Interpretation
Advanced maternal age resulted as an independent RF for caesarean section in Robson groups 1, 2A, 3, and 4A. We have recently shown11 that maternal age ≥40 years represents a RF on CS rate strongest than parity.
The presence of diabetes significantly increased the risk of CS delivery in groups 1 and 5 but its effect was not significant in the other groups. Our data are in contrast with the results of Zeki et al14 that showed that diabetes increases the CS rate in all groups except Group 1. These differences may be due to the different CS rate in the study populations (i.e. lower in our study). Despite the uncertainty of these evidences, diabetes has to be taken into consideration when comparing CS rates, since it is likely to play a role regardless of the Robson group.
Hypertensive disease was identified as a RF in groups 2A (OR 1.32) and 5 (OR 3.49), probably because the decision to perform a CS in labour is taken earlier if a nulliparous undergoes induction of labour for hypertension or she has a previous uterine scar. Also, hypertension represented the strongest RF for CS in group 10 (OR 3.7) probably because in this group the disease is more severe thus to anticipate the time of delivery is a common prudential clinical behaviour. In multiparous without a uterine scar, hypertension did not reach statistical significance. The association between severity of hypertension and CS has been previously shown6. However, in our study, similarly to what Gerli et al.15 showed in Robson’s classes 1 and 3, hypertension represents a RF for caesarean section in nulliparous but not in multiparous. Obesity increased the CS rates in nulliparous women at term (i.e. groups 1 and 2A) but not in multiparous at term (i.e. groups 3 and 4A). Also, in this case our results are similar to those of Gerli et al.15 in a population similar to ours. However, unexpectedly, in our population, obesity decreased the risk of CS in group 5 women attempting TOLAC. It is well known that increasing BMI has an inverse association with the likelihood of achieving vaginal birth16. A possible explanation could be that, given that caesarean sections in labour and in obese women, involves greater difficulties and complications, the decision is postponed, in an attempt to avoid surgery, with a consequent greater probability of obtaining a vaginal birth16; alternatively, it could be that only obese women who are more likely to be successful in vaginal delivery are admitted to TOL. Nulliparous women at term (group 1 and 2A) and multiparous women in spontaneous labour (group 3) with macrosomic fetuses were more likely to have a CS. Of note, macrosomia was one of the two RFs that more than doubled the risk of CS in group 3. The association between macrosomia and CS is well-known9and a recent trial10 proved that, when macrosomia is suspected, labour induction at an earlier gestational age increases the likelihood of vaginal delivery.
Obstetrics analgesia and ethnicity behaved as risk or protective factors depending on the Robson group. According to our data, obstetric analgesia increased the risk of CS in nulliparous women in spontaneous labour, while decreased the risk when labour was induced. In addition, it decreased the risk of CS in groups 5 and 10. A recent meta-analysis13 concluded that epidural analgesia has no impact on the risk of caesarean section. As a matter of fact, in our analysis we included every type of pharmacological pain control (i.e. intravenous drugs, epidural analgesia), so that it is difficult to make any comparison. Probably, more studies are necessary to define the role of obstetric analgesia as a risk or protective factor for CS rates in the different Robson groups.
Being immigrant, in our study, was a protective factor in groups 1 and 10 and a RF in group 4A. A metanalysis by Merry et al.17 that evaluated CS rates between immigrants and non-immigrants women revealed higher CS rate for Sub-Saharan African, Somali and South Asian women; higher emergency rates for North African/West Asian and Latin American women; and lower rates for Eastern European and Vietnamese women. However, they did not evaluated differences between Robson groups. A study by Minsart et al12 demonstrated an increased risk in group 1, 2, 3 and 4 for mothers from Sub-Saharan Africa compared with Belgian natives, while a reduced risk for East European women in group 1. A recent study by Linard et al18 demonstrated a higher CS rate for Sub-Saharan African women in group 5 with only one previous uterine scar, compared to French women. Anyway, it is not possible to compare these studies because the ethnic groups taken into consideration are different and data are heterogeneous.
The statistically significant reduction of CS rate in Robson groups 2A, 5, and 10 reflects the efforts made by our institution to avoid unnecessary CS. It is likely that the weekly audit system adopted in our institution, together with the use of standardized and updated protocols and the publication of annual reports, has played, and can continue to play, a decisive role in reducing CS rates. The implementation of the “trial of labour after caesarean section” helped to reduce the number of CS in Robson group 5, while the reduction in group 2A it is probably due to the fact that our labour induction protocol allows wait a longer period of time to declare the procedure failed19. On the other hand, we think that our CS rate in multiparous women (Robson groups 3 and 4A) and in nulliparous women in spontaneous labour was already adequate, according to international standards2, so that a further reduction could be harmful.