According to the World Health Organization, approximately 18 million people die of cardiovascular disease (CVD) each year, which accounts for more than 31% of all deaths worldwide, and with a trend toward younger people in recent years [1]. It has been reported that about 30% of the world’s population dies prematurely owing to a lack of physical activity, which accounts for approximately 9% of the total number of people who are physically inactive, thereby making it the largest global public health problem in the 21st century [2]. Physical inactivity is the leading cause of chronic non-communicable diseases and the fourth leading risk factor for increased mortality from non-communicable diseases [3]. Research has found that CVD has become the leading cause of death worldwide as the global population ages [4].

Autonomic function plays an important role in the development of CVD and is a key factor in cardiovascular health and prevention [5]. Among the indicators of autonomic function, heart rate variability (HRV), defined as the difference in adjacent RR intervals between beats in the electrocardiogram over time, responds to sympathetic- and parasympathetic-induced changes in the heart rate. Further, HRV is a non-invasive parameter used to evaluate cardiac autonomic function [6]. Studies have found that the ability of an individual to continuously adapt to changes in the microenvironment and their cardiovascular health is associated with high HRV excitability [7]. Conversely, autonomic dysfunction is associated with higher sympathetic function and lower parasympathetic activity [8]. A relative reduction in HRV is also an independent predictor of CVD risk and all-cause mortality [9, 10]. Studies have confirmed that the important pathophysiological role of HRV abnormalities is reflected in the early stages of essential hypertension, myocardial infarction, and chronic heart failure, leading to an increased risk of death, which is associated with coronary vasoconstriction and increased myocardial oxygen consumption [11, 12, 13, 14]. Thus, the prevention of HRV abnormalities caused by aging or senescence is a challenge that clinicians continue to address.

Regular exercise has been shown to reduce the risk of CVD and premature death and to have a positive impact on cardiovascular health and autonomic function [15, 16, 17, 18]. However, there is some controversy regarding the changes in HRV caused by exercise, which probably relates to the majority of studies illustrating that regular exercise significantly improves autonomic function in people with hypertension, diabetes, and obesity [19, 20, 21], whereas its effect on the general healthy population has remained inconclusive [22, 23]. Recently, several studies have confirmed that aerobic training (AT) improves HRV and reduces cardiovascular risks in healthy adults [24, 25, 26]. Conversely, there is no consensus on the effect of resistance training (RT) on HRV.

A reduced HRV is associated with reduced muscle mass and strength, and resistance exercise is effective in improving abnormal HRV in adults [27, 28, 29]. Some studies have shown that aerobic exercises, when combined with resistance exercises, can improve both endurance and strength at moderate training volumes and durations [30, 31]; however, their effects on autonomic function have not been systematically summarized. In recent years, high-intensity interval training (HIIT) has gained popularity owing to its time-saving and efficient properties and has been shown to improve sympathetic and vagal conditioning [32, 33]. However, it is not yet clear, which of these exercises, is the most appropriate. Therefore, the effects of exercise interventions on HRV in healthy adults must be studied.

Although a large number of randomized controlled trials (RCTs) and systematic reviews have explored the effects of exercise interventions on HRV in adults, indirect comparisons between different exercise interventions have not yet been conducted, and it remains unclear which exercise interventions are optimally effective. A network meta-analysis (NMA) overcomes the limitations of a traditional meta-analysis by allowing optimal ranking of different exercise interventions through performing direct and indirect comparisons. Therefore, the aim of this study was to use an NMA to assess the effects of aerobic, resistance, aerobic combined with resistance, and high-intensity intermittent exercises on HRV in adults, thereby providing an evidence-based basis for exercise to improve cardiac autonomic function and reduce the cardiovascular risks in adults.

In this study, we conducted a comparative analysis of the effects of several exercise interventions on HRV in adults. A total of 29 RCTs were found, which included four different interventions, and involved a total of 1317 participants. It was found that HIIT could increase the SDNN, RMSSD, and LF/HF power ratio in adults, while CBT and RT could increase the LF power and HF power.

Previous studies have shown that reductions in the SDNN and RMSSD are associated with an increased risk of cardiac mortality and morbidity [69, 70]. The RMSSD is an indicator of vagal tone [71]. Previous evidence has confirmed that exercise training may improve HRV in adults by decreasing sympathetic activity and increasing parasympathetic activity [72]. Heydari et al. [44] observed an improvement in the RMSSD in male adults after 12 weeks of HIIT (8 s sprints for 20 min, with recovery three times per week). Piras et al. [33] found that HIIT (20 min training followed by 1 min sprint and 2 min recovery, three times per week) increased the SDNN and RMSSD in adults. This type of training has also been shown to be more suitable for sedentary adults because of the short training time. The present NMA study also showed that HIIT was the best exercise for improving the SDNN and RMSSD. Both the direct and indirect comparisons of HIIT with the other interventions showed that HIIT was the most effective at improving the SDNN and that the difference was significant. Similarly, another meta-analysis showed that HIIT was more effective than other exercise interventions in improving HRV in adults and in increasing the SDNN and RMSSD [73]. A previous RCT also confirmed that HIIT was significantly better at improving the SDNN and RMSSD among sedentary Latin American adults than moderate-intensity continuous training (MICT) [56]. These data confirm the reliability of the results in the present study, suggesting that short periods of high-intensity exercise training may be the best-recommended form of exercise for subsequent national fitness campaigns.

Studies have shown that the LF power is associated with the activity of the stress receptor reflex system, by reflecting the complex regulation of the sympathetic and vagal nerves and, in some cases, the sympathetic nervous system tone [73, 74]. While HRV tends to decline with age, exercise increases sympathetic activity and decreases vagal activity, thereby inducing changes in the autonomic nervous system through complex metabolic and neurohumoral changes, which lead to central adaptation in the body [75]. Exercise interventions in adults can improve the decline in HRV with age and a sedentary lifestyle [76]. The results of this NMA study show that CBT is the most effective form of exercise in improving LF power. Li et al. [77] showed that both AT and CBT can improve the autonomic function of middle-aged and elderly people. Moreover, both of these training methods can increase vagal excitability and decrease sympathetic activity in perimenopausal female patients. However, CBT is significantly better than AT in improving the autonomic function in this population. In contrast, the results of the present NMA showed no significant differences in the direct and indirect comparisons. This finding may be partly related to the type of participants included in this study, the duration of the interventions, and the differences in age and gender.

In this study, RT was the most effective form of exercise for improving HF power, although the results of both the direct and indirect comparisons were not significant. In a study by Liu, the heart rate, HF power, LF power, and HF indices of the training group were not significantly different from those of the control group in all phases of the 12-week strength training cycle, which suggests that 12 weeks of RT does not have a significant effect on the vagal sympathetic tone [41]. In a study by Harris and Holley, in a population of prehypertensive individuals, 9 weeks of cyclic strength training did not significantly alter the resting heart rate, thereby suggesting that the training does not have any effect on parasympathetic dominance [78]. Caruso et al. [79] demonstrated that routine RT exacerbates atherosclerosis and that increased rates of atherosclerosis correlate with cardiac vagal pressure reflex sensitivity. Iellamo et al. [80] found a strong correlation between training intensity and autonomic function after 9 months of training. In their study, 1.5 months of training served as the first phase, 6 months of 75% maximal-intensity training as the second phase, and 3 months of 100% maximal-intensity training as the third phase. HRV and spontaneous stress reflex sensitivity were measured every 3 months in the second and third phases [80]. After 3 months of training at 75% intensity, there were no significant changes in the HF power, LF power, and LF/HF power ratio relative to the basal values. However, after 6 months of training at 75% intensity, there were significant increases in the RR interval, HF power, and high-frequency normalized units relative to the basal values. The increases in these indicators suggest that enhanced vagal tone occurs after 6 months of 75% intensity training. Taken together, the training duration and intensity may significantly influence HRV improvement. The inconsistencies in the intervention time and intensity in the present study resulting from the inclusion of original literature may be an important reason for the lack of any significant differences between the two groups.

The LF/HF power ratio is often used to reflect the relative activity of sympathetic and parasympathetic nerves [81]. For example, a strong sympathetic nervous system may indicate anxiety, fear, irritability, inattention, or hypervigilance. A strong parasympathetic nervous system may indicate muscle weakness, chronic neurasthenia, or depression. Some studies have shown that HIIT is more effective at improving the LF/HF power ratio than MICT [82]. This finding may be explained by the increase in vagal or pressure reflex-mediated sinus node modulation by HIIT. It has also been found that differences in the hemodynamic oscillations experienced during exercise may involve changes in the intrinsic heart rate, S–A node sensitivity [83], and/or myocardial phenotype [84]. HIIT has been shown to be an effective short-term strategy for improving cardiac autonomic function and may have important anti-arrhythmic effects [85]. The results of this NMA study also confirm that HIIT is the most effective treatment modality in improving the LF/HF power ratio and that it yields significant differences. Further, HIIT may be the best exercise modality of choice for improving the autonomic function of adults.

Our NMA and SUCRA ranking results suggest, that in healthy adults, HIIT is the most effective exercise modality for improving the SDNN, RMSSD, and LF/HF power ratio; RT and HF power; CBT and LF power. In subsequent follow-up experiments, several factors must be considered and reported, including age, sex, training intensity, frequency, and duration. In future NMA studies, subgroup analyses can also be conducted according to these factors to comprehensively explore the effects of various exercise patterns on HRV.

AT, aerobic training; RT, resistance training; CBT, combined training; HIIT, high-intensity interval training; HRV, heart rate variability; NMA, network meta-analysis; LF, low-frequency; HF, high-frequency; RCT, randomized controlled trial; MICT, moderate-intensity continuous training; SMD, standardized mean difference; CVD, cardiovascular disease.

FMY, YM, and SYL contributed equally to this study. FMY, YM, SYL, YLS, and CW designed this study. FMY, YM, and SYL conducted the literature search and screening. SYL, YLS, and CW extracted the data. FMY, YM, SYL, and CW analyzed the data and assessed the quality of the included studies. FMY, YM, SYL, YLS, and CW wrote the manuscript. YLS and CW managed the project and acquired the funding. All authors have read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

Not applicable.

This research was supported by a grant from the Fundamental Research Funds for the Central Universities (2020045).

The authors declare no conflict of interest.