Association of blood pressure control with atrial and ventricular ectopy in SPRINT

Abstract

Background

Previous epidemiological studies demonstrated that premature atrial contractions (PACs) and premature ventricular contractions (PVCs) detected by single 12-lead ECGs can predict incident cardiovascular disease and death. The determinants of cardiac ectopy remain unknown, with some evidence that hypertension may contribute.

Objective

To determine if intensive blood pressure (BP) control reduces the incidence of cardiac ectopy.

Methods

We performed a post-hoc analysis of the Systolic Blood Pressure Intervention trial, which randomised hypertensive participants to standard treatment (BP target <140 mm Hg) or intensive treatment (<120 mm Hg) with ECGs obtained at baseline, 2 years, 4 years and 5 years. The primary outcomes were incidence of ectopy (PACs or PVCs) as coded by Minnesota ECG classification, censoring for pacing, atrioventricular block, pre-excitation or atrial fibrillation/flutter. We performed Cox proportional hazards regression to determine the association of treatment group with outcomes.

Results

The analysis cohort comprised 3910 participants randomised to standard treatment and 3911 to intensive treatment, of whom 452 had ectopy on baseline ECG. After excluding those with baseline ectopy, there was no significant difference in the incidence of ectopy (incidence rate ratio 0.93 (95% CI 0.81 to 1.05)). There was no significant association between treatment group and ectopy incidence, with an unadjusted Cox HR of 0.93 (95% CI 0.82 to 1.07), and HR of 1 (95% CI 0.81 to 1.25) after adjusting for covariates.

Conclusion

Intensive BP control did not reduce the incidence of cardiac ectopy in patients with hypertension. Given the variable nature of PAC and PVC burden, further studies with continuous monitoring or more frequent sampling in larger populations are warranted.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Premature atrial and ventricular ectopic beats have known associative and likely causal relationships with atrial fibrillation, heart failure and cardiovascular hospitalisation. However, it is not known whether aggressive comorbidity management, including hypertension, can reduce the burden of these ectopic beats and potential subsequent morbidity.

WHAT THIS STUDY ADDS

• In a high-quality dataset of patients randomised to intensive versus standard blood pressure control in non-diabetic patients with at least one cardiovascular risk factor, intensive blood pressure control did not reduce the incidence of ectopic beats.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• While this study did not demonstrate an association between blood pressure control and ectopy on 12-lead ECGs, additional studies using continuous or ambulatory ECG monitoring are needed to identify relationships between modifiable risk factors and cardiac ectopy, which may be a preventable cause of cardiovascular morbidity.

Introduction

Premature atrial contractions (PACs) and premature ventricular contractions (PVCs) are commonly encountered in clinical practice, and their presence may indicate a tendency toward cardiovascular morbidity. In population-based cohort studies of individuals wearing ambulatory ECG monitors, PACs were associated with an increased risk of atrial fibrillation (AF) and stroke,^{1 2} while PVCs were associated with an increased risk of heart failure or death.^{3 4} Importantly, PACs and PVCs identified on even a single 12-lead ECG can predict an increased risk of AF, heart failure or death.⁵ Given evidence that successful ablation of PACs can prevent AF⁶ and that eradication of PVCs via ablation or medical therapy can normalise ejection fraction among individuals with systolic heart failure,⁷ it can be inferred that there is a possible causal relationship with these ectopic beats.

The aetiology of PAC and PVCs remains poorly understood. Hypertension is an important risk factor for AF, and, as PACs are such a potent predictor of AF,¹ it would seem plausible that higher blood pressure may promote PAC development. The few studies that have sought to determine predictors of these beats have suggested that prevention of the development of PACs and PVCs may therefore prove extraordinarily valuable, particularly if commonly used approaches in medicine might be applicable.^{7 8}

In the Systolic Blood Pressure Intervention (SPRINT) trial, intensive blood pressure control decreased fatal and non-fatal major cardiovascular events and death.⁹ However, the mechanisms of reduction in cardiovascular events are not fully understood. By leveraging the randomised controlled trial design, we therefore sought to mitigate confounders and determine whether intensive blood pressure might reduce the risk of incident PACs or PVCs.

Methods

We performed a non-pre-specified secondary analysis of the SPRINT trial, which randomised individuals with hypertension at risk of cardiovascular disease from 102 centres in the USA between November 2010 and August 2015 to either intensive or standard blood pressure control. At the time, it was not yet known whether more aggressive blood pressure control for individuals with hypertension would improve clinical outcomes and it was posed as the most important hypothesis to test by a 2007 National Heart, Lung and Blood Institute Expert panel. The study methods have been previously described.¹⁰

SPRINT enrolled patients at least 50 years of age with a systolic blood pressure between 130 and 180 mm Hg at risk of cardiovascular events, defined as one or more of: clinical or subclinical cardiovascular disease other than stroke, chronic kidney disease excluding polycystic kidney disease with estimated glomerular filtration rate of 20–60 mL/min per 1.73 m² of body-surface area, a 10-year risk of cardiovascular disease ≥15% by the Framingham risk score, or age 75 years or older. Those individuals with prior stroke or diabetes mellitus were excluded, the latter due to the ACCORD clinical trial which demonstrated intensive blood pressure control did not reduce the rate of cardiovascular events in hypertensive patients with diabetes.

In total, 9361 SPRINT participants were randomised in a 1:1 ratio to a systolic blood pressure target of either <120 mm Hg (intensive treatment group) or <140 mm Hg (standard treatment group) prior to exclusions. During SPRINT, 12-lead ECGs were obtained at baseline, 2 years, 4 years and the closeout visit at year 5. For each measurement, ECG abnormalities were over-read at a core facility blinded to randomisation assignment and classified using Minnesota ECG Classification. Of the 9361 participants, we excluded those who either did not have ECG data available or those with less than two study ECGs. To focus on the incidence of cardiac ectopy, those with any cardiac ectopy on the baseline ECG were excluded. For PAC-specific analyses, participants were excluded if a PAC was present on the baseline ECG; for PVC-specific analyses, participants were excluded if a PVC was present on the baseline ECG (cohort selection diagram in online supplemental figure 1).

The primary predictor was treatment assignment group (standard treatment or intensive treatment). The primary outcome of interest was any incident cardiac ectopy, defined as the presence of incident PACs or PVCs. Ectopic beats were identified by Minnesota ECG codes (8.1.1—PACs; 8.1.2—PVCs; 8.1.3—PACs and PVCs). Secondary outcomes included individual incidence of PVCs (8.1.1+8.10.3) and individual incidence of PACs (8.1.2+8.10.3). For all outcomes, participants were censored at (1) occurrence of outcome, (2) time of last ECG, (3) time of any ECG reporting pacemaker presence (Minnesota code 6.8) or (4) time of any ECG with ventricular pre-excitation (Minnesota code 6.4). For the PAC-based outcomes, we added additional censoring events for (1) complete/third-degree AV block (Minnesota code 6.1) or (2) AF or flutter (Minnesota code 8.3).

Demographic differences were assessed with the χ² test and two-sample t-test for categorical and continuous variables, respectively. We used multivariate logistic regression to evaluate the association of baseline characteristics with patients who had ectopic beats on their baseline ECG. Raw event rates and incidence rates per 1000 person-years were determined by treatment group. We then calculated incidence rate ratios (IRR) using Poisson regression and used Cox proportional hazards regression models to determine the association between treatment group and incidence of ectopic beats. Multivariable Cox proportional hazards models included covariates found to be statistically significant in comparing those with and without baseline (or prevalent) ectopy. Statistical analyses were performed using STATA V.18 (StataCorp).

Results

Of the 9361 initial SPRINT participants, there were 9028 who (1) did not withdraw consent and (2) had a baseline ECG performed. Of these participants, 733 had ectopic beats on their baseline ECG; 457 had PACs, 241 had PVCs and 35 had both PACs and PVCs. Participants with baseline ectopic beats were significantly more likely to be older, male, non-Hispanic or have smoked (baseline characteristics in online supplemental table 1). After covariate adjustment, age (OR 1.06 per year, 95% CI 1.03 to 1.09, p<0.001) and male sex (OR 1.81, 95% CI 1.3 to 2.52, p<0.001) had significant associations with the presence of ectopic beats on baseline ECG.

Of the 9028 eligible SPRINT participants noted above, 7821 participants with at least 2 ECGs were included in the analysis cohort (3910 standard treatment, 3911 intensive treatment; median follow-up 3.5 years). Baseline characteristics were well matched between the two groups, including for age, body mass index, gender, baseline smoking status and comorbid disease (table 1). Of the 7369 participants in the analysis cohort without ectopy on baseline ECG, incident ectopy occurred in 452 participants in the standard treatment group, corresponding to an incidence rate (IR) of 9.52 per 1000 person-years, compared with 418 participants in the intensive treatment group (IR 8.71 per 1000 person-years). There was no significant difference in the incidence of ectopy, with an IRR of 0.93 (95% CI 0.81 to 1.05). There was no significant association between treatment group and ectopy, with an unadjusted Cox HR of 0.93 (95% CI 0.82 to 1.07), and an adjusted HR of 1 (95% CI 0.81 to 1.25) and after adjusting for age, body mass index and baseline comorbidities. In the multivariable model, age was significantly associated with the incidence of ectopy, with an HR of 1.06 (95% CI 1.04 to 1.08, p<0.001) per additional year of age (figure 1).

Table 1. Baseline characteristics of analysis cohort.

	Standard treatment (n=3910)	Intensive treatment (n=3911)	P value
Age	67.6±9.2	67.6±9.2	0.78
Body mass index (kg/m2)	30.5±5.7	30.8±5.8	0.19
Gender			0.58
Female	1385 (35.4%)	1409 (36%)
Male	2525 (64.6%)	2502 (64%)
Race			0.19
White	2541 (65%)	2560 (65.5%)
Black	1236 (31.6%)	1187 (30.4%)
Asian	26 (0.7%)	41 (1%)
Other	84 (2.1%)	102 (2.6%)
Multiracial	22 (0.6%)	21 (0.5%)
Hispanic	412 (10.5%)	421 (10.8%)	0.74
Smoking status			0.34
Never	1778 (45.5%)	1786 (45.7%)
Current	494 (12.6%)	512 (13.1%)
Former	1638 (41.9%)	1613 (41.2%)
Atrial fibrillation	279 (7.1%)	320 (8.2%)	0.08
Cancer	463 (11.8%)	496 (12.7%)	0.27
Vascular disease*	697 (17.8%)	742 (19%)	0.19
Serum creatinine, mg/dL	1.06±0.36	1.07±0.34	0.67

^*Defined as patient-reported history of heart attack, angina or peripheral vascular disease.

Figure 1. Association of Blood Pressure Control with Atrial and Ventricular Ectopy on 12-leadECGs. In the Systolic Blood Pressure Intervention Trial (SPRINT) there was no significant association between the blood pressure treatment group and the incidence of premature atrial contractions (PACs) or premature ventricular contractions (PVCs) on routine monitoring 12-lead ECGs.

Premature atrial contractions

Of the 7520 participants in the analysis cohort without PACs on baseline ECG, we compared 3765 participants receiving standard treatment to 3755 participants receiving intensive treatment (baseline characteristics were still well balanced; online supplemental table 2). In the standard treatment group, there were 298 participants who had at least one ECG with a PAC, corresponding to an incidence rate of 6.03 per 1000 person-years, compared with 289 participants with an incidence rate of 5.81 per 1000 person-years in the intensive treatment group. There was no significant difference in the incidence of PACs, with an IRR of 0.96 (95% CI 0.82 to 1.14), an unadjusted HR of 0.98 (95% CI 0.83 to 1.16) and an HR of 1.05 (95% CI 0.77 to 1.43) after adjusting for age, body mass index and baseline comorbidities. In the multivariable model, there was a significant association between age and incident PACs, with an HR of 1.07 (95% CI 1.05 to 1.1, p<0.001) per additional year of age.

Premature ventricular contractions

Of the 7704 participants in the analysis cohort without PVCs on baseline ECG, we compared 3844 participants receiving standard treatment to 3860 participants receiving intensive treatment (baseline characteristics well balanced; online supplemental table 3). The incidence of PVCs was 4.31 per 1000 person-years in the standard treatment group (213 participants) and 4.31 in the intensive treatment group (214 participants). There was no significant difference in the incidence of PVCs, with an IRR of 0.97 (95% CI 0.79 to 1.19), an unadjusted HR of 0.98 (95% CI 0.8 to 1.2) and adjusted HR of 1.13 (95% CI 0.86 to 1.48). In the multivariable model, there was a significant association between age and incident PVCs, with an HR of 1.04 (95% CI 1.01 to 1.08, p=0.003) per additional year of age.

Discussion

In this secondary analysis of individuals with hypertension at risk of cardiovascular events, we found no significant association between intensity of blood pressure control and the incidence of PACs or PVCs on 12-lead ECGs. As observed in prior studies, age was strongly predictive of both prevalent and incident ectopy.

Previous studies have demonstrated an association between atrial and ventricular ectopy with AF, heart failure and death,^{4 5 11} but the aetiology and determinants of the presence of these beats remain largely unknown. In an earlier sub-study of SPRINT, intensive blood pressure control reduced the incidence of AF,¹² and the recently concluded ARREST-AF trial demonstrated that lifestyle and risk factor management including blood pressure control helped reduce recurrence of AF following catheter ablation.¹³ Therefore, if intensive blood pressure control reduced the incidence of ectopy, this could have served as a mediator for reduction in AF. Although hypertension has been observed to be a risk factor for ectopy, no study has yet demonstrated an intervention (such as more aggressive blood pressure control) that might influence the risk of common cardiac ectopy.^{14 15} However, we did not find a significant association between intensity of blood pressure control and incident PACs or PVCs.

There are multiple plausible explanations for this finding. First, given that 12-lead ECG measurements were performed a maximum of 4 times over the 5-year study period, we may be most likely to capture individuals with high burden PACs or PVCs. PAC burden is known to increase with left atrial enlargement and diminished left atrial function,¹⁶ and therefore the ECG cadence may only capture those individuals who already have manifest cardiac dysfunction. The most clinically relevant burdens of ectopy may be at the extremes (such as 20% of all beats, which would translate into some ectopy on a 10-second ECG), although such frequencies are likely rare in the general population.^{1 4 7} In such a case, serial ambulatory ECG monitoring may be better suited to capture the progression of ectopy as an indicator of subclinical disease. Second, various factors affect daily PAC and PVC burden, including but not limited to caffeine intake, physical activity and tobacco consumption,^{14 17 18} with unknown individual contributions at the time of 12-lead ECG ascertainment. Third, per the SPRINT trial protocol, beta-blockers and calcium-channel blockers were both recommended as available agents for blood pressure control, and patients in the intensive treatment group may have reached higher doses of these medications. These medications may have suppressive effects on PACs and PVCs^{8 19} and may therefore mask a potential signal. It is worth noting that the same trial has been used to demonstrate statistically significant differences in other ECG findings,²⁰ which suggests adequate power to detect such relationships, although the intermittent nature of PACs and PVCs likely reduced the chance of observing a difference.

There are clear benefits to reduction in PAC and PVC burden in preventing cardiovascular disease, and beyond immutable factors such as age and sex, or medication-based suppression, identifying reversible drivers or modifiable triggers (such as physical activity) may impact cardiovascular health and inform effective patient counselling. Simultaneously, well-designed studies to frame the effects of common lifestyle habits such as caffeine consumption¹⁸ are just as important for patient quality of life. A stepwise approach with high-quality data, where available, can serve as a nidus for impactful prospective studies. Focused clinical and lifestyle interventions identified by these studies could then inform clinical practice guidelines and quality improvement initiatives.

Limitations

Our study has limitations. First, this was a retrospective analysis, and SPRINT was not powered to compare rates of ectopic beats on 12-lead ECGs or designed to study the incidence of ectopy, which may be better assessed with continuous ECG monitoring. Second, SPRINT was limited to non-diabetic patients deemed to be at risk of cardiovascular disease, and so the findings may not generalise to all patients. Third, even SPRINT patients in the standard treatment arm may have had better overall blood pressure control than those in real-world populations, which may have limited the relative effect size on atrial and ventricular ectopy when compared with intensive treatment. Fourth, while some data are available about initial anti-hypertensive regimen choices, we are unable to account for the shifts in anti-hypertensive use and relative doses between patients, which may have unknown effects on long-term incidence of cardiac ectopy. Finally, although the current study leveraged randomised assignment of the predictor, exclusion of prevalent ectopy in order to focus on incident outcomes may have resulted in unmeasured confounding or residual confounding not sufficiently addressed with multivariable analyses.

Conclusions

In this post-hoc analysis of the SPRINT study, intensive blood pressure control was not associated with a reduction in incident atrial or ventricular ectopic beats on serial 12-lead ECGs. Atrial and ventricular ectopy are established predictors of cardiovascular morbidity and mortality, and additional large-scale studies with continuous monitoring data are needed to further characterise the role of intervenable risk factors.

Data availability statement

Data may be obtained from a third party and are not publicly available.