The current practice of neuraxial labour analgesia in most centres in North America, including our own institution, is evidence based. Most of the evidence that dictates our practice, however, originates in studies produced under very controlled conditions, and it is unknown to what extent the results of these trials are reproducible in the “real world” of clinical practice.1 Furthermore, most of these studies do not evaluate the process of labour and delivery as a whole experience, but rather focus on specific time frames (initiation of analgesia),2 characteristics of analgesia (sensory or motor block),3 mode of delivery,4 or incidence of side-effects and complications.5 In contrast, less evidence is available when it comes to addressing women’s expectations and preferences for analgesia during labour and delivery.6

The available audits in obstetric anesthesia are largely retrospective in nature and are converged on certain aspects of the technique and stages of labour,7 complications,5,8 or mode of delivery.9 In addition to the limitations of being retrospective, most of the audits do not include an objective assessment of patient satisfaction, which is a critical component while assessing the effectiveness of labour analgesia. Prospective observational studies that assess quality and safety of labour analgesia during the entire labour and delivery process and include a special focus on maternal satisfaction are uncommon in the medical literature. Although many studies have used maternal satisfaction as one of their outcomes, they lack in depth and breadth, and the results should be interpreted with caution. A more comprehensive assessment tool that captures the parturients’ point of view is under development and may contribute to a better understanding of women’s needs in the future.5,10

The assessment of the effectiveness of neuraxial analgesia encompassing the entire process of labour and delivery, including maternal satisfaction, may unveil problems, identify opportunities for improvement of clinical practice, and raise new research perspectives. The purpose of this study was to perform a prospective evaluation of the effectiveness of and maternal satisfaction with a patient-controlled epidural analgesia (PCEA) regimen using 0.0625% bupivacaine and fentanyl 2 μg·mL−1, initiated either as a combined spinal-epidural (CSE) or as an epidural anesthesia at a tertiary care teaching centre.

Methods

This was a prospective observational study approved by the Mount Sinai Hospital Research Ethics Board (REB 11-0233-E) and conducted with written informed consent. All women with a viable pregnancy requesting neuraxial analgesia for labour during the month of November 2011 were recruited. When comfortable after the procedure, women were approached for the consent process by the members of the research team not directly involved with their care. Patients presenting for termination of pregnancy or in labour for fetal demise were excluded.

Patients were managed as per departmental routine. A test dose of 2% lidocaine 3 mL followed by a loading dose of 0.125% bupivacaine 10 mL with fentanyl 50 μg was administered. Patient-controlled epidural analgesia was initiated with a maintenance solution of 0.0625% bupivacaine with fentanyl 2 μg·mL−1 with the following settings: bolus 5 mL, continuous infusion 10 mL·hr−1, lockout ten minutes, and a maximum dose of 20 mL every hour. According to patient request, the attending nurse could administer a one-time top-up of 0.125% bupivacaine 10 mL. If ineffective, the anesthesiologist would be called to evaluate the patient, troubleshoot the epidural, and decide whether to give further top-ups, replace the catheter, change the PCEA settings, or increase the concentration of the maintenance solution to 0.125% bupivacaine.

Pain was assessed hourly using a 0-10 verbal numerical pain scale (VNPS), where 0 = no pain and 10 = the most severe pain imaginable. After delivery, before being transferred to the postpartum floor, the patients completed a satisfaction questionnaire composed of ten statements to be classified according to a five-item Likert score (strongly disagree, disagree, neutral, agree, strongly agree). The statements focused on pain experienced during the insertion of epidural and during the first and second stages of labour, side effects, and overall experience with the labour analgesia technique.

Information was collected regarding demographic and obstetric data, characteristics of the neuraxial analgesia, side effects, and complications. All necessary information on the quality of analgesia and complications at all stages of labour and delivery was collected prospectively.

Statistical analysis

No formal sample size calculations were conducted prior to the study, and the study period of one month reflected the available resources.

The study population was summarized using descriptive statistical methods.

We examined the association between failure of analgesia during both first and second stages of labour (dependent variables) and all patient characteristics and variables collected (independent variables). Failure of analgesia or patient dissatisfaction with analgesia was defined as patients having disagreed or strongly disagreed (as per the Likert score) with either of the following two statements: a) I felt NO pain from the time I had my epidural until I started pushing; b) I felt NO pain when I was pushing. For each of the statements above (dependent variables), separately, the group of patients with successful analgesia was compared with the group with failed analgesia regarding all patient characteristics and variables collected using Chi square or Fisher’s exact test, as appropriate, for categorical variables, and Student’s two-sample t test or Wilcoxon Rank Sum test, as appropriate, for continuous variables.

Multivariable logistic regression was employed to determine the association of the independent variables with analgesia failure. Those variables identified as potentially associated with analgesia failure (P < 0.2) in the univariate analysis were included in the models. The adjusted odds ratio (AOR) based on the final models derived by a backward selection procedure with a criterion of P < 0.2 was reported. The data management and statistical analyses were performed using SAS® 9.2 (SAS Institute Inc., Cary, NC, USA).

Results

The recruitment process was very successful and allowed a comprehensive prospective evaluation of our practice. During the study period, 508 deliveries occurred, 332 of which were eligible for the study and were approached for enrolment. Ten women refused to participate, 322 were recruited, 28 were lost to follow-up after recruitment, and 294 were included in the data analysis. Most of the losses happened during weekends and night shifts when the labour floor was very busy and a fast patient turnover prevented us from capturing all the study variables. See flowchart in the Figure.

Figure
figure 1

Study recruitment flowchart

Full size image

Demographic data are presented in Table 1. The incidence of previous failed epidural was 2.1%, and 3.4% of women had back abnormalities. Most women (93.2%) received epidural analgesia, and the remainder (6.8%) received a combined spinal-epidural (CSE) technique. The mean (standard deviation) dilation at epidural insertion was 4.2 (2.0) cm, while the mean time from insertion of epidural to delivery was seven hours (ranging from 23 min to 27 hr). Most procedures were performed by residents. Most anesthesiologists preferred to use a solution of 0.125% bupivacaine as a loading dose. After the epidural placement, about 40% of the women declared having experienced a VNPS higher than 3/10 during the first or second stage of labour. Nearly 50% of the women required epidural top-ups during labour; 25% of them received only one top-up given by the nurse, and the remainder received two or more top-ups given by both nurses and physicians. The nurses’ top-ups were 0.125% bupivacaine 10 mL. The physicians used bupivacaine (66%), lidocaine (21%), and/or fentanyl (13%) in different doses and concentrations in their top-ups. Further details on the anesthetic technique, side effects, and complications are shown in Tables 2 and 3. The incidence of hypotension was 9.1%, and the incidence of pruritus, fetal bradycardia, and nausea or vomiting was 7.1% each. As for the technical complications, three (1%) dural punctures occurred, and the epidural catheter had to be replaced in 13 (4.4%) women. There was no detected case of intravascular injection of local anesthetic or seizure.

Table 1 Demographic data (n = 294)
Full size table
Table 2 Characteristics of neuraxial labour analgesia (n = 294)
Full size table
Table 3 Complications and side effects requiring treatment (n = 294)
Full size table

Sixty-three percent of women presented in spontaneous labour, and 52% of those required oxytocin augmentation. Thirty-seven percent were scheduled for induced labour. Malpresentation, defined as any presentation other than the occipitoanterior position, was detected in 6.5% of the parturients. The delivery mode was spontaneous vaginal in 72% of patients and assisted vaginal in 13%, either by vacuum extraction (10.5%) or forceps (2.5%). Forty-six (16%) women underwent Cesarean delivery, and 95.6% of the epidurals in situ worked well for the procedure; two (0.6%) women had their epidural converted to general anesthesia. The main indications for Cesarean delivery were atypical/abnormal fetal heart rate in 37%, arrest of dilation or descent in 58%, maternal fever in one woman, and severe vaginismus in another. The type of labour, mode of delivery, and the incidence of malpresentation are presented in Table 4.

Table 4 Obstetric data (n = 294)
Full size table

According to the responses to the satisfaction questionnaire, most women (90%) considered the placement of the epidural comfortable. One hundred and thirty-four (45%) women reported experiencing pain during the first or second stages, although most of them considered that the epidural worked well (85%). The detailed responses to the satisfaction questionnaire are shown in Table 5.

Table 5 Maternal responses to the satisfaction questionnaire (n = 294)
Full size table

The results of the univariable and multivariable logistic regression analyses related to analgesia failure during the first stage of labour are reported in Tables 6 and 7. The following variables were associated with anesthesia failure: patients’ overweight status (body mass index 25-30) (AOR = 2.56; 95% confidence interval [CI]: 1.1 to 5.97), need for top-ups given by the anesthesiologist (AOR = 5.08; 95% CI: 2.31 to 11.11), and induced labour (AOR = 2.4; 95% CI: 1.2 to 5.2). Nausea or vomiting was associated with lower odds of analgesia failure (AOR = 0.25; 95% CI: 0.07 to 0.94).

Table 6 Univariate analysis for the statement: I had NO pain from the time I had my epidural until I started pushing. Women who disagreed/strongly disagreed with the statement (Group II) were compared with the remaining women (Group I)
Full size table
Table 7 Multivariate analysis for the statement: I had NO pain from the time I had my epidural until I started pushing
Full size table

The results of the univariable and multivariable logistic regression analyses related to analgesia failure during the second stage of labour did not show any statistically significant associations.

Discussion

Overall, the epidural or CSE analgesia regimen offered at our institution led to a very low number of complications; however, we encountered a high incidence of women reporting pain (VNPS > 3) during the first or second stages of labour. Overweight women and those undergoing induction of labour were most likely to disagree with the statement that they felt no pain during the first stage of labour. Although a large proportion of patients requested multiple top-ups by nurses and physicians, the change to a stronger local anesthetic maintenance solution, readily available on the floor, occurred in just a few cases. Interestingly, however, most patients were satisfied with their pain management.

In order to capture the parturients’ point of view on the whole labour experience, our questionnaire focused on pain during the first and second stages separately, side effects, and overall experience. Patients completed the questionnaire within two hours of delivery before being transferred to the postpartum unit. Most patients reported being comfortable during placement of the epidural, but almost 40% of them disagreed with the statement “I felt NO pain from the time I had my epidural until I started pushing”. Similarly, 25% disagreed with the statement referring to lack of pain during the second stage. This information obtained from the questionnaire is in keeping with the pain scores collected during labour. Our findings are not consistent with those of other studies using dilute solutions of bupivacaine11-13 and probably imply that the environment of randomized controlled trials (RCT’s) may not completely replicate “real life” practice.1 When drugs or techniques proved useful in RCTs are incorporated into clinical practice, the results may be different.14 Carvalho et al. 11 compared different patient-controlled epidural analgesia (PCEA) regimens of bupivacaine for labour analgesia. In one of the study groups using nearly the same management protocol as ours (i.e., a continuous infusion of 0.0625% bupivacaine 10 mL·hr−1 with sufentanil 0.35 μg·mL−1, a bolus of 6 mL and a lock-out interval of eight minutes), 24% of women required rescue boluses, which is nearly half of that found in our study. Different explanations may be offered to justify the higher incidence of insufficient analgesia in our study: the smaller volume used in the loading dose (10 mL vs 15-20 mL), the different expectations regarding analgesia in the studied samples, or simply the inability to reproduce RCT findings in daily clinical practice. Two similar studies15,16 comparing continuous infusions of 0.125% and 0.0625% bupivacaine without a PCEA bolus also found that 40-60% of patients required supplemental doses when receiving the low concentration solution. In those cases, the lack of PCEA boluses could justify the findings.

Results from dose-finding studies have found that 0.048% is the minimum local anesthetic concentration (MLAC-EC50 [effective concentration in 50% of women]) of bupivacaine with fentanyl 2 μg·mL−1 in labour.17 Using the table of normal deviates, the effective concentration in 95% of patients (EC95) could be estimated to be 0.091%.18,19 Therefore, 0.0625% bupivacaine may not be sufficient to provide consistent analgesia to every woman, but rather to approximately 60% of them. This is important information, and perhaps our clinical practice, as reflected by the results of this study, proves this point. It should be highlighted, however, that MLAC studies are designed to obtain EC50, and EC95 is assumed from mathematical treatment; therefore, caution should be exercised when interpreting EC95 values. Furthermore, most MLAC studies express the concentration of local anesthetic required to initiate an epidural labour analgesia regimen. To date, one study by Ginosar et al.20 would be more applicable to the discussion of our results. In their study, the authors looked at the MLAC of bupivacaine required during maintenance of epidural analgesia, and in the context of the epidural infusion of fentanyl at 30 μg·hr−1, they determined the MLAC to be only 0.019% (95% CI 0.000 to 0.038). Nevertheless, these results are of limited value to suggest a useful concentration for clinical use, as they express EC50, not EC95.

A large retrospective study12 evaluating the need for anesthetic interventions in women who received bupivacaine in three different concentrations (0.04%, 0.0625%, and 0.125%) reported that the number of interventions was similar regardless of the concentration of local anesthetic used. While patients receiving lower concentrations required more supplemental analgesic doses, those receiving higher concentrations required more treatment for hypotension. Therefore, simply increasing the concentration of the maintenance solution for all women in our practice may not be the best solution, especially given that our PCEA regimen was satisfactory to most patients. Our results may suggest, however, that more education aimed at the individualization of PCEA regimens is needed for patients, nurses, and physicians. Despite the fact that almost 40% of women experienced pain during the first stage of labour and 24% required multiple supplemental top-ups, our study shows that the change to a maintenance solution of higher concentration occurred in only 7.8% of the cases.

Our multivariable regression model revealed that women who were most likely to disagree with the statement that they had no pain during the first stage of labour were those who were overweight and those undergoing induced labour. These results are in keeping with a previous survey reporting more pain in primiparous women and in those having induced or augmented labour.21 These results are also in keeping with those of Capogna et al.22 who showed greater analgesic requirement in prostaglandin-induced labour when compared with spontaneous labour. Furthermore, women who requested a top-up by the anesthesiologist in addition to that given by the nurse were more likely to disagree that they had no pain. Some of these women went on to request multiple top-ups, indicating insufficient local anesthetic concentration in the maintenance solution, but in the majority of cases, that did not prompt the change to a higher concentration. Our results suggest that a change in the maintenance solution should be contemplated once a top-up by the anesthesiologist is required.

Interestingly, despite the high incidence of insufficient analgesia, there was high overall satisfaction with the pain management. Satisfaction is a very complex concept and knowingly very difficult to assess. The first problem in measuring satisfaction is to understand how it is defined. So far, there is a lack of validated satisfaction questionnaires published for use in women during labour. In 1987, Lomas et al. made one of the first attempts to create a labour and delivery satisfaction index.23 Unfortunately, their instrument cannot be used to assess women’s views on specific aspects of intrapartum care. Further attempts to assess satisfaction were comprehensively reviewed by Hodnett.24 She identified four factors that appear to be paramount to maternal satisfaction: personal expectations, the amount of support from caregivers, the quality of the caregiver-patient relationship, and the involvement of the women in the decision-making. She further suggests that pain relief and medical interventions do not have a direct influence on satisfaction. This is in accordance with our results. Despite the incidence of 40% of pain during first stage, more than 90% of women reported overall satisfaction. Efforts continue to build a solid questionnaire to assess satisfaction during labour, and a quality index is being developed that would reflect the parturient’s point of view.10

The main limitation of our study is the short period of data collection; however, we had only the resources required to run this prospective project 24 hr per day for one month. Furthermore, as a single centre study recruiting patients over one month, the external validity of our findings may be limited. Women’s expectations toward labour analgesia may be influenced by many factors, including the information that is given to them during prenatal care and the interaction with health care professionals other than the anesthesiologist during labour and delivery. At our institution, the Department of Anesthesia provides bimonthly complimentary educational sessions on labour analgesia to all women; however, only a fraction of them will attend these sessions, and many will have obtained information from different sources.

We conclude that our neuraxial labour analgesia practice of PCEA with a dilute solution of 0.0625% bupivacaine and fentanyl 2 μg·mL−1 is effective for most women delivering at our institution and is associated with high patient satisfaction. Notwithstanding, some women will require multiple top-ups and a higher concentration of local anesthetic, and strategies to respond quickly to individualized needs should be readily available. Most importantly, anesthesiologists and nurses/midwives involved in their care should be educated on how to respond to such needs in the most appropriate way. Prospectively designed studies of labour analgesia such as this may offer a unique opportunity for testing the effectiveness and safety of evidence-based practice in routine clinical care. Prospective evaluation of our daily practice is challenging and requires considerable resources; however, these studies allow an objective documentation of women’s experience of pain during labour epidural analgesia, highlighting areas of deficiencies in clinical care and offering a unique opportunity to improve practice.