We included both randomized and non-randomized controlled trials that included patients with chronic heart failure that made the combined training protocol resistance training plus interval training and compare with interval training alone or a without intervention control group (WI).

Were included human patients of both sexes diagnosed with chronic heart failure, with no other congenital heart disease, atrial fibrillation or heart transplant implants.

IT is characterized when carried out by short intermittent periods of vigorous activity, alternated with recovery intervals or low-intensity exercise. RT is considered as exercise that requires muscle work against a load or resistance – this may be an external load or even body weight. Were considered an intervention of combined training when the RT and the IT were performed at the same session.

We were interested in compare the protocol of combined training (RT plus IT) with two other groups: (a) without intervention group (control group, non-exercise group, wait list control, and placebo); and (b) interval training alone group.

Comparisons were designed as:

• •

  Combined training vs. Without intervention group;

• •

  Combined training vs. Interval training.

The primary outcome of this study was the CRF, measured by change in VO2peak expressed in ml/kg/min at baseline and post intervention.

The searches were conducted in five databases: MEDLINE (Medical Literature Analysis and Retrieval System online) via PubMed, ScienceDirect, Sportdiscus, BIREME (Latin American and Caribbean Center on Health Sciences Information) and Scielo. The databases were selected according to the area of knowledge and scientific relevance worldwide. Furthermore, a search was performed on the references of each study found and selected in the searches.

Thus, the descriptors were selected based on index terms via DECS (BIREME health science subject descriptors) and MESH (Medical Subject Headings – controlled vocabulary used for indexing articles in PubMed). The following descriptors were chosen in English and Portuguese: heart failure; aged; cardiorespiratory fitness; oxygen consumption; resistance training; strength training; high-intensity interval training; and sprint interval training. The descriptors were combined with the Boolean terms “AND” and “OR”. The main search strategy used in all databases was heart failure OR aged AND cardior-respiratory fitness OR oxygen consumption AND resistance training OR strength training AND high-intensity interval training OR sprint interval training, respectively. There was no temporal delimitation of the articles found.

All duplicate articles were removed. The first screening was followed by reading the titles and abstracts of original articles whose summary was available for reading, on human samples that presented chronic heart failure, interval training intervention combined with resistance training and that presented an assessment of VO2peak, before and after intervention.

Subsequently, the studies were assessed for eligibility by reading the articles in full-text. All the stages, such as selection and election of the studies, were evaluated by two researchers (FJMJ) and (ICJ) independently. The reviewers compared the studies and, in case of disagreement, a third reviewer (NL) arbitrated. Any disagreements about a study's eligibility were decided by a consensus meeting.

The following variables were recorded in a database independently by each investigator: author, publication year, average age of the sample, number of individuals in the control group, the number of individuals in the experimental group, the physical training protocols (volume, load, duration, intensity and type of exercise) and maximum oxygen consumption (mean value and standard deviation) before and after intervention.

Once selected, all studies were analyzed for risk of bias, following the risk of bias of the Cochrane Collaboration tool, which assesses the risk of high, low or unclear bias.26 Two reviewers (FJMJ and ICJ) applied to the scale are composed of five criteria that are important for almost experimental and experimental studies, namely: (1) random sequence generation; (2) allocation concealment; (3) blinding of participants and personnel; (4) blinding of outcome assessment; (5) incomplete outcome data; and (6) selective reporting, the evaluation was standardized following the Cochrane Handbook of Systematic Reviews of Interventions.26 In the event of disagreement, a disagreement shall be settled by a third arbitrator (NL).

The quality of the evidence for each outcome was analyzed according to the Grading of Recommendations, Assessment, Development and Evaluation (GRADE),28 which consists of five items: (1) risk of bias; (2) inconsistency; (3) indirectness; (4) imprecision and (5) other considerations. The assessment was made by two reviewers (FJMJ and ICJ) and if any disagreement were found, it was discussed, if there was no agreement between the reviewers, a third reviewer (NL) arbitrated. The quality of the evidence was classified into four categories: high, moderate, low or very low, according to the information provided by the GRADE system and on the recommendations made by Balshem et al.28

Data from the change in VO2peak from post-intervention, characteristics of patients, as well as details of training protocols were analyzed and archived in a database for tabulation. Moreover, risk of bias was also assessed and described, as well as intergroup characteristics in the pre-intervention period, such as: age, gender, preserved or reduced ejection fraction, New York Heart Association (NYHA) classification and left ventricular ejection fraction (%LVEF). In the absence of information relevant to the primary outcome of the study, correspondent authors were contacted in order to request more detail. If there was no answer from the respective authors in relation to the request, the study was excluded from the final analysis.

The meta-analysis was conducted via Review Manager v 5.3 software, using a random effects model. The effect size of treatment was obtained by means of the difference of the variation between the pre- and post-intervention period, described by standardized mean difference (SMD) and the standard error (SE) adjusted for both CT group and the IT control group and the without-intervention group (WI).29 In addition, the analysis of heterogeneity among the studies was obtained through I2 test, in which I2 of <25%, 25–50% and >50% were considered small, medium and large inconsistency, respectively.30

Sensitivity analysis was performed following two procedures: (1) including only reporting of studies with complete’ training protocols; (2) by removing the classified studies with a high risk of bias.

Firstly, 411 studies were found following a combination of descriptors. Sequentially, 23 studies were excluded by duplicate examination. Of the remaining 388 studies 368 did not meet the proposed criteria, in that 144 were not original articles, 19 studies had no abstract available, 21 studies were conducted with animal samples and 184 were without a CT protocol.

Thus, 20 studies were assessed for eligibility and, where necessary, excluded for the following reasons: one study investigated the effect of supplementation, four studies had no evaluation of CRF and five were not composed of patients diagnosed with HF. Therefore, the search was finalized with 10 selected articles. The search steps, selection and eligibility of studies are outlined in Fig. 1.

Figure 1.

PRISMA flow diagram.

(0.28MB).

The selected articles were published between 2002 and 2017. All studies were performed on patients with CHF. A total of 188 patients underwent CT, 81 performed only IT and 132 formed the without intervention group (WI). Only one study was non-randomized controlled trial (RCT).38 In addition, 6 studies compared CT with IT31–36 and 6 studies compared CT with the WI group.35–40

The patients had a mean and standard deviation age of 51±13 years36 and 63±9 years.37 In six studies, patients had NYHA functional class≤III,31–36,38–40 in three NYHA functional class between II and III,33,38,40 and a NYHA functional class between II and IV.37 All studies indicated that patients had reduced ejection fraction.31–40 The main characteristics of the samples are presented in Table 1.

The risk of bias in the included articles was considered high, as shown in Fig. 2A. The main risk of bias was related to the allocation concealment.

Figure 2.

Risk of bias summary: review authors’ judgements about each risk of bias item for each included study (A) and about each risk of bias item presented as percentages across all included studies (B).

(0.56MB).

The risk of bias across studies is shown in Fig. 2B. Random sequence generation and incomplete outcome data were the most prevalent risk of bias across the eligible trials. However, allocation concealment, blinding of participants/personnel and blinding of outcome assessment were unclear in 80% of the studies.

Study quality was assessed using the GRADE system, which indicated moderate quality with serious or not serious methodological problems. Only one non-RCT38 was not included in the analysis of evidence for CT vs WI comparison. Quality assessment is shown in Table 2.

Table 2.

Quality of evidence (GRADE) between CT versus IT (A), and CT versus without intervention (B) for VO2peak.

Certainty assessment							No. of patients		Effect		Certainty	Importance
N°. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	CT	IT	Relative (95% CI)	Absolute (95% CI)
A. VO2peak CT vs IT (follow up: 12 weeks)
6	Randomized trials	Serious a	Not serious	Not serious	Not serious	None	82	88	–	SMD 0.25 higher (0.05 higher to 0.45 higher)	MODERATE	CRITICAL
B. VO2peak CT vs WI (follow up: range 12 weeks to 26 weeks)
5	Randomized trials	Serious a	Not serious	Not serious	Not serious	None	140	137	–	SMD 0.47 higher (0.32 higher to 0.63 higher)	MODERATE	CRITICAL

Note: CI, confidence interval; SMD, standardized mean difference; CT, combined training; IT, interval training; WI, without intervention.

a

Allocation concealment, blinding of participants and/or personnel and blinding of outcome assessment likely.

The CT protocols used in the studies showed similarity, in which 90% performed interventions lasting 12 weeks, except for one study that performed 26 weeks of intervention.40 All IT protocols were performed on a stationary bicycle, and the 30s/60s (effort/rest) ratio was predominant, with total duration of sessions last between 15 and 45min,31,32,34,35,37,38,40 except for two studies that used 120s/120s33,39 and a study that used 240s/180s.36 The intensities of the interval exercise ranged from 50%Wpeak,31,32,34,35,38,40 80–100% Wpeak37 and 40–80%VO2peak.33,36,39 Regarding TR, the majority of the studies used four exercises31,32,34,37,35 with the exception of those who used ten,39 six,33 five,38 three40 or two36 exercises per session. The load intensity ranged from 60 to 80% of one maximum repetition (MR)33,36–40 and 55–65% of two MR.31,32,34,35 Details of training groups and protocols are highlighted in Table 3.

The measurement of CRF was expressed by VO2peak (ml/kg min). All studies assessed the CRF using symptom-limited incremental exercise protocol on an electromagnetically braked cycle ergometer, with direct gas monitoring to assess VO2peak.

The standardized mean difference in the relative VO2peak was 0.47ml/kg/min [CI 95%: (0.32–0.63) p<0.00001; I2=20%] compared to the WI group (Fig. 3A), and 0.25ml/kg/min [CI 95%: (0.05–0.45) p<0.02; I2=7%] in favor of the CT compared with the IT protocol (Fig. 3B).

Figure 3.

VO2peak comparison between CT group (IT+RT) vs WI group (A) and CT group (IT+RT) vs IT (B). Abbreviations: CT, combined or concurrent training; WI, without intervention; IT, interval training; RT, resistance training.

(0.39MB).

However, there was no significant change in the result of the analysis when including only studies with a full description of exercise protocols, or when the studies had a high risk of methodological bias.