Abstract
Objective:
To determine whether patients who were newly prescribed antihypertensive therapy after the Eighth Joint National Committee (JNC 8) update were treated to a relaxed systolic blood pressure (SBP) goal compared with patients treated before the update.
Methods:
A retrospective cohort study approved by the Colorado Multiple Institutional Review Board. Patients aged 60–79 years, without diabetes or chronic kidney disease (CKD), newly treated for hypertension at a University of Colorado primary care clinics were included. The mean first-achieved and last-stable SBPs of patients newly prescribed antihypertensive medications from 1 January 2012 to 31 December 31 2013 (before cohort) were compared with patients newly prescribed antihypertensive therapy from 1 January 2014 to 1 October 2015 (after cohort). The mean number of antihypertensive medications at first-achieved SBP, the time to first-achieved SBP, and the class of initial antihypertensive medications were also evaluated.
Results:
A total of 128 patients were included, 64 patients in each cohort. The coprimary outcome of first-achieved mean SBP did not differ between the groups (131.3 mmHg versus 130.2 mmHg; p = 0.65). Last-stable mean SBP values were also similar between the groups (130.2 mmHg versus 129.5 mmHg; p = 0.74). Angiotensin converting enzyme inhibitors (ACE-I) were the most frequently initiated antihypertensive agent in both cohorts (43.8% versus 48.4%; p = 0.72).
Conclusions:
Our findings suggest that the JNC 8 recommendations did not alter SBP goals among patients aged 60–79 years newly treated for hypertension at University of Colorado primary care clinics.
Introduction
Hypertension is a major modifiable risk factor for cardiovascular disease (CVD), stroke, and chronic kidney disease (CKD). Nearly one-third of US adults aged ⩾20 years have recognized hypertension [Mozaffarian et al. 2016]. The prevalence of hypertension increases with age, with 65% of patients aged ⩾60 years meeting diagnostic criteria for hypertension [Mozaffarian et al. 2016]. A greater percentage of patients aged ⩾60 years have controlled hypertension (blood pressure <140/90 mmHg) compared with younger patients, but receive suboptimal management based on a blood pressure (BP) control rate of 54.1% in this population [Mozaffarian et al. 2016].
In 2014, two US hypertension guidelines were published – the American Society of Hypertension and the International Society of Hypertension (ASH/ISH) guideline and a report from panel members appointed to the Eighth Joint National Committee (JNC 8) [Weber et al. 2014; James et al. 2014]. Both of these guidance documents provided recommendations on BP thresholds for initiating antihypertensive treatment, preferred initial pharmacologic options, and goal BP values in specific patient populations [Weber et al. 2014; James et al. 2014]. While both recommended a less aggressive systolic blood pressure (SBP) goal of <150 mmHg in elderly patients, they recommended different ages cutoffs at which the SBP goal should be relaxed The JNC 8 report recommends a SBP target of <150 mmHg in adults aged ⩾60 years, without diabetes or CKD. This recommendation was given a Grade A recommendation, using a grading system developed by the National Heart, Lung, and Blood Institute’s Evidence-Based Methodology Lead [James et al. 2014]. Although graded as a strong recommendation, the age at which to relax the SBP goal differs from several other hypertension guidelines within and outside of the US which recommend relaxing the SBP goal to <150 mmHg for patients aged ⩾80 years [Leung et al. 2016; Mancia et al. 2013; Weber et al. 2014].
Several cardiovascular groups, including a minority of JNC 8 panel members, expressed concern that evidence supporting the relaxed SBP target is insufficient [Krakoff et al. 2014; Wright et al. 2014]. Moreover, the less aggressive SBP target may lead to under treatment of high-risk elderly patients. Additionally, several registry analyses have compared the proportion of patients meeting BP targets as recommended by JNC 8 with previous JNC 7 recommendations. These analyses found that 6.8–11.6% of adults aged ⩾60 years without diabetes or CKD would be reclassified as meeting BP targets by targeting a less aggressive SBP goal of <150 mmHg [Borden et al. 2014; Navar-Boggan et al. 2014; Miedema et al. 2015]. Given the controversy surrounding the relaxed SBP goal in patients aged ⩾60 years, and several other hypertension guidelines recommending a less aggressive SBP goal starting at age 80, the objective of this study was to determine whether patients newly prescribed antihypertensive therapy after the JNC 8 report were treated to a relaxed SBP compared with patients newly treated before the JNC 8 report.
Methods
Study design
This retrospective cohort study compared the SBP values of patients newly prescribed antihypertensive therapy for a new diagnosis of hypertension in one of seven University of Colorado Health (UCH) primary care clinics. These seven clinics included family medicine and internal medicine clinics, three of which are medical resident training sites, and most have clinical pharmacy services. Two primary endpoints were used as surrogate markers for SBP goals between cohorts, and include the mean first-achieved SBP and mean last-stable SBP. First-achieved SBP was defined as the SBP value, obtained during a primary care clinic visit in which hypertension was assessed after prescribing antihypertensive therapy, and no modifications were made to the antihypertensive treatment plan. Last-stable SBP was defined as the most recent in-office SBP reading in the study period, where hypertension was assessed and no medication modifications were made to the antihypertensive treatment plan. SBP values recorded during clinic visits where hypertension was not assessed (i.e. acute illness visits) were not included. Secondary endpoints included the time to first-achieved SBP, the number of antihypertensive medications prescribed to reach first-achieved SBP, the class of antihypertensive medication prescribed following hypertension diagnosis, and the number of antihypertensive medications at last-stable SBP.
Patient population
Patients aged 60–79 years with newly prescribed treatment for hypertension during the study period of 1 January 2012 to 1 October 2015 were eligible for this study. Additional inclusion criteria included a new diagnosis of hypertension (ICD-9 code 401.x) and at least one additional visit to a University of Colorado primary care clinic after being prescribed antihypertensive therapy, where hypertension was assessed and no changes were made to patients’ antihypertensive regimen. Exclusion criteria included a diagnosis of diabetes mellitus (ICD-9 code 250.x), CKD (ICD-9 code 585.x), or a kidney transplant (ICD-9 code V42.0). Patients receiving antihypertensive therapy for indications other than hypertension (i.e. atrial fibrillation, benign prostate hypertrophy, edema) prior to the new hypertension diagnosis were also excluded. A list of potentially eligible patients was generated from our electronic health record (EHR) based on study criteria.
Potentially eligible patients were assigned to one of two cohorts (Figure 1: Study cohorts by date). The before JNC 8 group consisted of patients newly prescribed antihypertensive therapy during the period 1 January 2012 to 31 December 2013, while patients newly prescribed antihypertensive therapy during the period 1 January 2014 to 1 October 2015 comprised the after JNC 8 group. The cohort time periods were chosen to provide an approximately equal time frame, and to allow time for the JNC 8 report to be disseminated to primary care providers. The online JNC 8 report was first available on 18 December 2013. The initial date of 1 January 2014 was chosen for the after JNC cohort, in the belief that enough time had passed since the online publication of the report for providers to be aware of the JNC 8 recommendations.
Study cohorts by date.
Patients in each cohort were randomly sorted, then EHRs were manually reviewed to ascertain if each patient met the criteria. The following patient data were extracted and collected for analysis: age, gender, ethnicity/race, current tobacco use, date of newly prescribed antihypertensive agent, initial in-clinic SBP, antihypertensive agent prescribed, unique number of antihypertensive medications at primary care visits, dates at which BP is assessed in clinic visit and where no changes were made to antihypertensive regimen, and SBP on eligible dates.
Statistical analysis
Primary and secondary continuous endpoints were compared with Microsoft Excel between cohorts using an unpaired t test. Categorical data were analyzed using two-tailed chi-square or Fisher’s exact tests, calculated with GraphPad Software (http://graphpad.com/quickcalcs/contingency2/). A sample size calculation (http://clincalc.com/Stats/SampleSize.aspx) determined that a total of 128 patients (64 per group) would be necessary to provide 80% power (α = 0.05) for detecting a clinically significant 6 mmHg SBP difference (standard deviation 12 mmHg) between groups, similar to the BP parameters reported in the Valsartan in Elderly Isolated Systolic Hypertension (VALISH) study [Ogihara et al. 2010]. Patient EHRs were reviewed until a total of 64 patients meeting inclusion criteria in each cohort were identified. This study received expedited approval by Colorado Multiple Institutional Review Board.
Results
Of the 671 patients reviewed, 128 patients met study criteria (Figure 2: Patients screened and included in analysis). Most patients (67.2%) were excluded for having a previous diagnosis of hypertension at their initial visit when establishing care at one of the primary care clinics. Baseline characteristics were similar between groups (Table 1), which included a majority of non-Hispanic white patients, mean age of 66 years, and low prevalence of reported tobacco use. Patients with known CVD were excluded from this study population, as they would have likely been receiving one or more antihypertensive agents as recommended by CVD secondary prevention guidelines [Fihn et al. 2012]. Patients in both cohorts had a similar initial SBP, which trended higher in the after JNC 8 group, but this was not statistically different (153.7 mmHg versus 157.7 mmHg; p = 0.07).
Patients screened and included in analysis.
Patient demographics.
Values are mean ± standard deviation.
DBP, diastolic blood pressure; JNC 8, Eighth Joint National Committee; SBP, systolic blood pressure; SD, standard deviation.
The coprimary endpoint (Table 2) of mean first-achieved SBP did not differ significantly between groups (131.3 versus 130.2 mmHg; p = 0.65). Additionally, no statistical difference in the second coprimary endpoint of mean last-stable SBP (Table 2) was found between study groups (130.2 versus 129.5 mmHg; p = 0.74). No significant different in diastolic BP was noted between the two cohorts.
Primary endpoints.
Values are mean ± standard deviation.
DBP, diastolic blood pressure; JNC 8, Eighth Joint National Committee; SBP, systolic blood pressure.
Table 3 reports the frequencies of each antihypertensive agent prescribed as initial therapy. Angiotensin converting enzyme inhibitors (ACE-I) were the most frequently prescribed initial antihypertensive agent in both the before and after cohorts, 43.8% versus 48.4% respectively (p = 0.72). Calcium channel blockers (CCBs) or thiazides were initiated as initial antihypertensive therapies in non-Hispanic black patients at a similar frequency in both the before and after cohorts, 57.1% versus 60.0% respectively (p = 1.0).
Secondary endpoints.
Values are means.
ACE-I, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCB, calcium channel blocker; HCTZ, hydrochlorothiazide; JNC 8, Eighth Joint National Committee; No., number; SBP, systolic blood pressure.
Discussion
This retrospective study of patients aged 60–79 years newly prescribed antihypertensive treatment found no difference in our primary endpoints of first-achieved SBP and last-stable SBP among patients treated before and after the JNC 8 report. Both cohorts reached a mean SBP of approximately 130 mmHg with minimal antihypertensive medications (mean 1.2 antihypertensive agents). This suggests a high degree of goal attainment, regardless of the BP goal value selected. Both cohorts demonstrated a considerable reduction in SBP (−22.4 mmHg in the before JNC 8 cohort and −27.5 mmHg in the after JNC 8 cohort), easily meeting an SBP goal of <140 mmHg and <150 mmHg. Based on this finding, the achieved SBP may not serve as the best surrogate for the SBP goal as providers may have relaxed their SBP goal for patients but not changed therapy, despite patients having a SBP dramatically lower than goal as long as they were not having complications from treatment. This is supported by the JNC 8 report recommendation that antihypertensive treatment need not be modified if patients achieve a SBP of <140 mmHg and have no adverse effects associated with antihypertensive treatment (Grade E – expert opinion) [James et al. 2014]. Thus, in patients with a SBP goal of <150 mmHg who achieved a SBP <140 mmHg, modification of antihypertensive therapy would not be necessary, unless they experience adverse effects. Patients in both cohorts achieved and maintained well-controlled SBP values which did not appear to be associated with any adverse effects that would require modifications to antihypertensive therapy.
Our results suggest that patients included in our study may have been treated to a more intense SBP goal than recommended by the JNC 8 report. Recently, the Systolic Blood Pressure Intervention Trial (SPRINT) found that targeting a more intensive SBP goal (<120 mmHg) compared with standard SBP (<140 mmHg) in patients with CVD risk factors resulted in a significantly lower rate of major CVD events and all-cause mortality [Wright et al. 2015]. Our study population differed from patients included in SPRINT in many ways. First, our study population did not include patients with CKD and known CVD, while the SPRINT trial included a high percentage of patients with CKD (28%) and known CVD (20%). Additionally, most patients in the SPRINT trial (>90%) were receiving antihypertensive treatment at baseline with a mean SBP of 140 mmHg. Our patient population included patients with newly diagnosed, untreated hypertension with a low CVD risk. Recently, a subgroup analysis of patients aged ⩾75 years from the SPRINT trial reported a >30% reduction in all-cause mortality and the primary composite endpoint of major cardiovascular events in the group treated to intensive SBP goal <120 mmHg, compared with SBP < 140 mmHg [Williamson et al. 2016]. Given evidence from the SPRINT trials, the JNC 8 recommendation to target an SBP of <150 mmHg in all individuals over the age of 60 is not consistent with the current clinical trial data.
A previous study conducted at the same academic health system found no difference in mean BP and the number of antihypertensive agents between a similar cohort of patients aged 60–79 years treated for hypertension 1-year before and 1-year after the JNC 8 report [Fixen et al. 2016]. Of note, patients in the previous study had an existing diagnosis of hypertension and were currently receiving antihypertensive therapy. Additionally, most patients (86%) were included in both the before and after cohort. Our current study excluded patients with a prior diagnosis of hypertension as well as those currently receiving antihypertensive medications. In addition, our study design did not permit patients to be included in both the before and after cohorts. Therefore, our results are a better measure of how the JNC 8 report has influenced primary care providers in treating newly diagnosed hypertension in an elderly population.
One possible explanation for the substantial reduction in SBP seen in both cohorts may be due to the health status of the patients studied. Patients in our study had a mean age of 66 years, which is younger than patients studied in previous isolated systolic hypertension trials (mean age range 70–84 years) [SHEP Cooperative Research Group, 1991; Staessen et al. 1997; Beckett et al. 2008; JATOS Study Group, 2008; Ogihara et al. 2010]. Our patients had few comorbidities, as our exclusion criteria did not include patients with diabetes or CKD, and patients with known CVD receiving recommended medical therapy would have been excluded. Of the five trials reviewed by the JNC 8 members that included patients aged ⩾60 years, four trials included patients (range 7–13.7%) with diabetes mellitus [SHEP Cooperative Research Group, 1991; Beckett et al. 2008; JATOS Study Group, 2008; Ogihara et al. 2010]. The Systolic Hypertension in Europe trial did not report the percentage of patients with diabetes, but did report that nearly 30% of included patients had a history of cardiovascular complications at baseline [Staessen et al. 1997].
It is possible that a portion of our study patients may not have had true hypertension, but instead presented with white-coat hypertension (WCH). It is estimated that up to 30% of patients with elevated in-clinic BP values may have WCH [Siu and U.S. Preventive Services Task Force, 2015]. In 2015, the US Preventive Services Task Force released recommendations to confirm the diagnosis of hypertension with the use of out-of-clinic BP values [Siu and U.S. Preventive Services Task Force, 2015]. A recent subgroup analysis examined the prevalence of WCH in patients with in-clinic hypertension (BP ⩾140/90 mmHg) who also underwent 24-h ambulatory blood pressure monitoring (ABPM) [Tanner et al. 2016]. In the subset of patients without diabetes or CKD, patients aged ⩾60 years had a significantly greater in office SBP compared with patients aged <60 years (149.9 mmHg versus 140.9 mmHg, respectively; p < 0.001), but similar SBP measured by ABPM (134.8 mmHg versus 132.2 mmHg; p = 0.22). The prevalence of patients in our study with WHC cannot be determined given the retrospective study design.
Both the ASH/ISH guidelines and the JNC 8 report recommend one of four classes of antihypertensive agents as initial therapy in non-Hispanic black patients. Our study found that ACE-I were the most frequently prescribed initial agent in both cohorts, and also noted several other trends in prescribed initial antihypertensive therapy. First, the frequency of beta-blockers as initial therapy in the before JNC 8 cohort trended down in the after JNC 8 cohort (6.3% versus 3.1%; p = 0.68), which is in line with recommendations against the use of beta-blocking agents as first-line antihypertensive agents without a compelling indication [Weber et al. 2014; James et al. 2014]. Second, we found an increased frequency of angiotensin receptor blockers (ARBs) use as initial therapy from 7.8% in the before cohort to 10.9% in the after cohort (p = 0.76). This trend may be explained by the fact that several ARBs became available as generic products after the release of JNC 8, making ARBs a more cost-effective option. Due to ASH/ISH and JNC 8 recommendations specifying initial antihypertensive therapy in non-Hispanic black patients, a prespecified subgroup analysis was conducted to determine if more non-Hispanic black patients in the after JNC 8 cohort were initially prescribed a CCB or thiazide compared with non-Hispanic black patients in the before JNC 8 cohort. A total of 12 non-Hispanic black patients were included in the patient sample, (7 in the before cohort and 5 in the after cohort) with no statistical difference between groups. Based on the small numbers of non-Hispanic black patients in our study, a future study including only non-Hispanic black patients may be useful to determine if initial antihypertensive therapy follows the recommendations from both US hypertension guidelines [Weber et al. 2014; James et al. 2014].
There are several limitations of this study, most notably our retrospective design. First, medication adherence could not be assessed due to our study design, which may have falsely increased our secondary endpoint of time to first-achieved SBP. In a study assessing adherence to antihypertensive medications among patients newly treated for hypertension, <10% of subjects had an adherence rate ⩾80% (portion of days covered) at 6 months [Mazzaglia et al. 2009]. Secondly, we cannot ensure that BP was measured using the recommended technique described by hypertension guidelines [Weber et al. 2014]. Current practice for BP measurement at most UCH primary care clinics involves a medical-assistant rooming the patient and obtaining vitals with an automated BP cuff, while the patient is sitting. Providers are free to remeasure BP manually, at their discretion, and may record this value in the EHR. Due to our retrospective design, it is therefore not possible to determine whether BP measurements were taken correctly or consistently at each primary care clinic. A third limitation of this study is our use of surrogate endpoints to evaluate SBP goals, since BP goals were not consistently documented in the EHR. Although we limited our assessment to only SBP values during clinic visits which were assessing patients’ hypertension disease state, patients who achieved a SBP below their goal value without adverse effects would be recommended to continue current antihypertensive therapy, based upon the JNC 8 corollary recommendation [James et al. 2014]. Lastly, our study included a limited number of primary care clinics, all of which are affiliated with a single academic health system and have clinical pharmacy services available. Several clinics have collaborative drug therapy management protocols, which allow pharmacists to manage and modify patients’ antihypertensive therapy in collaboration with primary care providers. Whether the availability of clinical pharmacy services influenced the SBP goal of patients included in this study cannot be determined. Future studies comparing the achieved SBP values for patients initiating antihypertensive treatment in primary clinics affiliated with an academic health center compared with primary care clinics not affiliated with academic health centers may be of value.
Conclusion
Our findings did not identify a change in overall intensity of BP lowering among patients with newly diagnosed hypertension. Our findings were limited to patients aged 60–79 years, without CKD or diabetes, who were receiving care from primary care providers of an academic health system. Overall, treatment of this population does not appear to have been negatively influenced by the JNC 8 recommendations to target a less aggressive SBP goal among these patients.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare that there is no conflict of interest.