Hypertension
Conditions
Keywords
Hypertension, Cardiovascular outcomes, Pragmatic clinical trial, Point of care research
Brief summary
The Diuretic Comparison Project aimed to evaluate whether chlorthalidone, as compared with hydrochlorothiazide, would reduce the risk of major nonfatal cardiovascular disease outcomes and non-cancer-related deaths in older patients with hypertension who were receiving hydrochlorothiazide at baseline. The investigators incorporated the pragmatic methods used by the Department of Veterans Affairs (VA) Healthcare System to provide a real-world assessment of the effectiveness of chlorthalidone as compared with hydrochlorothiazide in routine clinical care.
Detailed description
Thiazide-type diuretics have been in use for more than 50 years and are considered as the first-line treatment for hypertension. Of the more than 1 million Veterans prescribed a thiazide-type diuretic each year, more than 95% receive hydrochlorothiazide, and fewer than 2.5% receive chlorthalidone. However, indirect evidence has been accumulating for many years that chlorthalidone may be more effective than hydrochlorothiazide at preventing cardiovascular events. The Diuretic Comparison Project was conducted with a clinically integrated design (termed a point of care or pragmatic embedded trial). The key feature of our design was that, instead of employing local investigators, the investigators developed centralized trial procedures and implemented into the VA electronic health record (EHR) and healthcare delivery systems. This approach enabled us to conduct trial-related interactions and randomization with the use of data in the EHRs, centralize recruitment efforts without the use of site staff, eliminate the need for trial-related visits and procedures, and centralize data capture from administrative databases. This study was performed to answer a question of whether chlorthalidone is more effective than hydrochlorothiazide at preventing cardiovascular outcomes. Consent was obtained from eligible participants who were willing to participate and their primary care providers. With provider assented to the patient undergoing randomization, 3,523 older patients with hypertension were randomized across 72 VA health care systems. Patients across the US (including Puerto Rico and District of Columbia) were enrolled. Participants were randomly assigned to continue with existing hydrochlorothiazide treatment regimen (25/50mg daily) or switch to a dose equivalent chlorthalidone (12.5/25 mg daily). The primary outcome was the first occurrence of a composite outcome consisting of a nonfatal cardiovascular event or non-cancer related death. Nonfatal cardiovascular events included nonfatal myocardial infarction, stroke, hospitalization for heart failure, or urgent coronary revascularization for unstable angina.
Interventions
Trial activities following study randomization were considered as usual care. Prescriptions of the allocated drug were managed by the primary care providers and patients would receive the prescribed mediations through the VA outpatient pharmacy services.
DRUGChlorthalidone (CTD)
Trial activities following study randomization were considered as usual care. Prescriptions of the allocated drug were managed by the primary care providers and patients would receive the prescribed mediations through the VA outpatient pharmacy services.
Sponsors
VA Office of Research and Development
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
TREATMENT
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
65 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* over age 65 years * receiving hydrochlorothiazide from the VA pharmacy at a daily dose of 25 or 50 mg * most recent SBP ≥120 mm Hg and no records of SBP \<120 mm Hg in the past 90 days
Exclusion criteria
* impaired decision-making capacity rendering the patient unable to provide informed consent (i.e., if there is any question during the nurse's EHR chart review that the individual does not have the ability to make an autonomous decision or the PCP declines permission to randomize) * death expected within 6 months (inferred by PCP permission to randomize) * blood potassium level \<3.1 or 3.5 (if taking digoxin) meq/L in the past 90 days * blood sodium level \<130 meq/L in the past 90 days * enrolled in Medicare Part C (This exclusion only applied if there might be insufficient EHR data available to the study team)
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Time From Randomization to Composite Primary Outcome | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Time to study primary outcome was defined as years from randomization to the first occurrence of a composite endpoint, consisting of a nonfatal cardiovascular event or non-cancer related death. Nonfatal cardiovascular events included nonfatal myocardial infarction, stroke, hospitalization for heart failure, or urgent coronary revascularization for unstable angina. For participants who had a primary outcome, time to event was determined as the earliest admission or death date. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
| Proportion of Participants Had a Composite Primary Outcome | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | The primary outcome was the first occurrence of a composite endpoint consisting of a nonfatal cardiovascular event or non-cancer related death. Nonfatal cardiovascular events included nonfatal myocardial infarction, stroke, hospitalization for heart failure, or urgent coronary revascularization for unstable angina. Time to the first event was computed based on the earliest hospital admission or death dates. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Proportion of Participants Had Hospitalization for Heart Failure | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
| Proportion of Participants Had Nonfatal Myocardial Infarction | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
| Proportion of Participants Deceased and Not Related to Cancer | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
| Proportion of Participants Had Unstable Angina Leading to Urgent Coronary Revascularization | Collection of outcome data were performed from randomization until participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
| Proportion of Participants Had Nonfatal Stroke | Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants). | Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment. |
Countries
Puerto Rico, United States
Participant flow
Recruitment details
A total of 72 VA health care systems were enlisted in the trial. A total of 6188 primary care providers were approached for participation in the trial and 4128 (69%) consented. Patients managed by the consented provider were electronically screened for study eligibility. Centralized study recruiters telephoned potential participants and obtained oral informed consent if the patient agreed to participate (n=16595). Of these, 13523 patients were randomized (started participant flow).
Pre-assignment details
After participants provided verbal informed consent, the patient's provider was sent an electronic order to sign if the provider assented to the patient undergoing study randomization. Provider participation involved agreeing to have their patients approached for trial recruitment by study staff, signing the study drug order, and managing their patient per usual care after randomization. Providers were not randomized.
Participants by arm
| Arm | Count |
|---|---|
| Hydrochlorothiazide Participants remained on the existing hydrochlorothiazide treatment regimen (a daily dose of 25 or 50 mg) | 6,767 |
| Chlorthalidone Participants switched to an equivalent dose of chlorthalidone (a daily dose of 12.5 or 25 mg) | 6,756 |
| Total | 13,523 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 | FG002 |
|---|---|---|---|---|
| Overall Study | Death | 448 | 446 | 0 |
| Overall Study | Withdrawal by Subject | 7 | 12 | 0 |
Baseline characteristics
| Characteristic | Hydrochlorothiazide | Total | Chlorthalidone |
|---|---|---|---|
| Age | 73.5 years STANDARD_DEVIATION 5.3 | 72.4 years STANDARD_DEVIATION 5.3 | 72.4 years STANDARD_DEVIATION 5.4 |
| Age, Continuous | 72.5 years STANDARD_DEVIATION 5.3 | 72.4 years STANDARD_DEVIATION 5.3 | 72.4 years STANDARD_DEVIATION 5.4 |
| Body mass index | 31.8 kg/m2 STANDARD_DEVIATION 5.9 | 31.8 kg/m2 STANDARD_DEVIATION 5.8 | 31.7 kg/m2 STANDARD_DEVIATION 5.8 |
| Current smoker | 1437 Participants | 2957 Participants | 1520 Participants |
| Estimated GFR<60 ml/min/1.73 m2 | 1547 Participants | 3097 Participants | 1550 Participants |
| Ethnicity (NIH/OMB) Hispanic or Latino | 257 Participants | 494 Participants | 237 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 6268 Participants | 12549 Participants | 6281 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 242 Participants | 480 Participants | 238 Participants |
| History of diabetes | 3062 Participants | 6029 Participants | 2967 Participants |
| History of heart failure | 526 Participants | 1051 Participants | 525 Participants |
| History of myocardial infarction | 258 Participants | 488 Participants | 230 Participants |
| History of myocardial infarction or stroke | 722 Participants | 1455 Participants | 733 Participants |
| History of stroke | 495 Participants | 1029 Participants | 534 Participants |
| Number of antihypertensive drugs prescribed | 2.6 drugs STANDARD_DEVIATION 1 | 2.6 drugs STANDARD_DEVIATION 1.1 | 2.6 drugs STANDARD_DEVIATION 1.1 |
| Race (NIH/OMB) American Indian or Alaska Native | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) Asian | 13 Participants | 28 Participants | 15 Participants |
| Race (NIH/OMB) Black or African American | 1023 Participants | 2027 Participants | 1004 Participants |
| Race (NIH/OMB) More than one race | 145 Participants | 296 Participants | 151 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 361 Participants | 718 Participants | 357 Participants |
| Race (NIH/OMB) White | 5225 Participants | 10454 Participants | 5229 Participants |
| Receiving hydrochlorothiazide at a daily dose of 25 mg | 6402 Participants | 12781 Participants | 6379 Participants |
| Region of Enrollment Puerto Rico | 79 Participants | 156 Participants | 77 Participants |
| Region of Enrollment United States | 6688 Participants | 13367 Participants | 6679 Participants |
| Resided in rural area | 3079 Participants | 6122 Participants | 3043 Participants |
| Sex: Female, Male Female | 211 Participants | 431 Participants | 220 Participants |
| Sex: Female, Male Male | 6556 Participants | 13092 Participants | 6536 Participants |
| Systolic blood pressure | 139 mm Hg STANDARD_DEVIATION 14 | 139 mm Hg STANDARD_DEVIATION 14 | 139 mm Hg STANDARD_DEVIATION 14 |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 448 / 6,767 | 446 / 6,756 |
| other Total, other adverse events | 1,229 / 6,767 | 1,397 / 6,756 |
| serious Total, serious adverse events | 1,977 / 6,767 | 1,990 / 6,756 |
Outcome results
Primary
Proportion of Participants Had a Composite Primary Outcome
The primary outcome was the first occurrence of a composite endpoint consisting of a nonfatal cardiovascular event or non-cancer related death. Nonfatal cardiovascular events included nonfatal myocardial infarction, stroke, hospitalization for heart failure, or urgent coronary revascularization for unstable angina. Time to the first event was computed based on the earliest hospital admission or death dates. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of primary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Had a Composite Primary Outcome | 675 Participants |
| Chlorthalidone | Proportion of Participants Had a Composite Primary Outcome | 702 Participants |
Primary
Time From Randomization to Composite Primary Outcome
Time to study primary outcome was defined as years from randomization to the first occurrence of a composite endpoint, consisting of a nonfatal cardiovascular event or non-cancer related death. Nonfatal cardiovascular events included nonfatal myocardial infarction, stroke, hospitalization for heart failure, or urgent coronary revascularization for unstable angina. For participants who had a primary outcome, time to event was determined as the earliest admission or death date. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of primary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| Hydrochlorothiazide | Time From Randomization to Composite Primary Outcome | 1.5 Years | Standard Deviation 1 |
| Chlorthalidone | Time From Randomization to Composite Primary Outcome | 1.5 Years | Standard Deviation 1 |
Comparison: Primary analyses were performed with the use of unadjusted log-rank tests that were stratified according to VA health care system.p-value: 0.4595% CI: [0.94, 1.16]Log Rank
Secondary
Proportion of Participants Deceased and Not Related to Cancer
Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of secondary outcome was performance according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Deceased and Not Related to Cancer | 354 Participants |
| Chlorthalidone | Proportion of Participants Deceased and Not Related to Cancer | 359 Participants |
Secondary
Proportion of Participants Had Hospitalization for Heart Failure
Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of secondary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Had Hospitalization for Heart Failure | 232 Participants |
| Chlorthalidone | Proportion of Participants Had Hospitalization for Heart Failure | 242 Participants |
Secondary
Proportion of Participants Had Nonfatal Myocardial Infarction
Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of secondary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Had Nonfatal Myocardial Infarction | 140 Participants |
| Chlorthalidone | Proportion of Participants Had Nonfatal Myocardial Infarction | 142 Participants |
Secondary
Proportion of Participants Had Nonfatal Stroke
Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Outcome data collection was performed from study randomization until the participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of secondary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Had Nonfatal Stroke | 83 Participants |
| Chlorthalidone | Proportion of Participants Had Nonfatal Stroke | 83 Participants |
Secondary
Proportion of Participants Had Unstable Angina Leading to Urgent Coronary Revascularization
Secondary outcomes were the individual components of the primary outcome. Ascertainment of study outcomes was made with the use of administrative and clinical data obtained from VA EHRs through June 1, 2022, from records of Medicare claims obtained from the Centers for Medicare and Medicaid Services through 2021, and from National Death Index records through 2019. Trial outcomes were ascertained with the use of validated EHR phenotypes and, when needed, manual adjudication. Manually adjudicated outcomes were evaluated by investigators and staff who were unaware of group assignment.
Time frame: Collection of outcome data were performed from randomization until participants deceased, withdrawn, or reached end of study (Up to 5.4 years for the first patient enrolled and an average of 2.4 years for all participants).
Population: Analysis of secondary outcome was performed according to the intention-to-treat principle.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Hydrochlorothiazide | Proportion of Participants Had Unstable Angina Leading to Urgent Coronary Revascularization | 13 Participants |
| Chlorthalidone | Proportion of Participants Had Unstable Angina Leading to Urgent Coronary Revascularization | 20 Participants |