The search strategy yielded 260 citations, of which 18 were assessed in full-text. Of these, 9 randomized controlled trials were included in the systematic review, with a total of 2,220 women [2,8-15] (Figure 1). None of the studies excluded from the review were randomized controlled trials [1,16-22] and one referred only to the second stage of labor . Main characteristics of studies included in the review are presented in Table 1. The earliest studies were published in 1978 [2,8] and only two were performed in the last ten years [14,15]. All studies were published in English language and two were carried out in developing countries (India  and Brazil ). Randomization methods were not fully described in eight studies [2,8-14]. The allocation concealment was considered adequate in four studies [8,10,11,15] and unclear in five others [2,9,12-14]. Only one study described sample size calculation . We did not consider blinding to be feasible and, in fact, blinding was not reported in any study included in this review. After data pooling, the heterogeneity was high in three selected comparisons or outcomes. The results mentioned below and the heterogeneity levels are summarized in Table 2.
Duration of first stage of labor
We pooled the data from seven studies in which the duration of first stage of labor was recorded. A total of 2,166 patients were enrolled in those trials [2,10-15]. Two studies accounted for 78.2% (1,694 patients) of all patients enrolled [10,14]. The reviewers observed performance bias in one of these two studies (reporting on 627 randomized women)  because of delayed amniotomy in the study group. Another study included in this section randomized only patients with protracted labor . Despite these observations, results favored intervention (WMD (random) -0.83 hours; 95%CI -1.60 to -0.06) (Figure 2).
Mode of delivery
Eight studies examined the mode of delivery [2,8-11,13-15]. The overall cesarean section rate was 5.5% for the intervention group and 5.6% for the control group. (OR (fixed) 0.98; 95% CI 0.67 to 1.43) (Figure 3).
Use of analgesia
The use of analgesia was assessed in six studies [2,8,10,11,14,15]. A statistically significant result in favor of the treatment group was found in only one study . The overall use of analgesia was high (69.0%). (Pooled data: OR (random) 0.69; 95%CI 0.37 to 1.30) (Figure 4).
Three studies examined maternal comfort using different methods [11,12,15]. We judged that data pooling would be inappropriate in this case, as the conversion to a common scale was not feasible. In one study , a maternal comfort score developed by the authors themselves was used to evaluate maternal reactions to uterine contractions, including certain behavioral and physiological signs. The overall mean comfort score during the first stage of labor did not differ significantly between the two groups. Another study evaluated women's experiences and the results suggest more women rating their experiences positively in the study group compared to the control group . The third study used a visual-analog scale to evaluate maternal satisfaction during labor , resulting in no statistically significant difference between the two groups.
Four studies examined the need for labor augmentation [2,8,10,14]. No statistically significant results were observed, but all of them reported what could be a protective effect. (Pooled data: OR (fixed) 0.81; 95%CI 0.65 to 1.01) (Figure 5).
Child condition after birth
Six studies reported 5-minutes' Apgar scores [2,8-11,15]. One study reported a statistically significant difference in favor of the intervention group . (Pooled data: WMD (random) 0.11; 95% CI -0.07 to 0.28) (Figure 6).