Secondary Logo

Journal Logo

Is Concurrent Training Efficacious Antihypertensive Therapy? A Meta-analysis


Medicine & Science in Sports & Exercise: December 2016 - Volume 48 - Issue 12 - p 2398–2406
doi: 10.1249/MSS.0000000000001056

Aerobic exercise training and, to a lesser degree, dynamic resistance training, are recommended to lower blood pressure (BP) among adults with hypertension. Yet the combined influence of these exercise modalities, termed concurrent exercise training (CET), on resting BP is unclear.

Purpose This study aimed to meta-analyze the literature to determine the efficacy of CET as antihypertensive therapy.

Methods Electronic databases were searched for trials that included the following: adults (>19 yr), controlled CET interventions, and BP measured pre- and postintervention. Study quality was assessed with a modified Downs and Black Checklist. Analyses incorporated random-effects assumptions.

Results Sixty-eight trials yielded 76 interventions. Subjects (N = 4110) were middle- to older-age (55.8 ± 14.4 yr), were overweight (28.0 ± 3.6 kg·m−2), and had prehypertension (systolic BP [SBP]/diastolic BP [DBP] = 134.6 ± 10.9/80.7 ± 7.5 mm Hg). CET was performed at moderate intensity (aerobic = 55% maximal oxygen consumption, resistance = 60% one-repetition maximum), 2.9 ± 0.7 d·wk−1 for 58.3 ± 20.1 min per session for 19.7 ± 17.8 wk. Studies were of moderate quality, satisfying 60.7% ± 9.4% of quality items. Overall, CET moderately reduced SBP (db = −0.32, 95% confidence interval [CI] = −0.44 to −0.20, −3.2 mm Hg) and DBP (db = −0.35, 95% CI = −0.47 to −0.22, −2.5 mm Hg) versus control (P < 0.01). However, greater SBP/DBP reductions were observed among samples with hypertension in trials of higher study quality that also examined BP as the primary outcome (−9.2 mm Hg [95% CI = −12.0 to −8.0]/−7.7 mm Hg [95% CI = −14.0 to −8.0]).

Conclusions Among samples with hypertension in trials of higher study quality, CET rivals aerobic exercise training as antihypertensive therapy. Because of the moderate quality of this literature, additional randomized controlled CET trials that examine BP as a primary outcome among samples with hypertension are warranted to confirm our promising findings.

Supplemental digital content is available in the text.

1Department of Kinesiology, University of Connecticut, Storrs, CT; 2Institute for Collaboration on Health, Intervention and Policy (InCHIP), University of Connecticut, Storrs, CT; 3Department of Kinesiology, The University of Alabama, Tuscaloosa, AL; 4Department of Psychological Sciences, University of Connecticut, Storrs, CT; 5Instituto de Educação Física e Desportos, Universidade do Estado do Rio de Janeiro, Rio de Janeiro, BRAZIL; 6Henry Low Heart Center, Department of Cardiology, Hartford Hospital, Hartford, CT

Address for correspondence: Lauren M. L. Corso, M.S., M.L.S. (ASCP)CM, Department of Kinesiology, University of Connecticut, Gampel Pavilion Room 206, 2095 Hillside Rd., U-1110, Storrs, CT 06269-1110; E-mail:

Submitted for publication April 2016.

Accepted for publication July 2016.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (

Hypertension is the most common, costly, and preventable cardiovascular disease (CVD) risk factor, affecting one in three (80 million) American adults (29). Adults with hypertension are at disproportionate risk for developing CVD and are three to four times more likely to die from CVD than those without hypertension (13,29). Since 2010, the incidence of hypertension has not improved, underscoring the need for cost-effective, sustainable lifestyle intervention strategies to prevent, treat, and control hypertension that include regular participation in exercise (33).

Aerobic exercise training lowers blood pressure (BP) 5–7 mm Hg, whereas dynamic resistance training lowers BP 2–3 mm Hg among adults with hypertension (31). Even modest reductions in BP of ~5 mm Hg can reduce the risk of heart disease by 8% and stroke by 14%, substantiating the clinical importance of exercise as an antihypertensive lifestyle therapy (3). Accordingly, the American College of Sports Medicine (ACSM) recommends 30 min of moderate-intensity aerobic exercise on most, preferably all, days of the week supplemented by dynamic resistance training 2–3 d·wk−1 as an antihypertensive lifestyle therapy (31).

Despite the ACSM recommendations, it is not well understood how the combined effect of aerobic and dynamic resistance exercise training, termed concurrent exercise training (CET), influences resting BP among adults with high BP. More specifically, CET is defined as aerobic and resistance training performed in close proximity to each other (i.e., in a single session or on separate days) (10,25). Nonetheless, primary level CET trials often do not disclose the proximity of the aerobic and resistance exercise components, nor do they describe the order in which they are applied (i.e., aerobic performed before vs after resistance exercise). Thus, the working definition of CET remains loosely characterized, which may contribute to the inconsistencies in this literature (4,8,17,30,44).

To date, five meta-analyses (4,8,17,30,44) have examined the BP response to CET and report SBP/DBP reductions ranging from 0 to 4 mm Hg. These meta-analyses examined a variety of clinical populations; Chudyk et al. (4), Hayashino et al. (17), and Zou et al. (44) included adults with type 2 diabetes mellitus; Pattyn et al. (30) included those with the metabolic syndrome; and Cornelissen et al. (8) included adults absent of CVD or other chronic conditions. Cornelissen et al. (8), Chudyk et al. (4), and Pattyn et al. (30) included trials that examined CET only, whereas Hayashino et al. (17) and Zou et al. (44) examined CET studies that also involved dietary cointervention. Finally, only Cornelissen et al. (8) and Hayashino et al. (17) examined resting BP as a primary outcome. No meta-analyses focused exclusively on adults with hypertension per se; nonetheless, three did report the baseline BP of the participants. Of those meta-analyses that disclosed the baseline BP of their sample, only ~30% (n = 9) involved adults with hypertension (8,17,30). Despite four of the five meta-analyses reporting a measure of heterogeneity (8,17,30,44), moderator analyses were rarely performed in an attempt to explain the variability in the BP response to CET. Finally, these meta-analyses did not quantify the CET frequency, intensity, time, and type (FITT) of the intervention, nor did they examine how the proximity or order of the aerobic and resistance exercise components were implemented and how they may have influenced the BP response to CET (4,8,17,30,44).

Therefore, the purposes of our meta-analysis were to determine the efficacy of CET as antihypertensive therapy and to examine important potential moderators of the BP response to CET.

Back to Top | Article Outline


This meta-analysis is reported consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement (28).

Back to Top | Article Outline

Search Strategy and Selection Criteria

Aided by a medical librarian (JL), multiple electronic databases (PubMed, Cumulative Index to Nursing and Allied Health Literature, Web of Science, SportDiscus, and Scopus) were searched from their inception until January 31, 2015. The full search strategy appears in Supplemental Digital Content (see Document, Supplemental Digital Content 1, Systematic Search Terms and Full Search Strategy for each Electronic Database, Qualifying CET trials included the following: adult populations (>19 yr old), had a nonexercise control or comparison group, measured BP pre- and postintervention for the CET and control groups, and reported the FITT of the CET intervention. Cross-sectional studies, epidemiological study designs, weight loss, or pharmacological trials were excluded. Because hypertension rarely occurs in isolation (i.e., clusters with metabolic or CVD risk factors) (24), only studies that included populations with disease(s) or health conditions unrelated to CVD (e.g., cancer, HIV/AIDS) were excluded.

Potentially relevant reports were screened in triplicate (LML, HVM, and AB) first by title, then title and abstract, and last by full-text review. Reference lists of included studies, recent reviews, and meta-analyses were searched for additional qualifying reports.

Back to Top | Article Outline

Data Extraction: Coding and Reliability

Coded variables were extracted using a standardized coding form and coder manual previously developed by a team of experts (LSP, BTJ, and TBHM) and pilot tested. Two trained coders (LML and AB) independently extracted and entered study information with high reliability across all dimensions (mean Cohen’s k = 0.83 and Pearson’s r = 0.87) (18). Coding disagreements were resolved through discussion with a third independent party (HVM and LSP). Data were extracted as reported by study authors and included study characteristics (e.g., randomization and methodological study quality), sample characteristics (e.g., age and baseline BP), and features of the CET intervention (e.g., FITT).

Back to Top | Article Outline

Data Extraction: Methodological Study Quality

Methodological study quality was assessed with a modified version of the Downs and Black Checklist (12) and gauged as a percentage of items satisfied out of a possible 33-point total. This checklist has been widely used for health-related outcomes and is considered one of the most comprehensive quality assessment tools available (9). Details of the modified version of the Downs and Black Checklist appear in Supplemental Digital Content 2 (see Document, Supplemental Digital Content 2, Modified Downs and Black Checklist, Study quality was quantified as low (≤16 points, ≤50%), moderate (>16 to 23, 50% to 69%), or high (≥24, ≥70%) (5). Overall methodological study quality, study quality subscales (i.e., reporting, external validity, bias, confounding, and power), and individual quality items (i.e., BP-focused primary outcomes) were examined in analyses to determine their influence on the BP response to CET.

Back to Top | Article Outline

Effect Size Calculations

The standardized mean difference effect size (between-group, db) was used to quantify the effectiveness of CET as antihypertensive therapy, defined as the mean difference in resting SBP/DBP between the CET and the control groups post- versus preintervention divided by the pooled SD, correcting for small sample size bias and baseline differences (1,18). We disaggregated comparisons for trials with more than one CET intervention (1). Effect sizes were calculated for each CET and control group and analyzed separately for the different comparisons.

Twenty-five CET trials involved an “active content” control group; that is, the control group was randomized to usual care for chronic disease, minimal effect stretching routines, relaxation, or educational information sessions on health rather than a nonexercise wait-list control. To examine the potentially confounding effects that an active content control group (15) may have had on the between-group, db, effect size, we examined the within-group, dw, effect size, defined as the mean BP difference in resting BP post- versus preintervention for the independent CET and control group, divided by the preintervention SD (1,21).

Negative d values were set to indicate greater BP reductions were observed for the CET compared with the control group (i.e., between-group, db) or relative to baseline (i.e., within-group, dw). The magnitude of d was interpreted as 0.20, 0.50, and 0.80 for small, medium, and large BP reductions, respectively (7). Last, we transformed d values arithmetically to provide the equivalent BP change in millimeters of mercury for ease of clinical interpretation.

Back to Top | Article Outline

Publication Bias

We assessed the potential for publication bias using funnel plots (i.e., observed d values were plotted in contrast to their standard errors), which were visually examined for asymmetry and the presence of outliers. In addition, the methods for detecting publication bias of Begg and Mazumdar (2) and Egger et al. (14) were used to further examine suspected asymmetry. Two effect sizes from the same study (11), one for SBP and one for DBP, were determined to be outliers (i.e., more than three SD values than the mean). To reduce their influence on our results, they were windsorized to be the same magnitude as the next largest SBP and DBP effect sizes (16). Contour-enhanced funnel plots of the distribution of the within-group (dw) effect sizes appear with and without winsorization for SBP and DBP in Supplemental Digital Content 3 (see Figure, Supplemental Digital Content 3, Contour Enhanced Funnel Plots of Windsorized and Unwindsorized Effect Sizes,

Cochran’s Q (i.e., the Q statistic) was used to determine the presence (or absence) of heterogeneity (6). Higgins’s I2 statistic and corresponding 95% confidence interval (CI) were used to gauge the degree or extent of heterogeneity present in the sample (19,20). I2 values range from 0% (homogeneity) to 100% (greater heterogeneity); a CI that does not include 0% indicates that the hypothesis of homogeneity is rejected and an inference of heterogeneity is merited (19,20).

Back to Top | Article Outline

Moderator Analysis

In the presence of significant heterogeneity, separate moderator analyses were used to explain variability in the BP response to CET and control by examining theoretically driven a priori moderators. These moderators included the following: study characteristics (e.g., randomized vs nonrandomized controlled trial, number of CET arms, and methodological study quality), sample characteristics (e.g., age, medication use, and baseline BP), and features of the CET intervention (e.g., FITT and the order and proximity of aerobic and resistance exercise). Weighted regression models with a maximum likelihood estimation of the random-effects weights, the inverse variance for each dw, were used to explain variability in dw for SBP and DBP for the CET and control interventions. All interaction terms were generated with moderators and tested for significance. Continuous moderators were mean centered, and categorical variables were contrast coded before generating interaction terms or performing multiple moderator analyses (22). Windsorized within-group, dw, effect sizes for the CET and the nonexercise control or comparison group were evaluated separately in moderator analyses.

Back to Top | Article Outline

Multiple moderator models

We did not rely only on significant bivariate meta-regression analyses when examining a priori moderators in multiple moderator models. Significant or trending moderators were examined in conjunction with the model coefficient and R2 values (i.e., between-study variance explained by covariate) to examine the influence of individual moderators on the BP response to CET (43). To best minimize the influence of any methodological differences between trials, overall methodological study quality, study quality subscales, or individual quality (e.g., BP-focused primary outcomes) dimensions were controlled for and incorporated into multiple moderator models when feasible (22,43).

Back to Top | Article Outline

The moving constant technique

The moving constant technique (22) was applied to estimate the magnitude of the weighted mean effect sizes (dw+) at different levels of interest for individual moderators, including extreme values, and clinical thresholds (e.g., BP classification) (3). These estimates, or predicted

values, and their CIs statistically control for the influence of other moderators in the model, holding them constant at their mean values.

Back to Top | Article Outline

Additive models

For SBP and DBP, additive models were generated from the final multiple moderator models that represented the greatest potential antihypertensive benefit resulting from CET. Individual moderators were assessed within the same model at the level that yielded the greatest BP reduction (i.e.,

and 95% CI). Overall, these models identified the ideal sample or study profile that results in optimal antihypertensive therapy.

Back to Top | Article Outline

Statistical Computing

Results are presented as mean ± SD unless otherwise indicated. Differences in baseline characteristics between the CET and the control groups were examined using t-tests and one-way ANOVA. Analyses were performed using Stata 13.1 (Stata Corp, College Station, TX) (40) with macros for meta-analysis (26) and incorporated random-effects assumptions. Significance was set at P < 0.05.

Back to Top | Article Outline


Study characteristics

In total, 68 trials qualified for inclusion in our meta-analysis. Eight trials included multiple CET arms, yielding 76 interventions (k). The systematic search and selection process of the included CET trials is shown in Figure 1. A list of included trials is available in Supplemental Digital Content 4 (see Document, Supplemental Digital Content 4, Supplementary Reference List, Most of the included CET trials were randomized controlled trials (84.2%, k = 64) with only a minority examining BP as a primary outcome (14.5%, k = 11). Of note, 33% (k = 25) of interventions used an “active content” control comparison, whereas the remaining interventions (67%, k = 51) used a nonexercise control group (Table 1). A general description of each CET and control intervention appears in Supplemental Digital Content 5 (see Table, Supplemental Digital Content 5, Summary Table of Included Concurrent Exercise and Control Interventions,





Overall, included trials were of moderate methodological study quality, satisfying 60% or 20 points on the augmented Downs and Black Checklist, indicating this literature displayed considerable variability in scores (39.4% to 84.8%, 13 to 28 points). Only five trials were considered of higher quality (≥70% items satisfied). Of the five study quality subscales, trials were most likely to satisfy external validity (93.9% ± 21.56%), confounding (66.0% ± 16.8%), and bias (70.1% ± 10.4%) with substantial deficiencies in reporting (53.5% ± 17.3%) and power (12.5% ± 26.0%).

Back to Top | Article Outline

Sample characteristics

More than half of the CET interventions (k = 43) included samples that were absent of CVD-related chronic disease or health conditions, whereas the remaining interventions (k = 33) reported that they included populations with known chronic disease(s) and health conditions related to CVD that included type 2 diabetes, CVD, metabolic syndrome, and chronic kidney disease, or a combination of these chronic diseases and health conditions. Baseline sample characteristics did not differ between the CET and the nonexercise control or comparison groups (P > 0.05) (Table 2).



Back to Top | Article Outline

Features of the CET interventions

On average, CET was performed 2.9 ± 0.7 d·wk−1 at moderate intensity (aerobic exercise training = 55% maximal oxygen consumption; dynamic resistance training = 60% of one-repetition maximum), 58.3 ± 20.1 min per session for 19.7 ± 17.8 wk (Table 3). More than half of the included CET interventions failed to report the order (65.8%, k = 50) and proximity (57.9%, k = 44) of the individual aerobic and resistance exercise components. Of those that disclosed these details, 7.9% (k = 6) performed aerobic and resistance exercise on separate days (i.e., “combined training”). The remainder of interventions performed both modalities on the same day using the following protocols: circuit training (i.e., alternating between aerobic and resistance exercise) (6.6%, k = 5), aerobic first followed by resistance exercise (15.8%, k = 12), or resistance first followed by aerobic exercise (11.8%, k = 9). The majority of the CET interventions were supervised (64.5%, k = 49); 9.2% of the interventions incorporated a combination of supervised and unsupervised exercise, and 2.6% were unsupervised (k = 2). Eighteen (23.4%) interventions failed to report the level of supervision during the CET sessions. Adherence to the CET intervention ranged from 55% to 100% with an average attrition rate of 84.9%.



Back to Top | Article Outline

BP assessment methods

BP was most commonly assessed using automated methods (34.2%, k = 26) in the seated (25%, k = 19) position. Unfortunately, 46.1% (k = 35) of the trials failed to disclose details of the BP measurement. CET interventions rarely reported the use of standard laboratory measurement protocols (3.6%, k = 3), and only one study assessed BP under ambulatory conditions (36). Details related to BP assessment methods are available in Supplemental Digital Content 6 (see Table, Supplemental Digital Content 6, Blood Pressure Assessment Methodology Table,

Back to Top | Article Outline

The antihypertensive effects of CET

For the between-group comparisons, CET lowered SBP 3.2 mm Hg (db+ = −0.32, 95% CI = −0.44 to −0.20) and DBP 2.5 mm Hg (db+ = −0.35, 95% CI = −0.47 to −0.22) compared with control. Collectively, these effect sizes lacked homogeneity (I2 [95% CI]: SBP = 68.6% [60.3–75.2]; DBP = 65.7% [55.8–73.4]).

For the within-group (dw+) comparisons, CET lowered SBP 3.6 mm Hg (dw = −0.36, 95% CI = −0.44 to −0.27) and DBP 3.9 mm Hg (dw = −0.39, 95% CI = −0.49 to −0.29) post-versus preintervention. The nonexercise control or comparison group lowered SBP 1.1 mm Hg (dw+ = −0.11, 95% CI = −0.18 to −0.22), but DBP was not different post- versus preintervention (P > 0.05). The distribution of these effect sizes displayed significant heterogeneity for both the CET (I2 [95% CI]: SBP = 57.0% [44.4–66.7]; DBP = 70.1% [61.8–76.6]) and the control or comparison (SBP = 43.3% [25.3–57.0]) groups. The mean between-group (db+) and within-group (dw+) effect sizes and homogeneity statistics appear in Supplemental Digital Content 7 (see Table, Supplemental Digital Content 7, Table of Mean Effect Sizes,

Back to Top | Article Outline

Multiple moderator analyses

Multiple moderator models for SBP and DBP appear in Tables 4 and 5, respectively. SBP was reduced more among samples with hypertension (5.3 mm Hg) compared with prehypertension (2.9 mm Hg) and those with normal BP (+0.9 mm Hg) (P = 0.01). Furthermore, even larger SBP reductions were observed among samples with hypertension when trials examined BP as a primary study outcome (7.6 mm Hg) versus those that did not (3.1 mm Hg) (P = 0.01). We also observed greater SBP reductions among studies of higher (3.6 mm Hg) than lower (1.9 mm Hg) study quality (P = 0.01).





Additive model: CET elicited the greatest potential SBP reductions (9.2 mm Hg, 95% CI = 12.0–8.2) among samples with hypertension in higher-quality trials that examined BP as a primary outcome. These effects were not influenced by the order (i.e., aerobic performed before vs after resistance exercise) or proximity (i.e., in a single session or on separate days) of the aerobic and resistance exercise components but these were controlled in our final models (see Additive Model, Table 4). Among the control or comparison groups, greater SBP reductions were observed among the “active content” (1.9 mm Hg) than the nonexercise or wait-list control groups (0.0 mm Hg) (P = 0.01).

DBP was reduced to the greatest extent among samples with hypertension (5.6 mm Hg) compared with prehypertension (3.6 mm Hg) and those with normal DBP (1.5 mm Hg) (P = 0.01). Greater DBP reductions were also observed among studies of higher (4.8 mm Hg) than lower (2.3 mm Hg) study quality (P = 0.013).

Additive model: CET elicited the greatest potential DBP reductions among samples with hypertension in higher-quality trials (7.7 mm Hg, 95% CI = 14.0–8.3). Consistent with SBP, neither the order nor the proximity of the aerobic and resistance exercise components (see Additive Model, Table 5) proved statistically significant. Consistent with SBP, neither the order nor the proximity of the aerobic and resistance exercise components influenced the BP response to CET but were controlled for in our final models.

Back to Top | Article Outline


The purposes of our meta-analysis were to investigate the efficacy of CET as antihypertensive therapy and to examine important potential moderators of the BP response to CET. On average, we found that moderate-intensity CET performed ~3 d·wk−1 for ~60 min per session modestly reduced BP on average ~3 mm Hg compared with control. Notably, we found that BP reductions after CET exhibited a dose–response relationship such that SBP/DBP reductions were greatest among samples with hypertension (5/6 mm Hg) than with prehypertension (3/4 mm Hg) and normal BP (+1/2 mm Hg). The change in BP was examined as a primary study outcome versus studies in which it was not (for SBP only: 8 vs 3 mm Hg), and among trials of higher than lower methodological quality (4/5 vs 2/2 mm Hg). The greatest potential BP-lowering effects from CET occurred among samples with hypertension in trials of higher methodological quality focusing on BP as a primary outcome (~8–9 mm Hg).

Our findings demonstrate that CET appears to lowers BP to levels that are comparable with or even greater than aerobic exercise training for adults with hypertension (~5–9 mm Hg) (33,34). Of note, three trials in our meta-analysis directly compared the antihypertensive effects of aerobic exercise to CET and reported no group differences (38,39,41). The addition of resistance training to aerobic exercise (e.g., CET) may allow for simultaneous benefit in the cardiometabolic profile, cardiorespiratory fitness, and muscle endurance and strength for overweight adults (27). As a result, it appears that CET may be a more comprehensive approach to achieving health benefits than aerobic training alone. These results suggest that the current ACSM exercise recommendations for hypertension (35) should be expanded to include specifics of the FITT exercise prescription for CET.

Our meta-analysis is the first to examine methodological study quality as a moderator of the BP response to CET. Unlike previous meta-analyses (4,8,17,30,44), we quantitatively examined methodological study quality independently and interactively with other moderators to best determine the influence of study quality dimensions on the BP response to CET. We found greater BP reductions in studies of higher (≥70% items satisfied) than lower (≤50% items satisfied) methodological study quality (~5 vs ~2 mm Hg). Yet we found the quality of this literature to be of only moderate quality, leaving open the possibility that other potential risks of sample bias and threats to validity exist. One potential source of bias in this literature is that few interventions specifically examined the antihypertensive effects of CET. In fact, only 14% (k = 11) of interventions examined BP as a primary study outcome. Among these studies, BP reductions were more than twice the magnitude of those reported in trials with BP as a secondary outcome (~8 vs 3 mm Hg). Furthermore, because of poor disclosure of BP assessment methods, it is unclear how the BP assessment methodology may have influenced the BP response to CET (see Table, Supplemental Digital Content 5, Summary Table of Included Concurrent Exercise and Control Interventions,

Of note, one-third of the studies in our sample incorporated “active content” (32.9%, k = 25) versus “nonexercise” or wait-list control groups (69.1%, k = 51). Those with the active content comparison groups experienced small but significant SBP reductions of ~1 mm Hg, whereas DBP was not different pre- to postintervention among the “nonexercise” or wait-list control groups (0 mm Hg). The use of “active content” control groups is paradoxical because they may themselves act as interventions and may not be appropriate to use for the targeted comparisons (15). As a result, their use generally reduces the effectiveness of the intervention being studied, as gauged by between-group comparisons (15). Indeed, CET elicited BP reductions of ~1 mm Hg when compared with the active content control groups and ~5 mm Hg when compared with the nonexercise or wait-list control groups. We evaluated other potential sources of bias (i.e., randomized controlled trials vs not, number of CET interventions) and found that these did not influence the BP response to CET. Nonetheless, we acknowledge that additional unidentified sources of bias that may have affected BP outcomes may exist in this literature.

There were limitations to our meta-analysis. There were no FITT characteristics that emerged as moderators of the BP response to CET in our meta-analysis, which may be due to 86% (k = 6) of the included interventions that failed to fully disclose this important information. Moreover, nearly 70% (k = 26) of the primary level CET trials in our meta-analysis did not include samples with hypertension, calling into question the generalizability of this literature to adults with hypertension. Furthermore, only one study (42) reported the timing of the BP measurement after the last exercise training session; therefore, it is unclear if the authors appropriately isolated acute exercise effects from the more long-term training effects of CET on BP (32–34). These observations illustrate the need for additional primary level trials that address these limitations so that more definitive conclusions can be made about the antihypertensive effects of CET.

Our meta-analysis has important strengths. The trial selection process was comprehensive, including multiple databases. Furthermore, our search strategy was absent of publication year restrictions and language filters, and it permitted the inclusion of gray/unpublished literature. As a result, our search yielded by far the largest sample of CET trials meta-analyzed to date (k = 76), whereas previous meta-analyses included only 2 to 13 trials (4,8,17,30,44). Also, we performed theoretically driven a priori moderator analyses, which prior meta-analyses did not do. Of importance, our meta-analysis satisfied 94% of items on the AMSTARExBP (23), an augmented version of the Assessment of Multiple Systematic Reviews (37) that focuses on meta-analyses of exercise and BP interventions, indicating its high level of methodological study quality and adherence to contemporary meta-analytic standards.

In summary, our meta-analysis indicates that CET is an effective antihypertensive therapy. The potential BP-lowering effects from CET are equal to or greater than aerobic exercise among adults with hypertension. Our results, albeit positive, warrant additional CET studies that focus on samples with hypertension, investigate BP as a primary study outcome, use appropriate nonexercise control or comparison groups, and manipulate the FITT features of the CET regimen to determine the dose of CET that optimally reduces BP among adults with hypertension.

The results of the study put forth by the authors are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. This was work supported by the Institute for Collaboration on Health, Intervention and Policy (InCHIP) and the Office of the Vice President for Research, Research Excellence Program, University of Connecticut (Storrs, CT, USA). PVT Farinatti was supported by an individual grant from the Brazilian Council for the Scientific and Technological Development (CNPq). The authors have no conflicts of interest to declare. The present study does not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Becker BJ. Synthesizing standardized mean-change measures. Br J Math Stat Psychol. 1988;41(2):257–78.
2. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50(4):1088–101.
3. Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289(19):2560–72.
4. Chudyk A, Petrella RJ. Effects of exercise on cardiovascular risk factors in type 2 diabetes: a meta-analysis. Diabetes Care. 2011;34(5):1228–37.
5. Chudyk AM, Jutai JW, Petrella RJ, Speechley M. Systematic review of hip fracture rehabilitation practices in the elderly. Arch Phys Med Rehabil. 2009;90(2):246–62.
6. Cochran WG. The combination of estimates from different experiments. Biometrics. 1954;10(1):101–29.
7. Cohen J. Statistical Power Analyses for Behavioral Sciences. 2nd ed. Hillsdale: L. Erlbaum Associates; 1988.
8. Cornelissen VA, Smart NA. Exercise training for blood pressure: a systematic review and meta-analysis. J Am Heart Assoc. 2013;2(1):e004473.
9. Deeks JJ, Dinnes J, D’Amico R, et al. Evaluating non-randomised intervention studies. Health Technol Assess. 2003;7(27):iii–x, 1–173.
10. Dolezal BA, Potteiger JA. Concurrent resistance and endurance training influence basal metabolic rate in nondieting individuals. J Appl Physiol (1985). 1998;85(2):695–700.
11. Dos Santos ES, Asano RY, Filho IG, et al. Acute and chronic cardiovascular response to 16 weeks of combined eccentric or traditional resistance and aerobic training in elderly hypertensive women: a randomized controlled trial. J Strength Cond Res. 2014;28(11):3073–84.
12. Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health. 1998;52(6):377–84.
13. Egan BM, Li J, Hutchison FN, Ferdinand KC. Hypertension in the United States, 1999 to 2012: progress toward Healthy People 2020 goals. Circulation. 2014;130(19):1692–9.
14. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–34.
15. Ekkekakis P. Honey, I shrunk the pooled SMD! Guide to critical appraisal of systematic reviews and meta-analyses using the Cochrane review on exercise for depression as example. Ment Health Phys Act. 2015;8:21–36.
16. Ghosh D, Vogt A. Outliers: an evaluation of methodologies. JSM. 2012:3455–60.
17. Hayashino Y, Jackson JL, Fukumori N, Nakamura F, Fukuhara S. Effects of supervised exercise on lipid profiles and blood pressure control in people with type 2 diabetes mellitus: a meta-analysis of randomized controlled trials. Diabetes Res Clin Pract. 2012;98(3):349–60.
18. Hedges L, Olkin I. Statistical Methods for Meta-Analysis. 1st ed. Waltham (MA): Academic Press; 1985. p. 369.
19. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–60.
20. Huedo-Medina TB, Sanchez-Meca J, Marin-Martinez F, Botella J. Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychol Methods. 2006;11(2):193–206.
21. Johnson BT, Huedo-Medina TB. Meta-analytic statistical inferences for continuous measure outcomes as a function of effect size metric and other assumptions. Technical report EHC-2012-01-0011 R2 for the Agency for Healthcare Research and Quality. 2013: [Published online on AHRQ website].
22. Johnson BT, Huedo-Medina TB. Depicting estimates using the intercept in meta-regression models: the moving constant technique. Res Synth Methods. 2011;2(3):204–20.
23. Johnson BT, MacDonald HV, Bruneau ML Jr, et al. Methodological quality of meta-analyses on the blood pressure response to exercise: a review. J Hypertens. 2014;32(4):706–23.
24. Kannel WB. Risk stratification in hypertension: new insights from the Framingham Study. Am J Hypertens. 2000;13(1 Pt 2):3S–10.
25. Leveritt M, Abernethy PJ, Barry B, Logan PA. Concurrent strength and endurance training: the influence of dependent variable selection. J Strength Cond Res. 2003;17(3):503–8.
26. Lipsey MW, Wilson D. Practical Meta-Analysis. 1st ed. Thousand Oaks (CA): SAGE Publications; 2001. pp. 264.
27. MacDonald H, Farinatti PV, Lamberti L, Pescatello LS. Effects of concurrent exercise on hyeprtension: current consensus and emerging research. In: Pescatello LS, Brown M, editors. Effects of Exercise on Hypertension: From Cells to Physiological Systems. 1st ed. Switzerland: Springer International Publishing; 2015.
28. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol. 2009;62(10):1006–12.
29. Mozaffarian D, Benjamin EJ, Go AS, et al. Executive summary: heart disease and stroke statistics-2015 update: a report from the American Heart Association. Circulation. 2015;131(4):434–41.
30. Pattyn N, Cornelissen VA, Eshghi SR, Vanhees L. The effect of exercise on the cardiovascular risk factors constituting the metabolic syndrome: a meta-analysis of controlled trials. Sports Med. 2013;43(2):121–33.
31. Pescatello LS, Franklin BA, Fagard R, et al. American College of Sports Medicine Position Stand: exercise and hypertension. Med Sci Sports Exerc. 2004;36(3):533–53.
32. Pescatello LS, Kulikowich JM. The aftereffects of dynamic exercise on ambulatory blood pressure. Med Sci Sports Exerc. 2001;33(11):1855–61.
33. Pescatello LS, MacDonald HV, Ash GI, et al. Assessing the existing professional exercise recommendations for hypertension: a review and recommendations for future research priorities. Mayo Clin Proc. 2015;90(6):801–12.
34. Pescatello LS, MacDonald HV, Lamberti L, Johnson BT. Exercise for hypertension: a prescription update integrating existing recommendations with emerging research. Curr Hypertens Rep. 2015;17(11):87.
35. Pescatello LS, Arena R, Riebe D, Thompson PD. ACSM’s Guidelines for Exercise Testing and Prescription. 9th ed., 2013, p. 456.
36. Pickering TG. Principles and techniques of blood pressure measurement. Cardiol Clin. 2002;20(2):207–23.
37. Shea BJ, Hamel C, Wells GA, et al. AMSTAR is a reliable and valid measurement tool to assess the methodological quality of systematic reviews. J Clin Epidemiol. 2009;62(10):1013–20.
38. Sillanpää E, Laaksonen DE, Häkkinen A, et al. Body composition, fitness, and metabolic health during strength and endurance training and their combination in middle-aged and older women. Eur J Appl Physiol. 2009;106(2):285–96.
39. Sousa N, Mendes R, Abrantes C, Sampaio J, Oliveira J. A randomized 9-month study of blood pressure and body fat responses to aerobic training versus combined aerobic and resistance training in older men. Exp Gerontol. 2013;48(8):727–33.
40. StataCorp. 2013 Stata Statistical Software. Release 13. College Station (TX): StataCorp LP.
41. Tseng ML, Ho CC, Chen SC, Huang YC, Lai CH, Liaw YP. A simple method for increasing levels of high-density lipoprotein cholesterol: a pilot study of combination aerobic- and resistance-exercise training. Int J Sport Nutr Exerc Metab. 2013;23(3):271–81.
42. Vicente-Campos D, Mora J, Castro-Piñero J, González-Montesinos JL, Conde-Caveda J, Chicharro JL. Impact of a physical activity program on cerebral vasoreactivity in sedentary elderly people. J Sports Med Phys Fitness. 2012;52(5):537–44.
43. Viechtbauer W. Accounting for heterogeneity via random-effects models and moderator analyses in meta-analysis. J Psychol. 2007;215(2):104–21.
44. Zou Z, Cai W, Cai M, Xiao M, Wang Z. Influence of the intervention of exercise on obese type II diabetes mellitus: a meta-analysis. Prim Care Diabetes. 2016;10(3):186–201.


Supplemental Digital Content

Back to Top | Article Outline
© 2016 American College of Sports Medicine