Author + information
- Received March 14, 2016
- Revision received March 28, 2016
- Accepted March 29, 2016
- Published online June 21, 2016.
- Min Jung Ko, PhDa,
- Ae Jung Jo, MSca,
- Chan Mi Park, PhDa,
- Hyo Jeong Kim, MPHa,
- Yun Jung Kim, MPHa and
- Duk-Woo Park, MD, PhDb,∗ ()
- aDivision for Healthcare Technology Assessment Research, National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea
- bDivision of Cardiology, Asan Medical Center, University of Ulsan College of Medicine and National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea
- ↵∗Reprint requests and correspondence:
Dr. Duk-Woo Park, Division of Cardiology, Asan Medical Center, University of Ulsan College of Medicine, National Evidence-based Healthcare Collaborating Agency, 388-1 Poongnap-dong, Songpa-gu, Seoul 138-736, Republic of Korea.
Background Blood pressure (BP) targets from the SPRINT (Systolic Blood Pressure Intervention Trial) differ from targets of the 2014 hypertension (HTN) recommendations of the Eighth Joint National Committee.
Objectives The goal of this study was to estimate the proportion of hypertensive adults with who would meet BP goals under the SPRINT criteria and under the 2014 recommendations, and to determine related effects on cardiovascular morbidity and mortality.
Methods We used data from the Korean National Health and Nutrition Examination Survey of 2008 to 2013 (n = 13,346), as well as the Korean National Health Insurance Service health examinee cohort of 2007 (n = 67,965), to estimate the proportion of subjects meeting BP goals of each of the criteria. Using data from the Korean National Health Insurance Service health examinee cohort of 2007 (n = 67,965), we compared risks of major cardiovascular events (composite of nonfatal myocardial infarction, nonfatal stroke, or death from cardiovascular causes) associated with different BP control goals.
Results A substantially lower proportion of hypertensive adults met BP goals of the SPRINT criteria compared with the 2014 recommendations (11.9% vs. 70.8%, respectively). Ten-year predicted cardiovascular risks were lowest in the intensive control group (below SPRINT BP goals), intermediate in the less-intensive group (above SPRINT goals but below 2014 recommendation goals), and highest in the uncontrolled group (above 2014 recommendations) (6.15%, 7.65%, and 9.39%, respectively; p < 0.001). After multivariable adjustment, the less-intensive and uncontrolled groups had a greater risk of major cardiovascular events (hazard ratios 1.17 and 1.62, respectively; p value for trend, <0.001) than the intensive group.
Conclusions Substantially fewer hypertensive adults would meet SPRINT BP goals than would meet 2014 recommendation goals. Stricter BP control is associated with a decreased risk of major cardiovascular events.
Hypertension (HTN) affects >1 billion adults worldwide and is a critically important modifiable risk factor for cardiovascular morbidity and mortality (1). Given the high prevalence of HTN and its impact on cardiovascular disease (CVD) outcomes, optimal management of high blood pressure (BP) is a public health priority. The BP goals of the Eighth Joint National Committee (JNC 8) of 2014 (150/90 mm Hg for older adults ≥60 years of age and 140/90 mm Hg for younger adults <60 years of age, with diabetes, or with chronic kidney disease [CKD]) (2), were less stringent than those of the committee's previous (JNC 7) recommendations (3). Scientific evidence for the 2014 recommendations was primarily drawn from randomized controlled trials, but they have been highly debated (4–6).
In contrast with the less intensive BP targets of the 2014 recommendations, SPRINT (Systolic Blood Pressure Intervention Trial) found that aggressive lowering of systolic BP to <120 mm Hg, compared with the standard goal of <140 mm Hg, resulted in significantly lower rates of major cardiovascular events and mortality among adults with HTN but without diabetes (7). The recommended BP levels of the 2014 recommendations conflict with those of SPRINT, and it is unknown how this discrepancy affects overall achievement of BP control and CVD outcomes.
Using data from the Korea National Health and Nutrition Examination Survey (KNHANES), as well as the Korean National Health Insurance Service (NHIS) health examinee cohort, we estimated the proportion of adults who would have met BP goals under the 2014 recommendations and SPRINT criteria, assessed their cardiovascular risk factor burden, and compared risks of occurrence of major cardiovascular events associated with the BP targets under both criteria.
Data sources and study population
Two large nationally representative databases were used. To estimate the proportion of Korean adults with HTN who would meet BP targets under the 2014 HTN recommendations and SPRINT criteria, data collected from the KNHANES between 2008 and 2013 were used. KNHANES is a cross-sectional survey designed to represent the civilian noninstitutionalized Korean population. Details of the KNHANES regarding health interviews, standardized physical examinations, laboratory examinations, and definition of risk factors have been described previously (8). BP was measured at mobile examination centers by trained examiners. Systolic and diastolic BPs were measured 3 times, and the average of the second and third measurements was used for analysis (9,10). Patient-reported use of antihypertensive medication was assessed during interviews. Of 41,321 KNHANES participants ≥18 years of age, only adults with HTN in our analysis were included. We excluded participants who had missing data on BP measurements and those who had missing KNHANES sample weights (Online Figure 1). Baseline characteristics of all KNHANES participants (those included and excluded from our study) are summarized in Online Table 1.
To assess the clinical impact of the different BP targets of the 2014 recommendations and SPRINT, we used data derived from a representative sample cohort of the National Health Examination from 2007 with clinical follow-up for adverse cardiovascular events and death through December 31, 2013. The Republic of Korea has a universal health coverage system, in which the NHIS is a single-payer program, and it is mandatory for all residents in Republic of Korea; all insured individuals and their dependents are required to undergo a periodic (i.e., mostly biennial) general health examination. The Korean NHIS maintains a large health dataset and provides periodic updates on health-related risk factors and baseline biochemical data (11,12). The National Health Examination followed a standard procedure, the details of which were published previously (9). BP measurements were taken at local hospitals, each of which had met the internal and external quality control procedures of the Korean Association of External Quality Assessment Service. Among 162,593 participants ≥20 years of age from the 2007 NHIS health examinee cohort, adults with HTN were evaluated. Participants were excluded who had a history of CVD, those who had missing data on BP measurements, those who died before January 1, 2008, and those who had laboratory examination data outliers (Online Figure 1).
The institutional review board of the National Evidence-based Healthcare Collaborating Agency (Seoul, Republic of Korea) approved the study protocol. All KNHANES participants provided written informed consent for participation. The requirement for informed consent of the NHIS health examinee cohort was waived.
Definition of HTN and targeted BP goals
For our study, HTN was defined as physician-diagnosed, patient-reported BP treatment or a systolic BP ≥130 mm Hg, which is the same BP threshold used in the SPRINT trial (7). Among adults with HTN, we assessed the proportion who would meet BP treatment goals under SPRINT criteria (systolic BP <120 mm Hg) and those who would meet BP treatment goals under the 2014 HTN recommendations (18 to 59 years of age without CKD or diabetes, <140/90 mm Hg; ≥60 years of age without CKD or diabetes, <150/90 mm Hg; subjects with CKD, <140/90 mm Hg; and subjects with diabetes, <140/90 mm Hg) (2).
Clinical outcome events and assessments
The primary clinical outcome for the NHIS health examinee cohort was first occurrence of a major cardiovascular event, defined as the composite of nonfatal myocardial infarction (MI), nonfatal stroke, or death from cardiovascular causes. Secondary outcomes included individual components of the primary composite outcome and death from any cause. Events were assessed until December 31, 2013.
Cardiovascular events were ascertained according to the principal diagnosis of hospital admissions on the basis of International Classification of Diseases-10th Revision codes (for MI, I21, I22, and I23; for stroke, I60, I61, I62, and I63). We obtained information on vital status, which was prospectively recorded in the NHIS databases. Data on vital status and date of death were reconfirmed, and the cause of death was determined from the National Population Registry of the Korea National Statistical Office with the use of a unique personal identification number, in which central registration of death was conducted on the basis of death certificates (13). This approach provides a complete event ascertainment (because the NHIS and National Statistical Office are national organizations covering all Korean subjects). “Cardiovascular death” includes sudden cardiac death, death due to acute MI, death due to stroke, death due to heart failure, and death due to other cardiovascular causes.
Because of the stratified, complex, multistage, clustered probability-sampling design of the KNHANES, domain analysis was conducted with KNHANES-supplied sampling weights. These account for oversampling of subgroups and nonresponse rate to estimate the weighted percentages of adults in each segment of interest (8). We estimated the proportion of adults with HTN who would meet BP treatment goals under the 2014 recommendations and SPRINT criteria. Risk factor profiles and current BP treatment status were summarized for persons and grouped according to levels of BP control: 1) intensive group (below SPRINT BP goals); 2) less-intensive group (above SPRINT BP goals but below 2014 recommendations); and 3) uncontrolled BP (above 2014 recommendations). In addition, assuming that the mean estimation of 10-year risk of cardiovascular events approximates the actual rates of those events, the Framingham risk score was calculated to estimate the expected rates of cardiovascular events according to these various levels of BP control (14). Considering the disparity of a country- or ethnic-specific risk function, the predicted risks were also measured by using Korean-derived risk equations (a Korean CHD risk score) (15).
In the NHIS health examinee cohort, BP targets of the 2014 recommendations and SPRINT criteria were uniformly applied to classify groups according to levels of BP control. Risk factor profiles and comorbidities were summarized for persons in each group. For all cardiovascular events and deaths, the time to an event was calculated to obtain incidence rates per 1,000 person-years that were age-standardized after adjustment for sex. A Cox proportional hazards model was used to test for trends in the incidence of adverse cardiovascular events across BP control status groups (intensive, less-intensive, and uncontrolled). To account for potentially confounding clinical covariates and to adjust the established risk factors for CVD, multivariable models were adjusted for age at baseline, sex, presence or absence of diabetes, presence or absence of CKD, smoking status, presence or absence of hyperlipidemia, body mass index, use or nonuse of antihypertensive treatment, use or nonuse of aspirin, use or nonuse of statin therapy, and Charlson comorbidity index.
Because the JNC 7 recommendations were the standard of care for patients during the time period of the study cohort, we also compared the proportion of adults who met BP goals and potential clinical effects between the JNC 7 and SPRINT criteria. In addition, because there could be variations in BP values over time, a sensitivity analysis was performed by using time-updated Cox models with systolic BP as time-varying covariates among adults who had at least 1 more follow-up examination at 2-year intervals after the baseline visits. Lastly, similar analyses were performed if conventional BP criteria (systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg) were applied instead of the systolic BP level of ≥130 mm Hg used in SPRINT.
All reported p values were 2-sided, and those <0.05 were considered statistically significant. For all statistical analyses, SAS version 9.4 software (SAS Institute, Inc., Cary, North Carolina) was used.
KNHANES sample and estimates in the overall population
Of the 13,346 participants with HTN included in our final analysis, 1,677 (12.6%) would meet target BP goals on the basis of the SPRINT criteria, compared with 9,591 (71.9%) who would meet the goals of the 2014 recommendations; a summary of characteristics are provided in Online Table 2. Hence, 7,914 adults (59.3%) who previously met BP goals under the 2014 recommendations would no longer meet BP goals under the SPRINT criteria.
Estimates of the proportion of all Korean adults with HTN and those who would meet or not meet BP goals under the 2014 recommendations and the SPRINT criteria based on the KNHANES sample weights are shown in the Central Illustration. Among 11.0 million adults with HTN, an estimated 7.8 million (70.8%) would meet target BP goals under the 2014 recommendations, but only 1.3 million (11.9%) would meet BP goals under the SPRINT criteria. The proportions of adults with HTN who met BP goals under each set of criteria according to major subgroups are shown in Online Figure 2.
The clinical characteristics of adults with HTN and those stratified according to the levels of BP control status under 2 different criteria are shown in Table 1. The median systolic BP was 112 mm Hg in adults who met the SPRINT BP goals, 132 mm Hg in those who met the 2014 recommendation goals but not the SPRINT goals, and 150 mm Hg in those who did not meet the 2014 recommendation goals. Adults who met stricter BP goals were older, more likely to be female, and had a higher incidence of diabetes, CKD, and prevalent CVD. The predicted 10-year risk of CVD events, as estimated by using the Framingham risk score according to different levels of BP control are illustrated in the Central Illustration. The mean estimate of the 10-year risk of CVD events was lowest in adults who met the SPRINT goals (6.15; 95% confidence interval [CI]: 5.64 to 6.66), intermediate in those who met the 2014 recommendation goals but not the SPRINT goals (7.65; 95% CI: 7.34 to 7.96), and highest in those who did not meet the 2014 recommendation goals (9.39; 95% CI: 8.88 to 9.90) (p < 0.001). When the Korean CHD risk score was applied, incremental trends of CVD risks according to increasing BP levels were similar (Online Figure 3).
Potential effects of different BP targets in the NHIS health examinee cohort
When the different BP target criteria were applied to the NHIS health examinee cohort, among 67,965 adults with HTN, 49,855 (73.4%) met target BP goals according to the 2014 recommendations but only 5,639 (8.3%) met BP goals according to the SPRINT criteria. The clinical characteristics of groups stratified by status of BP control levels according to the 2 criteria are summarized in Table 2.
The mean duration of follow-up was 6.6 ± 1.20 years. During the follow-up period, 1,158 subjects had at least 1 major cardiovascular event; 513 experienced an MI, 664 experienced stroke, and 298 died of cardiovascular causes. A total of 1,982 deaths from all causes were observed. Standardized event rates and hazard ratios for clinical outcomes according to different levels of BP control under the 2014 recommendations and the SPRINT criteria are shown in Table 3. Age- and sex-adjusted risks for major cardiovascular events and for MI, stroke, or cardiovascular death were lowest in the intensive BP control group and highest in the uncontrolled BP group. After multivariable adjustment for potentially confounding clinical covariates, a significant linear trend suggesting an association between increased risk of events and lesser BP control was reported for major cardiovascular events (with the intensive group as the reference category, hazard ratios were 1.17 for the less-intensive group and 1.62 for the uncontrolled group; p for trend, <0.001) (Table 3, Figure 1). A significant linear trend was also observed for MI or stroke. However, there was no significant difference in cardiovascular or all-cause mortality.
Comparison between the JNC 7 recommendations and the SPRINT criteria
When the differences between the JNC 7 recommendations and the SPRINT criteria were compared, the proportion of adults who met JNC 7 BP goals was 57.2%, compared with 11.9% who met SPRINT criteria (Online Figure 4). The incremental trends of predicted CVD risks according to different levels of BP control were also similar (Online Figure 5). When the potential clinical effects of differences between the JNC 7 and the SPRINT criteria were determined in the NHIS health examinee cohort, there was a significant linear trend toward an increased risk of major cardiovascular events with worse BP control (Online Table 3, Online Figure 6). A linear trend was also significant for cardiovascular death, MI, or stroke.
Of 67,965 participants overall, 61,222 adults (90.1%) had at least 1 follow-up examination in the 2-year interval after the baseline visit. After adjusting for systolic BP as a time-varying covariate, the overall findings were similar (Online Table 4). When conventional BP criteria (systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg) were adopted as an alternative definition of HTN instead of the SPRINT criteria (systolic BP ≥130 mm Hg), a significant trend was observed for major cardiovascular events, MI, or stroke, but the minimal benefit difference between the SPRINT and 2014 recommendations did not occur (Online Table 5).
Our study includes 3 major findings. First, we estimated that among 11.0 million Korean adults with HTN, the proportion of adults who would meet BP goals would significantly decrease from 7.8 million (70.8%) under the 2014 recommendations to 1.3 million (11.9%) under the SPRINT criteria. Second, among different groups of incremental BP control status, the estimated 10-year risk of CVD events was lowest in the intensive BP control group (below SPRINT BP goals), intermediate in the less-intensive group (above SPRINT BP goals but below 2014 recommendation BP goals), and highest in the uncontrolled group (above 2014 recommendation BP goals). Third, when we estimated the actual event rates in the national health examinee cohort, there was a significant linear trend, suggesting an increased risk of major cardiovascular events in patients with less stringent BP control.
Despite greater cardiovascular protection with intensive BP lowering, achieving SPRINT-defined BP goals might not be easy or practical because the target BP was not met in more than one-half of the participants in the intensive-treatment group (7). Furthermore, given that the less intensive BP goals of the 2014 recommendations were only achieved in approximately 40% to 60% of patients with treatment-eligible HTN (9,16), a more aggressive and time-consuming approach would be required to achieve SPRINT BP goals (17). In our study, strict BP control of <120 mm Hg would be met in only about 12% of Korean adults with HTN. However, because SPRINT enrolled older adults ≥50 years of age without diabetes, prior stroke, heart failure, or CKD, SPRINT criteria in our general population may have limited applicability. Further research is required to determine how the SPRINT results might change practice patterns and BP control on a population-wide basis, as well as how such changes may importantly influence cardiovascular-related outcomes in clinical practice.
Some observational studies and clinical trials suggest that more intensive BP control (systolic BP <120 mm Hg) does not provide evident cardiovascular and mortality benefits compared with standard control (18,19). Conversely, recent systematic reviews and meta-analyses have suggested that there are additional benefits from lowering BP even below standard goals (20,21). The eagerly awaited results of SPRINT showed that intensive systolic BP control (<120 mm Hg) was associated with lower rates of cardiovascular events and mortality compared with standard control (<140 mm Hg). In our study, adults with HTN who met intensive BP goals had the lowest predicted 10-year risk of CVD events on the basis of the Framingham risk score and the Korean CHD risk score, compared with those who met less intensive BP goals or had uncontrolled BP levels. When the actual event rates were estimated in the NHIS health examinee cohort, similar findings were reproduced.
In the SPRINT trial, the principal benefits in the intensive group were a reduction in CVD mortality and heart failure. There was only a modest, nonsignificant effect of the intensive intervention on MI or stroke. In our study, the opposite was true; there was significant difference in risk of MI or stroke but no difference in cardiovascular or all-cause mortality. Although the exact reasons remain unclear, this discrepancy might be explained in part by differences in study design, population characteristics, clinical practice pattern, or race or ethnic groups. The generalizability of the SPRINT experience to multiple groups of various ethnic backgrounds warrants further investigations and is likely to be of considerable interest.
We did not evaluate potential adverse effects that are attributed to more-intensive BP control. Although a marked cardiovascular benefit with more-intensive control was evident in the SPRINT trial, there were more frequent serious adverse events of hypotension, syncope, electrolyte abnormalities, and acute kidney injury (7). Other studies also suggested the potential negative impact of treating patients with mild HTN, which could result in serious complications in elderly people (22,23). Therefore, beyond the BP target per se, several important factors should be considered for optimal BP management in the contemporary medical setting; for example, an integrated and systematic assessment of combined risk factors and baseline cardiovascular risk (24), concomitant preventive medical therapies, cost-effectiveness (25), clinician–patient discussions of the potential benefits and harms, or the clinical judgment of the treating physicians (2,3).
First, the estimated percentages in Korean adults with HTN are based on the KNHANES sample weights. Thus, our findings rely on the accuracy and representativeness of the KNHANES data. Second, because our study was observational, overall findings are explorative and hypothesis-generating. Our observations cannot on their own support the idea that lowering BP will directly result in lower rates of cardiovascular events; BP targets and actual BP lowering may not be equivalent in practice, and, given the low event rates in our cohort, the analyses were possibly underpowered. Third, to assess the applicability of the SPRINT results, we used an identical threshold (systolic BP ≥130 mm Hg) for the definition of HTN, which is more inclusive than the conventional definition of systolic BP ≥140 mm Hg and diastolic BP ≥90 mm Hg. The estimated proportions of subjects who would meet BP goals are altered by changes of such definitions. Fourth, we cannot accurately quantify the effect of the SPRINT criteria or 2014 recommendations on patients who were currently receiving BP therapy because we could not determine which BP values had been used to initiate treatment. Lastly, the applicability of the SPRINT results to our general population might be questionable. On the basis of the NHANES results, 7.6% of U.S. adults and 16.7% of adults with treated HTN met the SPRINT eligibility criteria, suggesting a lack of generalizability to populations not examined in the study (e.g., those with diabetes or prior stroke, younger adults <50 years of age) (26). However, because clinicians often need to generalize results beyond a trial population, we assessed the broader sample of “real-world” adults with HTN who were free of SPRINT exclusion criteria.
The proportion of Korean adults with HTN who met BP target goals would substantially decrease under the SPRINT criteria compared with 2014 recommendations. The risk of major cardiovascular events seems lower in persons with more-intensive BP control who met the SPRINT goals than in those with less-intensive BP control or uncontrolled BP. Further investigations are required to determine the applicability and the potential impact of the results of SPRINT to a less restrictive, general population.
COMPETENCY IN MEDICAL KNOWLEDGE: The BP target supported by the results of SPRINT is substantially lower than that recommended in the 2014 HTN guidelines developed by the panel appointed to JNC 8. On the basis of a nationally representative sample in Korea, the proportion of adults with HTN satisfying the respective goals would decrease from 70.8% (7.8 million) under the 2014 guidelines to 11.9% (1.3 million) under the SPRINT criteria. The 10-year predicted risk of cardiovascular events using the Framingham risk score differed significantly as well: 6.15% in those with BP below the SPRINT goal, 7.65% in those with BP above the SPRINT goal but below the 2014 guideline goal, and 9.39% in those with uncontrolled HTN.
TRANSLATIONAL OUTLOOK: Further research is required to assess the generalizability of these ratios to other populations and determine how best to apply more stringent criteria for control of HTN in clinical practice.
For supplemental tables and figures, please see the online version of this article.
This study was supported by the National Evidence-based Healthcare Collaborating Agency, Seoul, Republic of Korea (project number NA-15-002). In addition, this study used National Health Insurance Service (NHIS)–National Sample Cohort data (NHIS-2015-2-113). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- blood pressure
- coronary heart disease
- confidence interval
- chronic kidney disease
- cardiovascular disease
- Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure
- Korea National Health and Nutrition Examination Survey
- myocardial infarction
- National Health Insurance Service
- Received March 14, 2016.
- Revision received March 28, 2016.
- Accepted March 29, 2016.
- American College of Cardiology Foundation
- Krakoff L.R.,
- Gillespie R.L.,
- Ferdinand K.C.,
- et al.
- Borden W.B.,
- Maddox T.M.,
- Tang F.,
- et al.
- Kweon S.,
- Kim Y.,
- Jang M.J.,
- et al.
- Lee J.,
- Lee J.S.,
- Park S.H.,
- et al.
- Jee S.H.,
- Jang Y.,
- Oh D.J.,
- et al.
- Bress A.P.,
- Tanner R.M.,
- Hess R.,
- et al.