Comparative risk of renal, cardiovascular, and mortality outcomes in controlled, uncontrolled resistant, and non-resistant hypertension

Cohort Characteristics

A total of 470,386 individuals were identified for the study cohort (Figure 1) as previously described (9). The mean age (SD) was 65 (11) years with 45% males (Table 1). The race/ethnic composition was 43% white, 21% Hispanic, 13% black, and 8% Asian. The mean blood pressure of the entire study cohort was 133/75 mmHg. Resistant hypertension was identified in 60,327 (12.8%) individuals of whom 23,104 met our criteria for cRH and 37,223 met our criteria for uRH. Thus, uRH accounted for 61.7% of the resistant hypertension population and 7.9% of all hypertensive individuals.

Compared to the non-RH population, the RH population had a greater prevalence of comorbid conditions including DM (48% vs 30%), CKD (45% vs 24%), ischemic heart disease (41% vs 22%), and cerebrovascular disease (16% vs 9%) (P<0.001 for all). cRH and uRH were similar in age, BMI, race/ethnicity composition and comorbidities. Blood pressures were highest in uRH (154/79 mmHg) compared to both cRH (123/67 mmHg) and non-RH (132/75 mmHg).

Antihypertensive Drug Usage

Diuretics were the most frequently prescribed class of medications in all 3 groups (Figure 2). Individuals with cRH were more likely to be prescribed an angiotensin converting enzyme inhibitor or angiotensin receptor blocker (93%) compared to the non-RH (47%) and uRH populations (88%).

Outcomes

Overall 114,364 events occurred amongst the cohort and 18.5% experienced at least one event (Table 2). Mortality occurred in 43,580 (9.3%) individuals. In addition, there were 26,894 IHE, 12,306 CHF, 16,799 CVA, and 14,785 incident ESRD events (Table 2). The resistant hypertension population had greater proportions of those who reached any of the outcomes (31.0 vs 16.7%) compared to the non-RH population (P<0.001). The total number of all outcomes combined for each group was 87,217 in non-RH, 12,039 in cRH, and 15,218 in uRH. The rate of any outcome was highest in cRH (34.8%) followed by uRH (28.7%) and non-RH (16.7%).

After adjustment for age, gender, race, BMI, CKD, Charlson comorbidity index, and the comorbidities of DM, ischemic heart disease, CHF, CKD, and cerebrovascular disease, the resistant hypertension (cRH and uRH) population compared to non-RH population, had HRs (95% CI) of 1.24 (1.20–1.28), 1.46 (1.40–1.52), 1.14 (1.10–1.19), 1.32 (1.27–1.37), and 1.06 (1.03–1.08) for IHE, CHF, CVA, ESRD, and all-cause mortality, respectively (Table 3, Figure 3). Unadjusted event free survival was superior for non-RH compared to resistant hypertension for all the outcomes studied (Figure 4).

Controlled Resistant Hypertension (cRH) vs Non-Resistant Hypertension (non-RH)

The multivariable adjusted HRs (95% CI) for individuals with cRH compared to their counterparts with non-RH were 1.21 (1.16–1.26), 1.51 (1.43–1.59), 1.01(0.95–1.07), 1.16 (1.10–1.22), and 1.05 (1.02–1.09 for IHE, CHF, CVA, ESRD, and all-cause mortality, respectively (Table 3, Figure 3).

Uncontrolled Resistant Hypertension (uRH) vs Non-Resistant Hypertension (non-RH)

In both crude and adjusted models, individuals with uRH were at greater risk for all outcomes compared to non-RH (Table 3, Figure 3). Multivariable adjusted HRs (95% CIs) in those with uRH vs. those with non-RH were 1.26 (1.21–1.31), 1.42(1.35–1.50), 1.24(1.18–1.30), 1.45 (1.39–1.52), and 1.06 (1.03–1.09) for IHE, CHF, CVA, ESRD, and all-cause mortality, respectively (Table 3, Figure 3).

Uncontrolled Resistant Hypertension (uRH) vs Controlled Resistant Hypertension (cRH)

Compared to cRH, individuals with uRH were at greater risk for CVA (HR, 1.23; 95% CI, 1.14–1.31) and ESRD (HR, 1.25; 95% CI, 1.18–1.33). Risk for IHE (HR, 1.04; 95% CI, 0.99–1.10), CHF (HR, 0.95; 95% CI, 0.89–1.01) and all-cause mortality (HR, 1.01; 95% CI, 0.97–1.05), were similar between those with uRH and cRH (Table 3, Figure 3).

Intragroup Comparisons within non-RH, cRH, and uRH

Multivariable adjusted HR’s for ESRD demonstrated similar trends within non-RH, cRH, and uRH (Table 4). Males, Hispanics, DM, ischemic heart disease, and cerebrovascular disease were all associated with higher risk for ESRD. Similar to past observations, older age (>/=60 years) and BMI>/=30 were associated with paradoxically lower risk for ESRD. Older age increased risk for mortality. For IHE, CHF, and CVA outcomes, older age, male gender, CKD, DM, pre-existing ischemic heart disease and cerebrovascular disease were associated with increased HR’s within all 3 hypertension cohorts (Appendix Table 2).

The non-RH population had the strongest association with medication use and outcomes. Within non-RH, use of both ACEI and thiazide type diuretics were associated with lower risk for IHE, CHF, CVA, and mortality (figure 5, Appendix Table 3). Use of ACEI, ARB, and thiazide type diuretics were all associated with lower mortality risk within all 3 cohorts.

Potential Limitations

There are several potential limitations to our study that may confound the findings. We identified the study cohort from a 2 year window and followed individuals up to 5 years. In contrast to real world phenomenon where blood pressures vary over time, we somewhat arbitrarily identified a cohort as resistant hypertension. The assumption was that once individuals were identified as resistant hypertension that they would remain so throughout the observation period. Conversely, those with non-RH may have gone on to develop resistant hypertension. We also did not have information on the duration of hypertension. Due to our definition, the resistant hypertension population would have had to had some duration of follow up to have medicine titrated to at least 3 medicines introducing an immortal time bias. Non-RH individuals could have been identified based on only 2 visits. Therefore, the resistant hypertension population was more likely to have pre-existing vascular damage that may not have been identified at the start of the observation period compared with the non-RH population. By studying hypertension categories, we are somewhat trivializing the impact of blood pressure on our outcomes studied since the step wise risk of cardiovascular outcomes with higher blood pressures has been well described (45, 46). In addition, we cannot account for heterogeneous practice patterns. Certain patients may be deemed sicker by clinicians and received more comprehensive care such as in our cRH population. They may have been seen more frequently and also had better health maintenance care. The fact that there was only a modest (6%) mortality risk in the resistant hypertension population compared to the general hypertension population may further suggest this confounding.

Conclusion

Our comparative outcomes demonstrated that the resistant hypertension population was at greater risk for IHE, CHF, CVA, ESRD, and mortality. The uRH subgroup of resistant hypertension appears to demonstrate the greatest risk particularly for CVA and ESRD. The higher risks of resistant hypertension compared to non-RH for IHE and for CHF appear to remain non-modifiable, whereas the higher risk of CVA and ESRD were reduced in our observational study in those with cRH. Our findings underscore the hypothesis that resistant hypertension has a more adverse physiology and deserves better understanding in order to better manage this population.

Study Population

A retrospective, longitudinal cohort study of Kaiser Permanente Southern California (KPSC) members was performed between January 1, 2006 and December 31, 2010. The KPSC health care system is a prepaid integrated health plan providing comprehensive care to over 3.6 million members throughout Southern California, from Bakersfield to San Diego at 13 medical centers and over 200 satellite clinics. As of December 2010, there were over 2.4 million adult members within KPSC. The patient population is ethnically and socioeconomically diverse, reflecting the general population of Southern California (47). All KPSC members have similar benefits and access to health care services, clinic visits, procedures, and copays for medications. Complete health care encounters are tracked using a common EHR from which all study information was extracted.

Many of the details of the study population has been previously described (9). Briefly, individuals age 18 years and older with hypertension and a documented blood pressure measurement were identified in the time period between January 1, 2006 and December 31, 2007. Individuals were followed until they experienced any outcome or until the end of observation (December 31, 2010). Hypertension was identified by inpatient and outpatient International Classification of Diseases, Ninth Revision (ICD-9) codes specific to hypertension (401.xx, 402.xx, 403.xx, 404.xx, 405.xx). Inclusion into the study cohort required a minimum of 2 separately dated ICD-9 codes for hypertension. The accuracy of ICD-9 coding for the diagnosis of hypertension has been previously validated (48). The date of the second ICD-9 hypertension code was used as the index date. Blood pressure values closest in date to the index date were used. In those encounters with multiple blood pressure measurements, the lowest value was used for analysis to minimize white coat hypertension. Individuals who did not have a blood pressure measurement or those who were diagnosed with secondary hypertension (renovascular disease, adrenal disorders, Cushing’s syndrome, aortic coarctation, and secondary hypertension not specified) were excluded from the study cohort. Sleep apnea was not excluded as it is often coexistent with hypertension and not necessarily a causative factor.

Data Collection and Laboratory Measurements

All laboratory data, vital sign assessments (including blood pressure measurements), and diagnostic and procedure codes are collected in the EHR as part of routine clinical care encounters. Comorbidities, including diabetes mellitus (DM), ischemic heart disease, congestive heart failure (CHF), and cerebrovascular disease, were assessed based on inpatient and outpatient ICD-9 diagnoses coding. The Deyo adaption of the Charlson Comorbidity Index was also determined using ICD-9 diagnosis codes from inpatient and outpatient encounters as an overall measure of disease burden (49). Chronic kidney disease (CKD) was defined as an estimated glomerular filtration rate of less than 60 mL/min per 1.73m² estimated from serum creatinine levels using the Chronic Kidney Disease Epidemiology Collaboration Equation (50). Data on hospitalizations and diagnoses that occurred outside the healthcare system were available through administrative claims records.

Assessment of Medication Use

Antihypertensive medication data were retrieved from the KPSC pharmacy dispensing records (9). Prescription orders, pharmacy fills, and refills are tracked for KPSC members with pharmacy benefits. Individuals were determined to be on an antihypertensive medication if it was prescribed and dispensed for 7 or more days supply within the observation period at any time on or after the initial diagnosis of hypertension. Medications that were prescribed and filled for less than 7 days were not considered.

Antihypertensive medication classes included angiotensin-converting enzyme inhibitors (ACEI), alpha blockers, angiotensin receptor blockers (ARB), beta blockers, dihydropyridine and nondihydropyridine calcium-channel blockers, central acting agents, thiazide and loop type diuretics, potassium sparing diuretics, mineralocorticoid receptor antagonists, centrally acting alpha agonists, and direct renin inhibitors. Single pill combination drugs were classified into their different respective components. The total number of blood pressure medications was defined by the number of different antihypertensive medications taken by each subject and may have included multiple medications from the same drug class.

KPSC Hypertension Treatment Guideline

KPSC has a standardized hypertension management program which includes continuous processes to standardize blood pressure measurements. KPSC publishes and advocates an internally derived hypertension algorithm which is followed by the majority of physicians (26, 34). For patients with systolic blood pressure 140–159 and diastolic blood pressure 90–99 mmHg, the guideline suggests initiating treatment with either a diuretic alone or an angiotensin converting enzyme inhibitor (ACEI)/diuretic combination. The ACEI/diuretic combination was suggested for all patients with systolic blood pressure ≥160 and/or diastolic blood pressure ≥100 mmHg. If blood pressure was still not in control with full doses, the third medication was a beta blocker to full dose. If blood pressure was still not controlled, the fourth medication was a calcium channel blocker.

Definitions of Resistant Hypertension and Sub Categories

Individuals were classified as having resistant hypertension if their systolic blood pressure was ≥140mmHg and/or their diastolic blood pressure was ≥90mmHg while prescribed 3 or more different antihypertensive medications concomitantly; or prescribed 4 or more medications concomitantly regardless of blood pressure control. All other hypertensive individuals were categorized as non-resistant hypertension (non-RH). These criteria were based on based on the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC), the American Heart Association (27, 28), and the European Society of Hypertension/European Society of Cardiology (29) guidelines. Resistant hypertension individuals who were on 4 or more antihypertensive medicines with controlled blood pressure (<140/90 mmHg) were classified as controlled resistant hypertension (cRH). Among those with resistant hypertension, individuals were classified as having uncontrolled resistant hypertension (uRH) if their systolic blood pressure was ≥140mmHg or their diastolic blood pressure was ≥90mmHg while on 3 or more antihypertensive medications.

Outcomes

The primary outcomes evaluated were ischemic heart event (IHE), congestive heart failure (CHF), cerebrovascular accident (CVA), incident end-stage renal disease (ESRD), and all-cause mortality as separate competing outcomes. Any hospitalization with the primary or secondary diagnoses of IHE and CHF were used to identify these outcomes. IHE was defined as any code for myocardial infarction, angina, coronary occlusion, myocardial necrosis, and any procedure codes for percutaneous intervention or coronary artery bypass grafting (Appendix Table 1). CHF was defined as heart failure and cardiomyopathy. CVA was defined as stroke, central nervous system bleed, or cerebrovascular disease/accident not otherwise specified. ESRD was defined as treatment with dialysis or renal transplantation. The ESRD population is reliably captured within an internal KPSC database established primarily for billing purposes. Mortality information for the cohort was obtained from several complementary sources, including KPSC administrative sources, state vital statistics, and Social Security Administration death files. For the latter two sources, a probabilistic match was made based on name, address, birth date, Social Security Number (when available) and other demographic information. Because data from these latter sources may be delayed, the date of December 31, 2010 was used to censor follow-up. Each event was followed separately without competing outcomes. Thus, individuals were followed until the occurrence of that particular event, disenrollment from the health plan, end of the study period (December 31, 2010), or death. Follow-up was not censored when another event occurred with the exception of death.

Statistical Analyses

The demographic characteristics and comorbidities in cRH, uRH, and non-RH were evaluated and compared. χ² test was used for categorical variables and the nonparametric Kruskal-Wallis test for continuous variables.

The primary analysis was to compare risk for IHE, CHF, CVA, ESRD, and all-cause mortality among non-RH, RH, cRH, and uRH. Event rates and Kaplan-Meier survival curves were constructed to describe event free survival in the 4 groups. Cox proportional hazards regression modeling was used to estimate hazard ratios (HR) for each outcome separately. Multivariable HRs were calculated with adjustment for potential confounders including age group (18–59 or >=60), sex, race/ethnicity (black or non-black), body mass index (BMI, <30 or >=30), Charlson comorbidity index (CCI), and pre-existing comorbidities including DM, CKD, ischemic heart disease, CHF, and cerebrovascular disease. The multivariable Cox models were constructed to estimate HRs for four pairwise comparisons: 1) resistant hypertension vs. non-RH; 2) cRH vs non-RH; 3) uRH vs. non-RH; and 4) uRH vs cRH.

Intragroup comparisons were made within non-RH, cRH, and uRH to evaluate risk for outcomes among different age group (18–59 or >=60), gender, BMI (<30 or >=30), race/ethnicity, and co-morbidities. The effects of different medications on IHE, CHF, CVA, ESRD, and all-cause mortality were also assessed. Within non-RH, cRH, and uRH, multivariable HR’s were calculated for each outcome based on use of specific medication classes.

All statistical analyses were generated using SAS statistical software (version 9.2; SAS Institute, Inc., Cary, NC). Results with P< 0.05 were considered statistically significant. The study protocol was reviewed and approved by the KPSC Institutional Review Board and was exempt from informed consent.