• Study design

Every study included in this review was a randomized controlled trial.

• Participants

A total of 195 participants (134 men and 61 women) were included in these studies. Kiviniemi\_2007, Javaloyes\_2019, and Nuuttila\_2017 considered only male samples of 30, 17, and 32 participants, respectively. In the rest of the studies, the samples were composed of men and women: Kiviniemi\_2010 included 24 men and 36 women, Schmitt\_2018 incorporated 19 men and 5 women, and Vesterinen\_2016 assessed 20 men and 20 women. In the studies by Javaloyes\_2019 and Schmitt\_2018, the samples were composed of professional athletes (cyclists and Nordic skiers, respectively) while in the other four studies, the samples were of a nonprofessional level.

• Interventions

According to the types of comparisons contemplated in the present systematic review ((a) endurance training controlled by HRV versus no specific training intervention; (b) endurance training controlled by HRV versus other training intervention; (c) endurance training controlled by HRV (i) versus another training intervention (ii) versus another training intervention; (d) endurance training controlled by HRV (i) versus endurance training controlled by HRV (ii) versus other training intervention versus no specific training intervention. Kiviniemi\_2007, Javaloyes\_2019, Nuuttila\_2017, and Vesterinen\_2016 were classified in Comparison B, Schmitt\_2018 in Comparison C, and Kiviniemi\_2010 in Comparison D.

The interventions in the included studies focused on running (Kiviniemi\_2007, Kiviniemi\_2010, Nuuttila\_2017, and Vesterinen\_2016), skiing (Schmitt\_2018), and cycling (Javaloyes\_2019). They were from 6 to 15 weeks long. In most of the studies, three (Nuuttila\_2017) or four (Javaloyes\_2019, Schmitt\_2018, and Vesterinen\_2016) low-intensity preparation weeks were followed either by the experimental or control groups (standard training) before the intervention. An eight-week intervention was carried out in Javaloyes\_2019, Kiviniemi\_2010, Nuuttila\_2017, and Vesterinen\_2016, whereas Kiviniemi\_2007 considered four weeks of training and Schmitt\_2018 15 days. The assessment weeks were treated separately from the intervention period in Javaloyes\_2019, Kiviniemi\_2007, and Schmitt\_2018.

In every study, the experimental groups trained at moderate or high intensities according to their daily HRV scores. The control groups (standard training) followed a predefined training design at high, moderate, and low intensities (Javaloyes\_2019), high and moderate intensities (Kiviniemi\_2010 and Nuuttila\_2017), high and low intensities (Kiviniemi, 2007) or moderate and low intensities (Vesterinen\_2016). The control group (standard training) design was not explained in Schmitt\_2018.

• Outcomes

The primary outcome analyzed in the included studies was VO2max. The secondary outcomes were: ventilatory thresholds (Javaloyes\_2019, Kiviniemi\_2007) and power in the cycling test (Javaloyes\_2019); rMSSD or RR interval (Javaloyes\_2019, Kiviniemi\_2007, and Schmitt\_2018); basal heart rate (Nuuttila\_2017, Kiviniemi\_2010, and Schmitt\_2018); maximal heart rate in the ergometer test (Nuuttila\_2017); speed in the treadmill test (Kiviniemi\_2007, Nuuttila\_2017, and Vesterinen\_2016); maximal speed in the 10 m test (Nuuttila\_2017); time and lactate in the 3000 m test (Nuuttila\_2017); maximal load in the ergometer test (Kiviniemi\_2007 and Kiviniemi\_2010); and oxygen saturation and VO2 at the second ventilatory threshold (Schmitt\_2018).

Further details about participants, interventions, comparators, and outcomes are provided in Table 1.


### **Table 1.** Overview of the studies included in the review.

#### *IJERPH* **2020**, *17*, 7999



#### 4.1.3. Excluded Studies

As indicated in Figure 1, five studies were excluded from the qualitative analysis. Three studies were excluded because the VO2max was not considered as an outcome [24,25,34], and two studies were excluded because they were not RCTs [35,36].

#### *4.2. Risk of Bias in the Included Studies*

The risk of bias in the included studies is summarized in Table 2. This assessment was made following the Cochrane Collaboration guidelines [30]. In addition, publication bias was assessed using a funnel plot (Figure 2). The Egger test provided statistical evidence of funnel plot symmetry, suggesting the absence of a significant publication bias (*p* = 0.101).


**Table 2.** Risk of bias in the included studies.

**Figure 2.** Funnel plot of standard error by standard differences in means (17 comparison; black circle, HRV-guided training; white circle, traditional training).
