Observational studies are designed such that the researcher naturally observes subjects and measures variables without assigning treatments to the subjects. Therefore, observational studies have high external validity because they reflect actual clinical labor practices without any intervention (no selection bias, no treatment crossover).

In early observational studies, neuraxial analgesia appeared to be associated with prolonged labor times and increased rates of Cesarean sections and instrumental vaginal deliveries. These observations developed into a longstanding belief among obstetric providers that neuraxial analgesia delays labor and leads to increased rates of Cesarean sections and instrumental vaginal deliveries. However, interpretation of these studies is complex because of the presence of numerous confounding factors that can adversely affect labor outcome and patient selection bias [17]. For example, patients with complicated labor are more likely to experience severe pain, thus requiring more analgesics. These patients may consequently be more likely to choose neuraxial analgesia than patients with uncomplicated labor. Patients with a complicated labor may independently have a higher incidence of Cesarean section and instrumental delivery. In fact, Hess et al. [18] addressed this issue by investigating the association between severe labor pain and rates of Cesarean section. Patients who experienced three or more episodes of breakthrough pain during low-dose epidural bupivacaine/fentanyl labor analgesia had higher rates of Cesarean section in comparison to patients who experienced less pain (odds ratio 2.6, 95% CI 2.0 to 3.4). As such, patients' requests for analgesia might represent a marker of other comorbid risk factors for Cesarean section (e.g., dysfunctional labor, macrosomia, and malpresentation). Observational studies are not capable of reliably indicating whether neuraxial analgesia increases the risk of Cesarean section or instrumental vaginal delivery.

Impact studies are used to determine how a certain treatment affects patient outcomes. Known as sentinel event studies or before-after studies, these studies are designed to compare the incidence of patient outcomes before and after a sentinel event (e.g., introduction of a neuraxial analgesia in labor). This study design excludes cross-over between treatment groups and patients do not choose to participate in the study. Therefore, this study design provides high external validity. However, changes in hospital management and policies, or in patients' characteristics in the before and after treatment time periods, can influence the outcome [19].

In a large study, Yancey et al. [20] investigated the impact of the administration of neuraxial analgesia on Cesarean section at the Tripler United States Army Hospital before and after 1993. As a result of the availability of neuraxial labor analgesia in 1993, the rate of epidural analgesia increased from 1 to 80% over 1 year. Despite this increased rate of neuraxial analgesia, the incidences of Cesarean sections (19.0 vs. 19.4%) and instrumental vaginal deliveries (11.1 vs. 11.9%) remained unchanged. Another observational study investigating this topic was conducted in the National Maternity Hospital in Dublin, Ireland. Here, Impey et al. [21] compared the obstetric consequences, using retrospective analysis, of 1,000 nulliparous females in spontaneous labor at term who delivered in 1987 with comparable groups who delivered in 1992 and 1994 at the same hospital. Cesarean section (4% in 1987, 5% in 1992, and 4% in 1994) and instrumental vaginal delivery rates remained unchanged, although the epidural analgesia rate increased during this time period (10% in 1987, 45% in 1992, and 57% in 1994). A meta-analysis including nine impact studies involving 37,753 women was conducted by Segal et al. [22]. There were no significant increases in the rate of Cesarean section (mean change 0.67%, 95% CI 2.0 to 0.74) or instrumental vaginal delivery (mean change 0.76%, 95% CI 1.2 to 2.8) before and after the implementation of neuraxial analgesia.

Randomized controlled trials are the gold standard design of clinical trials and are used to test the outcome of medical interventions. However, in addition to their limitations, including low external validity and low sample sizes, randomized controlled trials have several weaknesses in the labor analgesia field. For example, neuraxial analgesia has superior analgesic effects in comparison to other forms of analgesia. Also, the use of randomized controlled study designs in the labor analgesia field includes potential limitations such as not including a placebo group or blinding, and higher cross-over rates. Nonetheless, several randomized controlled trials and meta-analyses have compared the effects of neuraxial analgesia to non-neuraxial analgesia on labor outcome. Sharma et al. [23] conducted an individual patient meta-analysis (n = 2,703) comparing Caesarean section rates in female patients randomized to epidural analgesia or systemic opioid analgesia. Their results suggested that the administration of neuraxial analgesia does not increase the risk of Cesarean section (odds ratio 1.04, 95% CI 0.81 to 1.34). A 2011 meta-analysis of 38 randomized trials comparing all modalities of epidural (including combined-spinal-epidural), non-epidural analgesia, and no analgesia during labor concluded that epidural analgesia did not significantly increase the risk of Cesarean section (RR 1.10, 95% CI 0.97 to 1.25). Further, the majority of studies (n = 33) compared epidural analgesia with systemic opioids. Five of the included studies compared epidural analgesia with no analgesia during labor [5]. Thus, results of randomized controlled trials and meta-analyses of these trials strengthen the conclusion that neuraxial analgesia does not increase the rate of Cesarean section deliveries.

In contrast, systematic reviews of randomized controlled trials comparing neuraxial analgesia to systemic opioid analgesia conclude that neuraxial analgesia is associated with an increased risk of instrumental vaginal delivery. In an individual patient meta-analysis (n = 2,703) reported by Sharma et al. [23] and a 2004 meta-analysis (n = 2,962) by Liu and Sia [24], the adjusted odds ratios for instrumental vaginal delivery were 1.86 (95% CI 1.43 to 2.40) and 1.63 (95% CI 1.12 to 2.37), respectively. Similarly, a meta-analysis in 2011 demonstrated that neuraxial labor analgesia was associated with an increased risk of instrumental vaginal birth (RR 1.42, 95% CI 1.28 to 1.57) [5]. However, the cause of the increased rate of instrumental delivery in patients with neuraxial analgesia is uncertain. One possible explanation is that the presence of epidural analgesia engenders a difference in obstetric management [25,26]. For example, the relaxation of the abdominal wall muscles secondary to neuraxial analgesia could encourage the use of instrumental deliveries. Also, there were significantly more instrumental deliveries in teaching hospitals for patients administered epidurals, which could have been a result of instructing residents how to perform instrumental vaginal deliveries.

The effects of neuraxial analgesia on the duration of the first stage of labor were varied, as neuraxial analgesia prolongs the second stage of labor. A 2005 Cochrane review found no difference in the duration of the first stage of labor among patients administered epidural analgesia versus those with systemic opioid or no analgesia [27]. However, the individual meta-analysis of studies by Sharma et al. [23] demonstrated that the first stage of labor is prolonged by approximately 30 min in nulliparous patients with epidural analgesia. Cambic and Wong [19] speculated that these inconsistent results are likely a result of variations in the frequency of cervical examination or the impact of confounding factors such as fluid blouses, which can influence uterine activity. A 2011 Cochrane review concluded there was no difference in the duration of the first stage of labor (mean difference 18.51 min, 95% CI 12.91 to 49.92; 11 trials, 2,981 patients). However, the second stage of labor was longer (mean difference 13.66 min, 95% CI 6.67 to 20.66; 13 trials, 4,233 patients) in those receiving neuraxial analgesia compared to those receiving systemic opioids or no analgesia [5]. It appears that the magnitude of prolonged labor has no clinical significance. In fact, several studies showed that a prolonged duration of the second stage of labor is not associated with maternal or fetal morbidity so long as the fetal status is within normal limits, the mother is well hydrated with adequate analgesia, and there is progress in fetal head descent [28-30].

In summary, much evidence suggests that neuraxial analgesia does not increase the frequency of Cesarean section deliveries in comparison to systemic opioids. There is, however, conflicting evidence with regard to instrumental delivery rates. Whereas randomized controlled trials have concluded that neuraxial analgesia is associated with an increased risk of instrumental vaginal birth, impact studies have identified no such difference. Neuraxial analgesia prolongs the second stage of labor; however, this appears to be of no clinical significance.