Classification of primary caesarean sections in labor and its usefulness for analysis of Slovenian perinatal data

Caesarean section (CS) rates have increased significantly worldwide during the last decades without a concomitant decrease in neonatal morbidity or mortality (1, 2, 3, 4). With growing knowledge of short- and long-term complications associated with CS, many efforts have been made to control the rise in CS rates (2, 3). The safest and most effective approach to achieve this is to avoid the first, i.e. primary CS (3).

Variation in rates of CS in nulliparous women with cephalic foetuses presenting in spontaneous labour at term indicate that differences in clinical practice affect the number of CS significantly (5, 6). A classification that would allow transforming crude numbers of primary CS into useful information on differences in clinical practice patterns could be used to design individualized approaches required to safely reduce the rates of primary CS.

In 2013, Robson et al. proposed a classification of CS in labour shown in Table 1 (7). They classified indications for CS in labour into foetal distress and dystocia. Primary CS for foetal distress was defined as a CS in labour for foetal distress when no oxytocin is used. All other CS were classified as a form of dystocia. Sub-groups of dystocia depend on whether progress of labour (cervical dilatation) is less than 1 cm/h (inefficient uterine action) or more than 1 cm/h (efficient uterine action). Inefficient uterine action was then subdivided into poor response (despite maximum treatment with oxytocin), inability to treat adequately (for foetal reasons), inability to treat adequately (because of the uterus over-contracting), or no treatment (oxytocin not given for other reasons).

The aim of our study was to determine whether this classification of indications for CS in labour can be retrospectively applied using data from a national perinatal database and whether this could help recognizing differences in clinical practice in different maternity units that could explain different CS rates.

Adapted from ref. 7: Robson M, Hartigan L, Murphy M. Methods of achieving and maintaining an appropriate caesarean section rate. Best Pract Res Clin Obstet Gynaecol. 2013;27:297–308; * maximum dose of oxytocin refers to individual unit’s protocol.

We evaluated NPIS data for the period 2013 through 2014. Since 1987, NPIS registers all deliveries in Slovenia at ≥22 weeks of pregnancy or when birth weight is equal to 500 g or above. Registration is mandatory by law in the country’s 14 maternity units and more than 140 variables are entered into a computerized database by the attending midwife and doctor. Patient demographics, family, medical, gynaecologic and obstetric history, data on current pregnancy, labour and delivery, postpartum period, and neonatal data are collected. The complete list of variables with definitions and methodological guidelines has been published online by the Slovenian Institute of Public Health (8). To assure quality of data collected, controls are built in the computerized system, data is audited periodically, and comparisons are made with international databases, such as the Vermont Oxford network in which Slovenia participates.

A 2-year period was chosen for the present analysis to avoid changes in clinical practice that may occur over a longer time and long enough to provide a meaningful analysis due to small number of deliveries in some units.

All nulliparous patients with spontaneous onset of labour at ≥37 weeks with single cephalic foetuses (Group 1 according to the Robson’s Ten Group Classification System) were classified into the seven categories according to classification in Table 1 (9). We also compared rates of several maternal outcomes (3rd or 4th perineal tear, postpartum haemorrhage >500 ml, need for transfusion) and neonatal outcomes corrected for lethal anomalies (neonatal mortality of live born babies who died within 28 completed days from birth), Apgar score <7 at 5 minutes of life, perinatal asphyxia and severe perinatal asphyxia (diagnosed according to criteria of Sarnat and Sarnat (10), and Erb’s palsy).

Chi-square test was used for statistical comparison between units. A P value ≤0.05 was considered statistically significant. The software used for statistical analysis was IBM SPSS Statistics for Windows Version 21.0 (Armonk, NY: IBM Corp.).

There were 41,246 deliveries in Slovenia during the study period. Of these, 8,332 (20.2%) were CS. In Group 1, according to the Robson’s Ten Group Classification System (nulliparous patients with spontaneous onset of labour at ≥37 weeks with single cephalic foetuses) there were 13,572 deliveries and 1,567 (12.0%) CS.

Group 1 contributed 19.1% to the overall CS rate. Table 2 presents numbers and percentages of CS in Group 1 classified according to the proposed classification of CS in labour for every delivery unit in Slovenia.

Rates of CS in Group 1 differed significantly among different maternity units (from 4.1% (unit 3) to 20.9% (unit 14)). The most common indication for CS was cephalopelvic disproportion. Overall, 26.5% of all primary CS in Group 1 were performed for cephalopelvic disproportion (dystocia with efficient uterine action and without foetal malposition). However, incidences of diagnosis of cephalopelvic disproportion differed significantly among units: from 11.2% in unit 7 to 45.9% in unit 4 (odds ratio (OR) 6.72; 95% confidence interval (CI) 3.10–14.71; P<0.001).

Significant differences were seen in other groups of CS in labour as well. Foetal distress without labour augmentation was the indication for CS in 10.7% of primary CS overall, ranging from 1.2% in unit 13 to 30.4% in unit 3 (OR 36.75; 95% CI 4.23–319.44; P<0.001). There were also significant differences in ranges of CS for inability to reach the maximum dose of oxytocin due to foetal intolerance (no such cases recorded in units 1, 3, 4, 9, and 12 and 6.7% of primary CS in unit 8).

Significant differences were also observed in proportions of CS in Group 1 due to an inability to reach maximum oxytocin dose because of the uterus over-contracting or not following unit’s protocols. Proportion of all primary CS in Group 1 in this sub-group ranged from 0% in unit 12 to 45.9% in unit 13.

When comparing unit 3 (the unit with lowest overall CS rate in Group 1, i.e. 4.1%) and unit 14 (the unit with highest overall CS rate in Group 1, i.e. 20.9%), the proportion of primary CS for foetal distress without oxytocin was significantly higher in unit 3 (30.4% vs. 6.8%, OR 6.04; 95% CI 1.70–21.50; P=0.003). Different types of dystocia were, therefore, diagnosed relatively less frequently in unit 3, as was also the case with other units with lower overall primary CS rates.

A significant proportion of cases could not be classified (overall 25.0%, range 12.5% to 55.1% among units). The reasons for inability to classify a case were that either the dose of oxytocin reached, or foetal presentation were not entered in the database.

Table 3 presents maternal and neonatal outcomes in each maternity unit in the Robson’s Ten Group Classification System Group 1. There were statistically significant differences in all outcomes studied except neonatal mortality. Unit 3 (the unit with lowest CS rate) had higher rates of maternal morbidity, however, neonatal mortality and morbidity were higher in some other units. Neonatal or maternal morbidity were not lower in unit 14, the unit with highest CS rate.

The main finding of the present study is that classifying primary CS into foetal distress and several dystocia sub-groups reveals significant differences among maternity units. Such classification could, especially when supported by evaluation of perinatal outcomes, allow a meaningful analysis of CS rates in different hospitals leading to evidence-based initiatives to improve clinical practice.

Various forms of dystocia are diagnosed less frequently in units with lower rates of primary CS, while the proportion of primary CS for foetal distress without labour augmentation is relatively larger in such units. One of the forms of dystocia is the inability to achieve efficient uterine action due to foetal distress with oxytocin augmentation. Such cases were significantly less frequent in units with lower primary CS rates compared to units with higher rates. Since units with lower CS rates did not have higher rates of neonatal mortality and morbidity, it can be argued that primary CS could be safely reduced in certain units with better training in foetal heart rate interpretation (11). However, maternal adverse outcomes, such as incidence of severe perineal trauma and postpartum haemorrhage, were higher in units with lower CS rates and this has to be taken into account when designing strategies for reduction of CS rates. There also seem to be significant differences in diagnosing cephalopelvic disproportion and uterine hyperstimulation. Better training and/or supervision could, again, have an important impact on rates of primary CS in certain units.

A high-dose regimen for labour induction and augmentation with oxytocin has been recommended for all maternity units in Slovenia (12). It includes an initial oxytocin infusion of 2 to 5mU/min with increments every 20–30 min until a maximum dose of 40mU/min is reached (12). Results of the present study indicate that a significant number of maternity units do not follow the proposed regimen (different proportions of CS in which maximum oxytocin dose has not been reached despite inefficient uterine action and no foetal distress). Although conflicting results have been reported from trials comparing low-to high-dose oxytocin regimens, some authors have found that high-dose oxytocin for labour augmentation is associated with a decrease in CS (13, 14, 15). Some units could, therefore, potentially reduce their CS rates with a stricter adherence to the proposed oxytocin regimen.

The study has several limitations. Small numbers in certain sub-groups make comparisons between units difficult. The main limitation is, however, the inability to classify a significant proportion of CS. Nevertheless, we showed that classification of CS in labour using NPIS data can yield important information for future perinatal audit in the country. Moreover, one of the purposes of using classification systems is also to analyse the quality of data collection. Our results indicate that future efforts to improve collection of data on maximum oxytocin dose reached during labour should be undertaken. This is even more important when considering the significant differences in proportions of CS that could not be classified in different units, suggesting different standards of data collection. Differences in sub-groups of primary CS between units found in our study should, therefore, be interpreted with caution. Still, even when only comparing units with similar proportions of non-classified cases, differences are still apparent.

In conclusion, our study suggests that a classification system of primary CS in labour can provide important information on different clinical practices in different units. Knowledge of these differences can lead to the development of effective strategies to safely reduce primary CS rates. In addition, our data also showed that efforts to improve collection of data are needed in order to apply this classification to all maternity units in the country.