Reply: Response to Letter to the Editor “Labor pain, birth experience and postpartum depression”

To the Editor,

We are grateful for the opportunity to comment on the letter from Blix, Kaasen, and Eri regarding our study [1]. We appreciate their interest and welcome a critical discussion.

With regards to the study population, we understand the difficulties in keeping track of the numbers, as this is a complex longitudinal cohort study with various data sources and response rates. In our approach, we tried to be as descriptive and transparent as possible. Since we used data from different time points and sources, we considered it most appropriate to define the study population women with complete data from all sources. This is reflected in Figure 1, and characteristics of the study population provided in Table 1. Blix et al. ask about characteristics of each sample separately. This would have increased the number of tables considerably and lead to more confusion than enlightenment for the reader, in our opinion. Further, there were virtually no differences in characteristics between the various samples, making the extra reporting redundant. Blix et al. also appear to be confused about the number of participants included in Tables 2–5. As we describe in the table legends, they include data from all sources, but with missing data imputed for the two questions asked shortly after birth. Quite correctly, this involved imputed values in 45% of the sample. We do agree that this merits discussion; however, given the fact that the imputed and the non-imputed analyses showed the same results, we rather consider it strength and a support of our conclusions than a limitation. We decided to present the missing-adjusted models as we believe these results are the least biased, but the crude models with no adjustment for missing are included as supplementary tables.

Figure 1:

Overlap among severe pain during labor (≥8), not very good birth experience (≥3), and post-partum depression (≥10) in women giving birth.^[1]

Next, we agree that the fact that most women experienced very intense labor pain while reporting a good birth experience, is an interesting finding. We would also like to thank Blix et al. for noticing an error in Figure 3. As correctly pointed out, the y-axis (birth experience) was mistakenly reversed. We have now provided a corrigendum where the figure is corrected [2], and we apologize for this error. Indeed, most women reported a good birth experience and high pain intensity. Still, we consider a more thorough discussion of this finding beyond the scope of this paper. Our aim was to study the associations between birth pain/experience and post-partum pain and depression. What Figure 3 illustrates is that more women with a poor birth experience also reported high pain intensity. As we reported, birth experience was associated with post-partum depression, but we also observed that they all overlapped. To illustrate this even better we present a diagram (Figure 1) showing the overlapping distribution between severe pain (NRS ≥8), not a very good birth experience (≥3, median split), and postpartum depression (EPDS ≥10).

Further, Blix et al. miss a discussion of instrumental vaginal deliveries, episiotomies and tears, and the association with epidural. In retrospective analyses, like this, it is necessary to be careful with interpretation of associations. Tears are likely to affect persistent pain, but in our model, this association was not statistically significant. Meta-analyses have concluded that epidural may increase use of vacuum and forceps, but we argue that it is more relevant to consider the evidence showing epidural early in labor to be safe and not increase the risk of instrumental vaginal or cesarean delivery [3]. In our study, most women had epidural late in labor, a practice that has changed during the last decade. Probably, a new data collection today would be different since most birth clinics have acted on the evidence that early epidural is safe. Controlled trials to test if epidural affects the risk of vaginal tears are challenging, as other analgesic treatments are ineffective and the crossover rate to epidural will be substantial. A recent successful trial in this regard, is the RCT by Freeman et al. [4]. They compared epidural with i.e., remifentanil and found no difference in mode of delivery. Unfortunately, tears were not reported specifically. Future studies addressing risk of tears specifically could thus compare epidural with remifentanil.

Finally, Blix et al. claim that epidural analgesia has a negative influence on birth experience, and refer to a study from their own group [5]. They studied low risk pregnancies suitable for vaginal delivery in a midwifery-lead birth unit, but randomized them to deliver in the midwifery unit or in a normal unit or a special unit in the same birth clinic. The midwifery unit did not offer epidural analgesia or augmentation with oxytocin. This information was known to the included women. Among these planned, vaginal, low-risk births, 29% were still transferred from the midwifery unit to one of the obstetric units to have epidural pain relief and/or operative delivery. The analyses revealed that epidural analgesia was associated with a poor birth experience. As no one wishes a complicated birth, and all included women expected to deliver without medical interventions, this is an expected result. Furthermore, there was no information about intra-partum pain. We thus wonder whether the women who were given epidural experienced more severe pain, and whether epidural as such is rather a marker for pain than an independent risk factor for a poor birth experience. A number of studies have focused on birth experience; surprisingly few have studied intra-partum pain as a factor. Nevertheless, a population based study from Iceland that analyzed predictors of childbirth pain experience concluded that several factors were associated with a better experience, among them having effective pain relief with epidural analgesia [6].

The discussion of oxytocin is interesting, and many publications address the effects of oxytocin within the central nervous system. However, lack of clinical evidence limits the significance. Oxytocin receptor agonists may have analgesic effects. Some human studies even suggest that i.e., administration of oxytocin receptor agonists give pain relief, which is unexpected since neither oxytocin nor carbetocin (the two available drugs for human use) are known to pass the blood brain barrier [7], [8]. Of relevance for the relationship between epidural analgesia and administration of oxytocin, the studies comparing early versus late epidural is best suited to assess the impact on oxytocin augmentation. Two studies have relevant information and neither the study by Wang et al. [3] nor Wong [9] found increased use of oxytocin when epidural was initiated early compared to late.

Research, guidelines, and clinical experience should guide midwives advising a laboring woman to make her decisions about treatment options, including method of pain relief. The increasing use of epidural analgesia in Scandinavia indicates that most midwives and pregnant women have acknowledged this as effective and safe.