims data were used to identify patients initially prescribed chlorthalidone or hydrochlorothiazide. A total of 214 patients prescribed chlorthalidone 25 mg were matched with 428 patients prescribed hydrochlorothiazide 25 mg (1:1 potency ratio) and 214 patients prescribed hydrochlorothiazide 50 mg (1:2 potency ratio). Mean systolic blood pressure/diastolic blood pressure values at least 30 days after initial prescription were lower with chlorthalidone (132.2/74 mm Hg) compared with hydrochlorothiazide 25 mg (137.0/77.5 mm Hg) and hydrochlorothiazide 50 mg (138.6/78.5 mm Hg) (P<.05 for all comparisons). Goal systolic blood pressure/diastolic blood pressure values were achieved in a higher percentage of patients prescribed chlorthalidone (45.0%/78.3%) than with either hydrochlorothiazide 25 mg (32.1%/63.9%) or hydrochlorothiazide 50 mg (32.8%/68.9%) (P<.05 for all comparisons). Mean serum potassium was 3.94 mEq/L with chlorthalidone 25 mg, 4.13 mEq/L with hydrochlorothiazide 25 mg (P<.01 vs chlorthalidone), and 3.96 mEq/L with hydrochlorothiazide 50 mg. These findings indicate that chlorthalidone 25 mg is associated with a better antihypertensive response than hydrochlorothiazide 25 mg or 50 mg, without clinically significant differences in serum potassium.

Hypertension is a major modifiable risk factor for atherosclerotic cardiovascular (CV) disease.1 Thiazide diuretics are first‐line antihypertensive agents for most patients with hypertension. Two such agents, chlorthalidone, considered a thiazide‐like, and hydrochlorothiazide, considered a thiazide‐type, have been utilized for more than 4 decades.2 Although both are available in generic formulations, prescriptions for hydrochlorothiazide dominate the US antihypertensive market, while use of chlorthalidone is less common. Reasons for this prescribing pattern are unclear but may result from concerns about greater risk for hypokalemia with chlorthalidone, the relative lack of chlorthalidone containing fixed‐dose combination products, and uncertainty among prescribers about interchangeability between these agents.3

Both hydrochlorothiazide and chlorthalidone were approved by the US Food and Drug Administration more than 50 years ago. Soon after approval, smaller trials documented comparable efficacy of chlorthalidone and hydrochlorothiazide, but at much higher doses than are currently used.4 Several years later, the study advisory board for the landmark multicenter Multiple Risk Factor Intervention Trial (MRFIT), which originally allowed patients to be treated with either hydrochlorothiazide or chlorthalidone, recommended that all patients be given chlorthalidone exclusively because of the unfavorable trend in mortality in hydrochlorothiazide‐treated patients.5 Three subsequent landmark trials, the Systolic Hypertension in the Elderly Program,6 the Verapamil in Hypertension and Atherosclerosis Study,7 and the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial,8 provided evidence solidifying the long‐term evidence‐based benefits of chlorthalidone on reducing CV events.

Although hydrochlorothiazide is more frequently utilized than chlorthalidone, there is a paucity of long‐term evidence demonstrating similar reductions in CV events with hydrochlorothiazide. A retrospective observational cohort study of MRFIT suggested that chlorthalidone reduces CV events more than hydrochlorothiazide.9 These data, consistent with the 2011 National Institute for Health and Clinical Excellence guideline recommendations for the clinical management of primary hypertension, suggest chlorthalidone as well as indapamide as a preferred diuretic for patients with hypertension.10 In addition, on a mg‐per‐mg basis, chlorthalidone is more effective in lowering blood pressure (BP) than hydrochlorothiazide.2, 11 Hence, chlorthalidone is recommended ahead of hydrochlorothiazide in patients with resistant hypertension.12

This comparative effectiveness evaluation of chlorthalidone and hydrochlorothiazide used integrated claims data and electronic medical records. Two ratios of potency for chlorthalidone vs hydrochlorothiazide were used to assess outcomes in a real‐world managed care population.

Methods

We hypothesized that clinical response with chlorthalidone and hydrochlorothiazide would not be different in the management of hypertension and would result in comparable clinical effectiveness with use of a potency adjustment that accounts for greater per‐mg potency with chlorthalidone. The study objective was to compare key clinical markers of effectiveness and drug toxicity in patients treated with chlorthalidone and hydrochlorothiazide in a managed care population.

Study Design

This was a retrospective analysis of patients diagnosed with hypertension who were prescribed chlorthalidone or hydrochlorothiazide. Patients were enrollees of a large health plan with a current membership of more than 200,000 members from 50 counties in the southern region of the United States.

The health plan's electronic health record (EHR) database was used to extract data from January 1, 2005, to December 31, 2012. Data extracted included medical coding (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD‐9‐CM]), clinical measurements, and prescribing data. These data were used to identify diagnoses, determine clinical parameters, and ascertain antihypertensive medication prescribing. The study index date for each patient was defined as the day of first prescription for chlorthalidone 25 mg, hydrochlorothiazide 25 mg, or hydrochlorothiazide 50 mg. The first prescription could have been as a single‐pill or fixed‐dose combination formulation. The studied period was 13 months after the index date. Follow‐up clinical outcome measures were evaluated during the period between the first 30 days after the index date through 13 months after the index date.

Patients were enrolled if they met predetermined study criteria. Inclusion criteria were age 18 to 89 years at the index date; at least one prescription for chlorthalidone 25 mg, hydrochlorothiazide 25 mg, or hydrochlorothiazide 50 mg; a minimum of 13 months of continuous enrollment in the health plan after the index date; high adherence with hydrochlorothiazide or chlorthalidone; and a minimum of 6 months of continuous enrollment before the index date without any prescription for chlorthalidone or hydrochlorothiazide. High adherence was defined as a calculated medication possession ratio equal to or greater than 80% during the study period. All patients were required to have at least one hypertension diagnosis (using ICD‐9‐CM code 401.xx indicating essential hypertension) any time prior to the index date, on the index date, or during the study period. Patients were excluded if any of the following criteria were present: a prescription claim for both hydrochlorothiazide and chlorthalidone during the study period, switch from hydrochlorothiazide to chlorthalidone or vice versa during the study period, switch from hydrochlorothiazide or chlorthalidone monotherapy to a fixed‐dose combination that included either hydrochlorothiazide or chlorthalidone during the study period, change in prescribed dose of either hydrochlorothiazide or chlorthalidone from the dose at the index date during the study period, or heart failure (using ICD‐9‐CM code 428.xx) during the study period. Patients with heart failure were excluded because antihypertensive drug therapies are often used in these patients for purposes beyond lowering BP. Study approval was provided by the Colorado Multiple Institutional Review Board.

Patient Cohorts

Health records and claims data were used to identify patients prescribed chlorthalidone 25 mg, hydrochlorothiazide 25 mg, and hydrochlorothiazide 50 mg who met study criteria. Patients prescribed chlorthalidone 25 mg were matched with two samples of hydrochlorothiazide patients to provide two comparisons as follows: 1:1 potency ratio with patients prescribed chlorthalidone 25 mg daily (n=214) matched with patients prescribed hydrochlorothiazide 25 mg daily (n=428); and 1:2 potency ratio with patients prescribed chlorthalidone 25 mg daily (n=214) matched with patients prescribed hydrochlorothiazide 50 mg daily (n=214).

Propensity Matching

After matching by potency, patients were further matched on propensity scores to minimize potential observable differences between different treatment groups.13 The nearest neighbor propensity score matching was developed.14 Logisitic regression was used to estimate the propensity scores using the following variables: age, sex, unique number of other hypertensive medications, presence of diabetes and kidney disease, and the chronic condition indicator (excluding hypertension).15 Variables were selected for propensity matching based on their potential to explain discrepancy in prescribing of medications. Accordingly, the binary treatment variable (chlorthalidone 25 mg daily) was regressed on age, sex, unique number of other hypertensive medications, presence of diabetes and kidney disease, and the chronic condition indicator (excluding hypertension) using logistic regression. Then, using the predicted probabilities from the logistic regression, each patient in the chlorthalidone 25 mg group was matched with 2 patients in the hydrochlorothiazide 25 mg group and 1 patient in the hydrochlorothiazide 50 mg group. The sample size for hydrochlorothiazide 50 mg was small, so 1‐to‐1 matching was utilized. The hydrochlorothiazide 25 mg sample size was large and enabled 1‐to‐2 matching.

Outcome Measurements

Clinical outcomes of effectiveness and toxicity included systolic BP (SBP), diastolic BP (DBP), serum potassium concentrations, serum glucose concentrations, and serum urate concentrations. These outcomes were collected during the follow‐up period of 12 months after the 30‐day washout period. The first occurrences of these clinical outcomes available during this time frame were used for assessment of the study objective.

Statistical Analysis

All analyses were performed using SAS 9.2 (SAS Institute Inc, Cary, NC). Two sets of analyses were developed. The first analysis (1:1 potency ratio) compared patients prescribed 25 mg of chlorthalidone with patients prescribed 25 mg of hydrochlorothiazide. The second analysis (1:2 potency ratio) compared patients prescribed 25 mg of chlorthalidone with patients prescribed 50 mg of hydrochlorothiazide.

Descriptive statistics were used to explain the cohorts. Outcomes between the two cohorts were compared with formal test statistics reported. We also compared demographic characteristics such as age, sex, race, smoking status, and insurance type as relevant characteristics between cohorts. All unadjusted analyses were performed on the matched samples.

In addition to propensity matching, adjusted regression analyses were used to examine differences in effectiveness for clinical outcomes between the two treatment groups. Separate regression analyses were developed for outcome variables. The ordinary least‐square method was used to examine the differences in continuous clinical outcomes such as SBP and DBP. Logistic regression was used to examine the effect of treatment (chlorthalidone vs hydrochlorothiazide) on the odds of goal attainment. All regressions were controlled for explanatory variables (thiazide, age, sex, race, region, plan type, chronic comorbidity index, diabetes or kidney disease, glomerular filtration rate, concomitant other hypertension medications, adherence, and number of unique hypertension medication classes at baseline). Regression analysis was used on the matched samples and included additional control variables. The advantage of using regression analysis is the ability to estimate incremental effects of all input (control) variables on the outcome variable.

Results

Overall, 3793 patients met the inclusion criteria: 214 were prescribed chlorthalidone 25 mg, 3216 were prescribed hydrochlorothiazide 25 mg, and 363 were prescribed hydrochlorothiazide 50 mg. To establish the two potency ratio groups, all 214 of the patients prescribed chlorthalidone were matched with 428 patients prescribed hydrochlorothiazide 25 mg to create the 1:1 potency ratio group. The same 214 patients prescribed chlorthalidone were also matched with 214 patients prescribed hydrochlorothiazide 50 mg to create the 1:2 potency ratio group. Baseline demographics and insurance types between chlorthalidone and hydrochlorothiazide in both potency ratio groups are shown in Table 1. The most significant differences were that in both groups, patients prescribed chlorthalidone were more likely to be prescribed combination antihypertensive therapy than patients prescribed hydrochlorothiazide. Adherence during the study period, estimated based on medication possession ratio, was 95.8, 94.5, and 94.5 for chlorthalidone 25 mg, hydrochlorothiazide 25 mg, and hydrochlorothiazide 50 mg, respectively (P>.5 for comparisons of chlorthalidone and hydrochlorothiazide in both groups).

Table 1.

Baseline Characteristics in the 1:1 Potency Ratio and 1:2 Potency Ratio Comparison Groups

Characteristic	Chlorthalidone 25 mg	Hydrochlorothiazide 25 mg	Hydrochlorothiazide 50 mg
Total cohort, No.	214	428	214
Mean age, y	58.8	59.1	58.6
Female sex, %	53.3	54.7	53.7
Male sex, %	46.7	45.3	46.3
Race, %
White	84.6	81.5	74.8
Black	10.7	8.4a	16.4
Other	4.7	10	8.9b
Insurance coverage, %
Commercial	72.0	70.1	72.9
Medicare	25.7	26.2	24.2
Both	2.3	3.7	2.8
Hypertension monotherapy, %	6.5	59.8a	29.9b
Hypertension combination pharmacotherapy, %	93.5	40.2a	70.1b
Other antihypertensive agents used, No. (%)c
ACE inhibitor	25 (33.3)	50 (31.3)	17 (20.7)
ARB	5 (6.7)	8 (5.0)	2 (2.4)
β‐Blocker	34 (45.3)	58 (36.3)	31 (37.8)
CCB	8 (10.7)	32 (20.0)	25 (30.5)
Other	3 (4.0)	12 (7.5)	7 (8.5)
Mean chronic conditions indicator	2.09	2.05	2.12
Diabetes mellitus, %	14.5	14.0	13.1
Chronic kidney disease, %	0.9	0.9	0
Mean glomerular filtration rate, mL/min/1.73 m2 d	75.1 (n=140)	75.9 (n=304)	72.8 (n=154)

Abbreviations: ACE, angiotensin‐converting enzyme; ARB, angiotensin receptor blocker; CCB, calcium channel blocker.

^a P<.05 for chlorthalidone 25 mg vs hydrochlorothiazide 25 mg.

^b P<.05 for chlorthalidone 25 mg vs hydrochlorothiazide 50 mg.

^cOnly a subset of the total number of patients in each of the study groups took other hypertensive agents.

^dComplete data needed to calculate this parameter were available for only a portion of the study population

The first BP between 30 days and 13 months after the index date for chlorthalidone or hydrochlorothiazide therapy is depicted in Table 2. The mean number of days between the index date and this BP measurement were 122.9 days for chlorthalidone 25 mg, 101.1 days for hydrochlorothiazide 25 mg, and 119.1 days for hydrochlorothiazide 50 mg. Mean systolic BP and DBP values were significantly lower in patients prescribed chlorthalidone than hydrochlorothiazide in both groups. Mean systolic BP was 4.8 mm Hg lower with chlorthalidone 25 mg than with hydrochlorothiazide 25 mm Hg. Similarly, mean systolic BP was 6.4 mm Hg lower with chlorthalidone 25 mg than with hydrochlorothiazide 50 mm Hg. Differences in mean DBP followed a similar pattern, and also demonstrated greater reductions with chlorthalidone 25 mg compared with hydrochlorothiazide 25 mg (3.5 mm Hg mean difference) or hydrochlorothiazide 50 mg (4.5 mm Hg mean difference). Patients prescribed chlorthalidone were significantly more likely to have attained their SBP or DBP goal than patients prescribed hydrochlorothiazide. Among patients prescribed chlorthalidone 25 mg, 40.6% attained both SBP and DBP goal, which was significantly higher than in patients prescribed hydrochlorothiazide 25 mg (27.0%) and hydrochlorothiazide 50 mg (28.8%).

Table 2.

Comparison of Clinical Markers of Blood Pressure

Clinical Marker	Chlorthalidone 25 mg (n=180)	Hydrochlorothiazide 25 mg (n=355)	Hydrochlorothiazide 50 mg (n=177)
Mean SBP, mm Hg	132.2	137.0a	138.6a
Mean DBP, mm Hg	74.0	77.5a	78.5a
Patients at SBP goalc	45.0%	32.1%a	32.8%b
Patients at DBP goald	78.3%	63.9%a	68.9%b
Patients at both SBP and DBP goal	40.6%	27.0a	28.8%b

Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure.

^a P<.01, vs chlorthalidone 25 mg.

^b P<.05, vs chlorthalidone 25 mg.

^cSBP goal defined as <140 mm Hg, except <130 mm Hg if diabetes.

^dDBP goal defined as <90 mm Hg, except <80 mm Hg if diabetes.

Adjusted analyses demonstrated that SBP and DBP values were lower in patients prescribed chlorthalidone 25 mg than in patients prescribed hydrochlorothiazide 25 mg or hydrochlorothiazide 50 mg (Table 3). After adjusting for explanatory variables, patients prescribed chlorthalidone 25 mg had SBP values that were 3.2 mm Hg and 4.6 mm Hg lower than in patients prescribed hydrochlorothiazide 25 mg and hydrochlorothiazide 50 mg, respectively. Patients taking chlorthalidone 25 mg also had DBP values that were 2.6 mm Hg and 3.7 mm Hg lower compared with patients receiving hydrochlorothiazide 25 mg and hydrochlorothiazide 50 mg, respectively. Adjusted analyses on goal attainment rates demonstrated that patients prescribed chlorthalidone had higher odds of attaining goal BP values. This difference trended toward statistical significance when chlorthalidone 25 mg was compared with hydrochlorothiazide 25 mg (37% and 58% higher odds for SBP and DBP, respectively) and was statistically significant when chlorthalidone 25 mg was compared with hydrochlorothiazide 50 mg (65% and 137% higher odds for SBP and DBP, respectively).

Table 3.

Adjusted Results for Clinical Markers of Effectiveness During the Follow‐Up Period

	Chlorthalidone 25 mg Compared With Hydrochlorothiazide 25 mg	Chlorthalidone 25 mg Compared With Hydrochlorothiazide 50 mg
Mean incremental mm Hg difference in SBP	−3.16 (−5.25 to −1.07)	−4.6 (−7.4 to −1.8)
Mean incremental mm Hg difference in DBP	−2.59 (−3.84 to −1.34)	−3.68 (−5.41 to −1.95)
Odds ratio of both SBP and DBP goala attainment	1.38 (0.98 to 1.94)	1.81 (1.12 to 2.94)

Values in parentheses are expressed as 95% confidence interval. Chlorthalidone 25 mg (n=180), hydrochlorothiazide 25 mg (n=355), and hydrochlorothiazide 50 mg (n=177).

^aSystolic blood pressure (SBP) goal is defined as <140 mm Hg, except <130 mm Hg if diabetes is present, and diastolic blood pressure (DBP) goal is defined as <90 mm Hg, except <80 mm Hg if diabetes is present.

Clinical markers of drug toxicity are depicted in Table 4. Patients prescribed chlorthalidone 25 mg had lower mean serum potassium concentrations compared with patients prescribed hydrochlorothiazide 25 mg. However, serum potassium values were similar with chlorthalidone 25 mg compared with hydrochlorothiazide 50 mg. Mean serum glucose values were not different among patients prescribed any of the three different thiazides. Serum urate values were available in 21 patients, and serum sodium values were available in four patients during the study time frame, therefore analyses based on these parameters were not conducted.

Table 4.

Clinical Markers of Toxicity

Clinical Marker	Chlorthalidone 25 mg	Hydrochlorothiazide 25 mg	Hydrochlorothiazide 50 mg
Mean serum potassium, mEq/L	3.94 (n=143)	4.13a (n=299)	3.96 (n=152)
Mean serum glucose, mg/dL	111.0 (n=114)	107.3 (n=237)	109.6 (n=125)

^a P<.01, vs chlorthalidone 25 mg.

Discussion

The relationship between elevated BP and CV events is well established.16 Clinical evidence definitively demonstrates benefits of reducing BP with effective antihypertensive drug therapy.17, 18 Diuretics continue to be recommended as first‐line therapy in a broad range of patients in contemporary clinical guidelines.19, 20, 21, 22 Common diuretics for the management of hypertension include thiazide‐type (eg, hydrochlorothiazide) and thiazide‐like (eg, chlorthalidone) diuretics. Many patients with hypertension who are treated with antihypertensive drug therapy may not achieve their goal BP.23 Therefore, many patients may benefit by using a more potent drug therapy option when there is a difference within a particular drug class.

Our findings indicate that treatment with chlorthalidone was associated with greater reductions in BP and higher rates of achieving goal BP values than treatment with hydrochlorothiazide. Based on the adjusted analyses, this correlated to approximately 3 mm Hg to 4 mm Hg lower SBP and DBP values with chlorthalidone compared with hydrochlorothiazide. This was seen both with a mg equivalent 1:1 potency ratio and with a 1:2 mg potency ratio that accommodates the greater antihypertensive effects with chlorthalidone. This difference is clinically meaningful. On a patient‐specific basis, the greater effectiveness with chlorthalidone could be the difference between being at BP goal and requiring additional antihypertensive therapy. Patients prescribed chlorthalidone had higher rates of attaining goal BP values. In the adjusted analyses, chlorthalidone 25 mg had 38% higher odds of attaining both SBP and DBP goals compared with hydrochlorothiazide 25 mg, and 81% higher odds of attaining both SBP and DBP goals compared with hydrochlorothiazide 50 mg.

Although chlorthalidone and hydrochlorothiazide are very well tolerated, drug toxicity related to metabolic adverse effects can manifest as hypokalemia, hyperglycemia, hyponatremia, and/or hyperuricemia. The surrogate clinical measures of drug toxicity used in this analysis (serum concentrations of potassium, glucose), demonstrated little to no difference between the two study drugs. The only surrogate clinical measure of drug toxicity that was more prevalent in patients treated with chlorthalidone compared with patients treated with hydrochlorothiazide was slightly lower mean serum potassium values with chlorthalidone compared with hydrochlorothiazide 25 mg. Importantly, this surrogate marker was similar with chlorthalidone 25 mg compared with hydrochlorothiazide 50 mg.

Our findings demonstrating greater antihypertensive effects with chlorthalidone are consistent with other reports. Research comparing antihypertensive effects of hydrochlorothiazide and chlorthalidone in trial‐like settings has demonstrated that chlorthalidone is more potent than hydrochlorothiazide on a mg‐per‐mg basis.11 These studies involved small sample sizes of fewer than 50 patients. Ernst and colleagues24 conducted a randomized 8‐week clinical trial in 30 patients comparing chlorthalidone with hydrochlorothiazide on 24‐hour ambulatory BP.24 The authors found that chlorthalidone had a greater antihypertensive effect than hydrochlorothiazide.

A 2010 meta‐analysis of the dose response between chlorthalidone and hydrochlorothiazide on SBP demonstrated that chlorthalidone produces greater reductions in SBP in the low‐dose range of 12.5 mg to 25 mg.25 On a mg‐per‐mg basis, chlorthalidone is at least 1.5 to 2 times as potent as hydrochlorothiazide based on a few small, direct comparative trials where hydrochlorothiazide was administered at twice the dose of chlorthalidone and achieved a similar lowering in SBP.2 Data from a 2012 meta‐analysis suggest an even greater difference in potency with chlorthalidone being approximately 4 times as potent as hydrochlorothiazide on a mg‐per‐mg basis.26 Many clinicians may not be aware of the differences in antihypertensive effects between chlorthalidone and hydrochlorothiazide. However, our data further emphasize this important discrepancy.

Hydrochlorothiazide is available as a 12.5‐mg capsule and as 25‐mg and 50‐mg tablets (both scored). However, chlorthalidone is available only as a 25‐mg tablet (unscored) and a 50‐mg tablet (scored). Dosage formulation availability likely influences the prescribing of these two agents. Real‐world clinicians must prescribe chlorthalidone and assume either a 1:1 or 1:2 potency ratio for chlorthalidone vs hydrochlorothiazide when comparing interchangeability of these agents. We used both of these potency ratios in this comparative effectiveness evaluation and demonstrated greater reductions with chlorthalidone using both ratios. Most of the current data comparing chlorthalidone with hydrochlorothiazide are small or are retrospective analyses of clinical trials, which are artificial environments in controlled settings. This analysis represents confirmation of findings in the real‐world setting.

The most commonly used doses of hydrochlorothiazide and chlorthalidone are 25 mg daily, although hydrochlorothiazide 50 mg daily has been shown to reduce BP to a greater extent than 25 mg.27 However, hydrochlorothiazide 50 mg is less frequently used because of potential increased risk for hypokalemia. Of note, based on meta‐analysis data, risk of hypokalemia is similar between hydrochlorothiazide and chlorthalidone when adjusting for the different mg‐per‐mg potency.25 However, no direct comparisons have confirmed this finding. Our data demonstrated that lower serum potassium values were associated with chlorthalidone 25 mg than with hydrochlorothiazide 25 mg. However, this difference was small, was not likely clinically significant, and was only seen with the 1:1 potency ratio comparison.

Study Limitations

There are several limitations to our data. This was a retrospective analysis of one EHR database and may not be reproducible or generalizable to all populations. In our database, medication use was defined based on EHR documented prescribing, and BP values were based on what was entered into the EHR. Accuracy of BP measurements that are documented in the EHR cannot be assured. The time that BP was reassessed after achievement of goal BP was not systematically evaluated in all groups. There is a possibility that variability in the time of these assessments could have influenced the BP measurement. While we had a reasonable sample size, it is possible that we may have lacked the power to see differences in some of the study endpoints. Although our propensity matching incorporated several variables, the dose of other concurrent antihypertensive medications was not able to be included. Race can influence the antihypertensive response of hydrochlorothiazide and chlorthalidone, especially when used as monotherapy. For example, black patients may have larger reductions in BP with thiazide diuretics than nonblack patients. It was not possible to capture the race of patients included in this study within the EHR and claims database that were used. Therefore, it is possible that differences in race could have influenced our results. These limitations are typical for these types of database analyses, but nonetheless, our data represent real‐life comparative effectiveness.

Combination antihypertensive drug therapy was prescribed commonly in our analysis, especially with chlorthalidone. Considering that these are real‐world data, patients prescribed chlorthalidone may possibly have had more severe hypertension, and this possibility is reinforced by the preferential use of chlorthalidone in the treatment of patients with resistant hypertension. Differences in severity of disease could explain some of the differences in BP or goal attainment rates seen in our comparisons. However, because this study is a retrospective cohort analysis, possible differences in disease severity do not refute the overall findings that prescribing of chlorthalidone was associated with better markers of antihypertensive treatment (BP values and attainment of goal BP values) than hydrochlorothiazide.

Conclusions

Our data found that greater reductions in BP and higher BP goal attainment rates were associated with chlorthalidone treatment than with hydrochlorothiazide. This association was seen when chlorthalidone 25 mg was compared with hydrochlorothiazide 25 mg and even with hydrochlorothiazide 50 mg. Although chlorthalidone was associated with more antihypertensive effectiveness, it was not associated with a clinically significant change in the common diuretic‐associated adverse effect of serum potassium depletion. Clinicians can be confident with the expected antihypertensive effectiveness when prescribing chlorthalidone instead of hydrochlorothiazide in patients with hypertension.

Disclosures

All authors do not have any conflicts of interest. Takeda Pharmaceutical Company Limited provided financial support for this research.

J Clin Hypertens(Greenwich). 2015;17:134–140. DOI: 10.1111/jch.12453. © 2014 Wiley Periodicals, Inc.