Blood Pressure Responses to Hypertension Treatment and Trends in Cognitive Function in Patients With Initially Difficult‐to‐Treat Hypertension: A Retrospective Subgroup Analysis of the Observational Study on Cognitive Function and SBP Reduction (OSCAR) Study

Abstract

J Clin Hypertens (Greenwich). 2012;14:78–84. ©2012 Wiley Periodicals, Inc.

The Observational Study on Cognitive Function and SBP Reduction (OSCAR) provided opportunities to examine the influence of eprosartan on trends in cognitive performance in a large population of patients with difficult‐to‐treat hypertension (DTTH). A total of 4649 patients diagnosed retrospectively with DTTH, defined as systolic/diastolic blood pressure (SBP/DBP) ≥140/90 mm Hg despite use of at least 3 antihypertensive drugs during the month preceding the baseline visit comprised the intention‐to‐treat (ITT) cohort. The patients were given eprosartan‐based antihypertension therapy (EBT; 600 mg/d). Blood pressure and cognitive function parameters included significant (P<.001) differences for DTTH vs non‐DTTH patients such as older age, body mass index, SBP and pulse pressure (PP), and lower Mini‐Mental State Examination (MMSE) score. After EBT for 6 months, SBP/DBP in DTTH was 138.8±12.2/81.9±7.4 (ΔSBP−26±15.7; ΔDBP−11.4±9.8); PP was 57.0±10.8 (ΔPP−14.5±13.8) (all P<.001 vs baseline and non‐DTTH group). A total of 2576 patients (87.4%) responded to EBT (ie, SBP <140 mm Hg and/or ΔSBP ≥15 mm Hg, or DBP <90 mm Hg and/or ΔDBP ≥10 mm Hg); 1426 DTTH patients (48.4%) achieved normalized SBP/DBP (ie, SBP <140 mm Hg and DBP <90 mm Hg). ΔPP in DTTH‐isolated systolic hypertension (ISH) was −18.0±13.3 mm Hg (P=.003 vs DTTH‐systolic‐diastolic hypertension). End‐of‐EBT mean MMSE was 27.5±3.0 (P<.001 vs baseline). Blood pressure responses after EBT coincided with stabilization/improvement of MMSE in this retrospective investigation in DTTH patients. The average improvement in MMSE in DTTH patients was similar to that in non‐DTTH patients. EBT effects on PP may be relevant to the evolution of MMSE in DTTH‐ISH patients.

The Observational Study on Cognitive Function and SBP Reduction (OSCAR) was designed to examine the safety and tolerability of once‐daily therapy with eprosartan in a very large community‐dwelling population of patients with arterial hypertension recruited in 28 countries and managed in routine primary care. This study provided opportunities to examine the influence of the angiotensin receptor blocker (ARB) eprosartan on trends in cognitive performance in a large population of patients with high blood pressure (BP).

The principal findings of OSCAR have been reported previously. ¹ A 6‐month period of antihypertensive therapy based on eprosartan was associated with a significant reduction in mean systolic BP (SBP) and a significant improvement in mean Mini‐Mental State Examination (MMSE) score (P<.0001 for both outcomes). In multiple linear regression, cognitive decline was demonstrated to be independently and inversely correlated with SBP reduction (odds ratio, 0.77; 95% confidence interval [CI], 0.73–0.82).

The intention‐to‐treat (ITT) cohort of OSCAR comprised 25,745 adult patients with hypertension. This large number of patients enabled us to identify several large subgroups for further investigation, with the intention of adding to what are still sometimes limited epidemiologic data on specific types of patients.

Recent revisions to extant European hypertension guidelines and first reports of new treatment options for this condition ² , ³ , ⁴ are indications that resistant hypertension is a pressing clinical priority. Poorly controlled high BP may be a factor in cognitive decline in older patients. Observations from several thousand patients in a real‐world survey may be instructive for the better management of this condition. Accordingly, we report here our findings in a retrospectively identified cohort of almost 3000 patients from the OSCAR population classified as having difficult‐to‐treat hypertension (DTTH) at baseline.

Methods

The rationale, design, and methodology of OSCAR have been reported in detail. ⁵ This 6‐month observational study was conducted in 28 countries (Australia, Bahrain, Belgium, Belorussia, Canada, France, Germany, Greece, Hong Kong [PRC], Hungary, Iran, Ireland, Jordan, South Korea, Kuwait, Lebanon, Malta, The Netherlands, Pakistan, The Philippines, Poland, Qatar, Russia, Spain, South Africa, Saudi Arabia, Thailand, and the United Arab Emirates).

The study population comprised men and women 50 years and older with newly diagnosed hypertension (SBP ≥140 mm Hg) who were eligible for treatment with eprosartan 600 mg/d once daily by virtue of (1) being previously untreated, (2) having an insufficient response to previously prescribed antihypertensive drugs, or (3) being intolerant to previously prescribed antihypertensive drugs.

For the present analysis, patients were classified as having DTTH at baseline if they had SBP >140 mm Hg and/or diastolic BP (DBP) >90 mm Hg at baseline despite having been prescribed at least 3 antihypertensive medications during the 4 weeks immediately preceding the baseline visit.

Eprosartan‐based therapy (EBT) was initiated at 600 mg/d, with provision for additional medication to be introduced after 1 month, at the discretion of individual investigators.

Data for each patient were collected at baseline, 1 to 3 months post‐baseline, and at 6 months, using a bespoke standardized record form. Other visits were permitted during this period as required by the attending physician’s usual schedule for newly diagnosed patients with hypertension.

The Mini‐Mental State Examination (MMSE) ⁶ was used for the assessment of cognitive function. Instructions and information relating to the conduct of the MMSE were given to all participating practitioners. The test itself was performed at baseline and at the end of the study using validated local‐language editions as appropriate to each country.

The primary efficacy parameters of OSCAR were the mean absolute change from baseline in SBP and the mean absolute change from baseline in MMSE score, defined as the end‐of‐study score minus the baseline score. Adverse events and suspected adverse drug reactions were recorded during follow‐up and characterized according to their severity and relation to study medication.

Statistical Considerations

The present subgroup analysis used the same core statistical principles and methods as the overall study. Thus, analyses were based on the ITT population, which comprised all patients who received at least one dose of study treatment and who provided eligible relevant values at baseline and from at least one follow‐up visit. The change in MMSE score between the baseline and post‐baseline visits was examined by means of a paired‐difference t test; a similar test was used to analyze absolute changes in BP during the study.

Nominal qualitative variables were analyzed using the χ² test, ordinal qualitative variables were analyzed using the Wilcoxon test or the Mantel–Haenzel test, and quantitative variables were analyzed by analysis of variance (ANOVA). MMSE scores and BP variables were compared between visits using covariance analysis, with baseline value as the adjusted variable. The results are presented as means with standard deviations (SDs), with the exception of results derived from ANOVA, where standard errors of the mean were calculated.

All statistical tests were performed using SAS 9.1.3 software (SAS Institute, Cary, NC).

Results

Demographics and Baseline Characteristics

A total of 4649 patients met the definition of DTTH at baseline: the determination of a subgroup ITT population of 2948 DTTH patients is illustrated in Figure 1. Baseline characteristics of the DTTH‐ITT population are summarized in Table I, where they are compared with data from the residuum of the overall ITT cohort (non‐DTTH group). DTTH patients tended to be older and heavier and to have a higher body mass index than their non‐DTTH peers. Indicators of adverse cardiovascular status were more widespread among the DTTH patients, as was a diagnosis of diabetes (Table I). Among countries with large national cohorts, Poland, Hungary, Germany, and Belgium made proportionately large contributions to the DTTH population; Greece, Australia, and France contributed proportionately fewer DTTH patients (Table II).

Figure 1.

 Derivation of patients with difficult‐to‐treat hypertension (DTTH) at baseline in the Observational Study on Cognitive Function and SBP Reduction (OSCAR). MMSE indicates Mini‐Mental State Examination.

Table I.

 Baseline Characteristics of the Resistant Hypertension ITT (DTTH‐ITT) and Residuum ITT Populations

	DTTH‐ITT Cohort (n=2948)	Non–DTTH‐ITT Cohort (n=22,797)
Age, mean (SD), y	66.9 (9.5)a	63.8 (9.4)
Men/women	1460 (49.8)/1474 (50.2)	11,550 (50.8)/11,187 (49.2)
Body mass index, mean (SD), kg/m2	28.8 (4.6)a	27.6 (4.0)
Systolic blood pressure, mean (SD), mm Hg	164.8 (14.5)a	161.5 (12.7)
Diastolic blood pressure, mean (SD), mm Hg	93.3 (9.8)	93.0 (8.7)
Pulse pressure, mean (SD), mm Hg	71.5 (14.2)a	68.5 (12.6)
Isolated systolic hypertension	827 (28.1)	5843 (25.9)
Systolic‐diastolic hypertension	2108 (71.5)	16,608 (73.7)
Mean MMSE score,b mean (SD)	26.5 (3.6)a	27.1 (3.3)
Tobacco use
Smokers	487 (17.8)	4576 (21.4)
Ex‐smokers	815 (29.8)	5662 (26.4)
Nonsmokers	1429 (52.3)	11,185 (52.2)
Hypercholesterolemia	2189 (74.8)	13,739 (60.6)
Diabetes mellitus	1171 (40.1)	4833 (21.3)
Diabetic nephropathy	255 (8.8)	576 (2.5)
Myocardial infarction	479 (16.5)	869 (3.8)
Coronary artery disease	1149 (39.5)	2533 (11.2)
Congestive heart failure	297 (10.3)	514 (2.3)
Elevated plasma creatinine	466 (16.0)	1275 (5.6)
Transient ischemic attack	414 (14.3)	1442 (6.4)

Abbreviations: DTTH, difficult‐to‐treat hypertension; ITT, intention to treat; MMSE, Mini‐Mental State Examination; SD, standard deviation. ^a P<.001 vs residuum. ^bMaximum possible score=30. Values are expressed as number (percent) unless otherwise indicated.

Table II.

 Distribution of the Resistant Hypertension (DTTH) Population of OSCAR Among Participating Countries^a

Country (Total No. of Patients)	DTTH‐ITT Cohort, No. (n=2948)	DTTH in National Cohort, No.	Contribution to Total DTTH Cohort (% of 2948)
Australia (1899)	123	6.5	4.17
Bahrain (17)	4	23.5	0.14
Belgium (8095)	1,123	13.9	38.09
Canada (378)	59	15.6	2.00
France (3541)	154	4.3	5.22
Germany (4013)	570	14.2	19.34
Greece (3831)	255	6.7	8.65
Hong Kong (4)	0	0.0	0.00
Hungary (740)	171	23.1	5.80
Ireland (28)	2	7.1	0.07
Jordan (78)	8	10.3	0.27
Kuwait (149)	10	6.7	0.34
Lebanon (131)	19	14.5	0.64
Malta (82)	11	13.4	0.37
Netherlands (303)	21	6.9	0.71
Pakistan (51)	8	15.7	0.27
Poland (666)	279	41.9	9.46
Qatar (85)	20	23.5	0.68
Russia (299)	27	9.0	0.92
Saudi Arabia (264)	23	8.7	0.78
South Africa (93)	18	19.4	0.61
South Korea (863)	33	3.8	1.12
United Arab Emirates (129)	10	7.8	0.34

Abbreviations: DTTH, difficult‐to‐treat hypertension; ITT, intention to treat; OSCAR, Observational Study on Cognitive Function and SBP Reduction. ^aBelarus, Iran, The Philippines, Thailand: no DTTH patients reported. Spain: no data.

Mean SBP and mean pulse pressure (PP) were significantly (P=.001) higher at baseline in the DTTH population than in the non‐DTTH group; there was no significant difference for DBP. Systolic‐diastolic hypertension (SDH) and isolated systolic hypertension (ISH) accounted for 71.5% and 28.1% cases, respectively, in the DTTH subset.

All patients in the DTTH group were prescribed at least 3 drugs before the start of OSCAR; 25.7% of patients were receiving ≥4 drugs. Usage of individual classes of antihypertensive drugs at baseline was as follows: diuretic, 2204 (74.8%); β‐blocker, 1176 (60.2%); calcium antagonist, 1442 (48.9%); angiotensin‐converting enzyme inhibitor, 563 (19.1%); centrally acting drug, 315 (10.7%); α‐blocker, 216 (7.3%); and angiotensin II receptor blocker (ARB), 50 (1.7%). Principal recorded combinations are illustrated in Figure 2.

Figure 2.

 Major pre‐study combinations of antihypertensive drugs recorded at baseline in the difficult‐to‐treat hypertension (DTTH) population of the Observational Study on Cognitive Function and SBP Reduction (OSCAR). (a) Distribution of major drug combinations in 2189 patients recorded as receiving 3 drugs during the 4 weeks preceding inclusion in OSCAR. (b) Distribution of major drug combinations in 593 patients recorded as receiving 4 drugs during the 4 weeks preceding inclusion in OSCAR. ACEI indicates angiotensin‐converting enzyme inhibitor; CCB, calcium channel blocker; ARB, angiotensin receptor blocker. Note: Data from an additional 166 patients recorded as receiving ≥5 treatments are not illustrated.

Mean baseline MMSE score in the DTTH group was 26.5±3.6 (range 12–30), significantly lower (P<.001 by ANOVA) than in the non‐DTTH group (27.1±3.4).

BP Response

Mean systemic BP (SBP/DBP) decreased significantly during the study in the DTTH subgroup (from 164.8 mm Hg at baseline to 138.8 mm Hg at the end‐of‐study visit, P<.001 vs baseline). The reduction in mean SBP in the DTTH and non‐DTTH subgroups was similar (change from baseline in DTTH group, −26.0±15.7 mm Hg; change from baseline in non‐DTTH group, −25.8±13.9 mm Hg; P<.001 for both comparisons). The reduction in mean SBP was 6 mm Hg larger in DTTH patients with SDH (n=2108) than in DTTH patients with ISH (n=827) (∼28 mm Hg vs ∼22 mm Hg) (P<.001 vs baseline for SDH and ISH subsets; P=.075 for comparison of SDH vs ISH by analysis of covariance [ANCOVA]). The mean reduction in PP was −14.5±13.8 mm Hg (P<.001 vs baseline) and was larger in DTTH patients with ISH (−18.1±13.3 mm Hg) than in DTTH patients with SDH (−13.3±13.6 mm Hg; P<.001 vs baseline for both subsets; P<.003 by ANCOVA).

Normalization of BP, defined as SBP <140 mm Hg and DBP <90 mm Hg, was recorded for 1426 initially DTTH patients (48.4%), a significantly lower proportion than in the non‐DTTH ITT population (61.3%; P<.001 by χ² test).

A BP response to treatment, defined as fulfillment of the normalization criteria and/or reductions of ≥15 or ≥10 mm Hg in SBP or DBP, respectively, was achieved in 2576 DTTH patients (87.4%) compared with 92.1% in the non‐DTTH group (P<.001 by χ² test).

Evolution of MMSE Score

The overall mean MMSE score in the DTTH population after 6 months of EBT was 27.5±3.0 (P<.001 vs baseline). The MMSE average increment was thus 0.96±2.05. This increment in mean MMSE score did not differ significantly from that recorded in the non‐DTTH subgroup (Δ+0.79±1.79). Principal BP and MMSE data are summarized in Table III.

Table III.

 Principal Blood Pressure and MMSE Data for the DTTH Population

	Baseline	End of Treatment	Change From Baseline
SBP	164.8±14.5	138.8±12.2	−26.0±15.7
DBP	93.3±9.8	81.9±7.4	−11.4±9.8
MMSE	26.5±3.6	27.5±3.0	0.96±2.1

Abbreviations: DBP, diastolic blood pressure; DTTH, difficult‐to‐treat hypertension; MMSE, Mini‐Mental State Examination; SBP, systolic blood pressure. Values are expressed as (mean±standard deviation).

Safety and Tolerability

All 4649 DTTH patients were included in the safety population. A total of 101 adverse drug reactions were recorded in 72 DTTH patients (1.5%), 58 of which were classified as related to study medication. Most events (n=69) resolved without sequelae. Ten events were classified as severe and 9 were classified as serious; data were missing for 10 events. These adverse drug reactions have already been reported. ¹

Neither of the 2 deaths recorded during the study in DTTH patients were attributed to study medication.

Discussion

The primary results of OSCAR ¹ were supportive of the proposition that antihypertensive drugs that act via the renin‐angiotensin system (RAS) may have potential in preventing, delaying, or decelerating the onset and progression of cognitive decline.

The present analysis focused on a subset of OSCAR participants identified as having DTTH at baseline. Our retrospective examination of this cohort was motivated by the consideration that sustained high BP despite antihypertensive therapy has potential for adverse longer‐term consequences beyond familiar cardiovascular risk. In particular, increased arterial stiffness in the cerebrovascular circulation may contribute to cognitive decline. ⁷ , ⁸ , ⁹

DTTH was defined for this purpose as BP not satisfactorily controlled despite the use of at least 3 antihypertensive drugs. This was a pragmatic definition chosen to facilitate a retrospective investigation and should not be conflated with current definitions of “resistant hypertension.” Our definition of DTTH involves no consideration of lifestyle, the adequacy of drug dosing or adherence to therapy, the appropriateness of the ≥3 antihypertensive drug combination, faulty techniques in routine BP measurement in the office setting, or the possibility of white‐coat hypertension or pseudo‐hypertension. All of these factors need to be considered as part of the current best practice in the recognition of resistant hypertension. Retrofitting a model for the identification of true resistant hypertension ¹⁰ to the OSCAR data was beyond the resources of the present analysis.

In aggregate, some 11.5% of the OSCAR patients were classified as having DTTH at baseline, and the reported prevalence was ≤10% in 13 countries. Despite our use of a notionally less demanding definition than “true” resistant hypertension, this prevalence is lower than some published estimates for resistant hypertension, especially those for the United States. ¹¹

A substantial reduction in mean arterial BP was recorded in the DTTH subpopulation of OSCAR. The magnitude of this reduction was in line with that reported in some other recent evaluations of hypertension therapy based on ARBs or angiotensin‐converting enzyme inhibitors, ¹² , ¹³ and also that of Hypertension in the Very Elderly Trial (HYVET), ¹⁴ a substudy in which the MMSE was used to explore the impact of BP control on measures of cognition in patients older than 80 years. ¹⁵

The reduction in BP in our DTTH cohort during the period of observation coincided with a 1‐point increase in average MMSE. Care must be taken not to impute cause and effect to these changes in a post hoc analysis. It must also be recognized that the clinical relevance of a mean 1‐point change in MMSE during 6 months is open to question for a variety of methodologic reasons. ¹⁶ , ¹⁷ , ¹⁸ This is nevertheless a notable change over a relatively short space of time, especially in a patient population subject to a degree of ceiling effect and the upward trend implied in these data is a contrast to the findings of Okonkwo and colleagues, ¹⁹ who described progressive decrements in cognitive functioning in a population similar in age to our DTTH group (and with cardiovascular disease) despite their patients having lower baseline BP. MMSE testing in the Study on Cognition and Prognosis in the Elderly (SCOPE) in hypertensive patients at similar intervals to OSCAR also demonstrated a modest decline in performance, albeit in an older patient population. ²⁰ The observed mean improvement of 0.96±2.05 in our cohort is larger than the annual change reported for the control group of the study by Clark and colleagues and the similar change interpreted by Lopez and colleagues, ²¹ as indicating therapeutic benefit from cholinesterase inhibitors in patients with Alzheimer disease.

The unfavorable overall cardiovascular risk profile of our DTTH patients (Table I) was broadly in accordance with observations in resistant hypertension. ²² Cognitive function, as represented by MMSE score, was also lower at baseline in the patients with DTTH than in OSCAR comparators. We are unable to securely conclude if this difference was due to age, BP, cardiovascular risk profile, some combination of these influences, or additional factors. It is perhaps of note that median baseline MMSE in the DTTH subset of OSCAR patients was similar to HYVET, despite our patients being on average some 15 years younger. ²³ This could be an indication of an adverse effect of high BP and general cardiovascular status on cognition in our patients. Contributing factors might include diabetes, ²⁴ , ²⁵ metabolic syndrome, ²⁶ , ²⁷ and obstructive sleep apnea. ⁷ There is evidence for a role of angiotensin in these phenomena and as a corollary reason to believe that ARBs may be advantageous in these situations. ²⁸ , ²⁹ , ³⁰ , ³¹ , ³²

In both ISH and SDH patients in the DTTH subgroup, there was a substantial reduction in PP, determined principally by the SBP response. The similar improvement in MMSE score in both these sets of patients leads us to conjecture that any influence of hypertension on cognitive function in DTTH patients is exerted principally via SBP (and hence PP). This would be consistent with reports of reduced central PP and/or aortic pressure wave velocity or augmentation index (acting as proxies for systemic arterial stiffness) with antihypertensive therapies (including eprosartan) that target the RAS. ³³ Associations have also been reported between structural and functional deficits in the brain and PP or reduced cerebral blood flow. ³⁴ , ³⁵ , ³⁶ , ³⁷ These observations suggest plausible routes for positive effects of EBT on preservation of cognitive function, via direct effects on arterial pressure and through remodeling effects on vascular structure. Neuroprotective effects of ARB therapy have also been suggested. ³⁸

The possible significance of these observations for public health is indicated by reports that control of hypertension—and in particular SBP—in mid‐life may contribute to a reduction in the risk of dementia at older ages. ³⁹ , ⁴⁰ (Although the contrary conclusions of Johnson and colleagues ⁴¹ must also be noted.)

The overall safety and tolerability profile of eprosartan in OSCAR was replicated in this subgroup analysis and was in accordance with earlier assessments arising from randomized controlled trials. ⁴²

Conclusions

Use of EBT in a large group of patients retrospectively identified as having DTTH was associated with a substantial reduction in mean arterial BP and with an increase in the average MMSE score. These observations are consistent with the premise that use of eprosartan may delay or prevent cognitive decline in people with high BP. However, we concur with the sentiments of the HYVET group ⁴³ in regarding these findings principally as an indication of the need for further prospective investigation of the possible relation of antihypertension therapy and longer‐term trends in cognitive function.

Acknowledgments and disclosures:  RJP reports having received funds from the sponsor to conduct the research study and being involved in the editorial‐publication board. ES reports no conflict of interest. AOC reports no conflict of interest. JP‐B is an employee of Abbott Product Operations AG, sponsor of the study. AS is the CEO of EVIDENCE Based Communication (EBC), a France‐based medical communication company. EBC received fees from the sponsor in respect of the study conduct and analysis. AP reports funding by Abbott Products Operations, receipt of an honorarium from Abbott Products Operations, and reimbursement for attending Abbott Products Operations symposia. All the named authors contributed to the formulation of the study and the development of the study protocol. RJP, ES, AOK, and AP were involved in the conduct of the study in their respective countries. RJP, ES, AOK, J‐PB, and AP contributed to the analysis of the data and the development of the manuscript. AS contributed to the Methods section of the manuscript.

Funding sources:  The OSCAR study is supported financially by Abbott Products Operations AG, Hegenheimermattweg 127, 4123 Allschwil, Switzerland. Preparation of this report was assisted by Hughes associates, 3 Collins St, Oxford, OX4 1XS, UK.