Pelvic incidence (PI), first defined by Legaye et al. [1], is the angle between the line perpendicular to the sacral endplate at its midpoint and a line connecting this point to the axis of the femoral head. PI, the sum of sacral slope and pelvic tilt [2], is fundamental for the regulation of pelvic spine parameters [3,4]. PI increases gradually during childhood as a result of walking on two legs [5]. After the bone matures, PI remains constant for every individual in adulthood if sacroiliac joint movement does not occur [6]. However, during vaginal delivery, the sacroiliac joint might widen [7]. A well-known effect of estrogen is the proper ligamentous relaxation around the pelvis for delivery [8]. The estrogen concentration increases during pregnancy and peaks in the third trimester [9]. According to a previous study, during pregnancy, the strain force on the lumbar facet joints is more than double the average load [10], and the risk of degenerative spondylolisthesis increases with increasing parity [11]. A recent report from a database in the United Kingdom demonstrated that PI and lumbar lordosis mismatch increases with increasing parity [12].

By contrast, the association between pelvic parameters and vaginal delivery, during which bony changes in the birth canal occur, remains unknown. Thus, the major objective of this study was to use cohort analysis to investigate the association between vaginal delivery and pelvic parameters. This is the first study to elucidate the relationship between vaginal delivery and pelvic parameters.

Of the 320 women who volunteered for the study, 60 were excluded on the basis of the exclusion criteria (Fig. 1). We then excluded 47 patients who were younger than 39 or older than 81 years. The 213 (81.9%) who were included were classified into three groups according to the number of vaginal deliveries (none, one or two, or three or more) (Fig. 2). The ages of the women ranged from 48 to 80 years, and their mean age was 71.1±7.2 years. The mean parity was 2.4 (range, 0–6), and the mean number of vaginal deliveries was 2.2 (range, 0–6). Of the participants, 90 (42.3%) had three or more vaginal deliveries, 105 (49.3%) had one or two vaginal deliveries, and 18 (8.45%) had delivered by cesarean section (CS).

Table 1 shows the demographic characteristics and spinopelvic parameters of the three groups. Age, body mass index, and bone mineral density did not differ significantly among the three groups. Regarding spinopelvic parameters, PI was significantly higher in the women who delivered vaginally three or more times than in those who delivered vaginally once or twice and those who did not deliver vaginally (p<0.001). Sacral slope was significantly higher in the women with three or more vaginal deliveries than in those with none (p<0.001). PI–lumbar lordosis mismatch increased with the number of vaginal deliveries; however, the difference was not significant. Other spinopelvic parameters, including C7 SVA, thoracic kyphosis, and lumbar lordosis, did not differ significantly among the three groups. We measured all parameters with good reproducibility. For C7 SVA, thoracic kyphosis, lumbar lordosis, sacral slope, pelvic tilt, and PI, the intraobserver intraclass coefficients were 0.913, 0.996, 0.926, 0.936, and 0.918, respectively, and the interobserver intraclass coefficients were 0.991, 0.990, 0.964, 0.962, and 0.947, respectively.

Table 2 shows the differences in demographic characteristics and spinopelvic parameters for women who delivered by CS and those who delivered vaginally. PIs were significantly higher in women who delivered vaginally than in those who delivered by CS (p<0.001), as were sacral slopes (p=0.023). PI–lumbar lordosis mismatch was greater in women who delivered vaginally than in those who delivered by CS; nevertheless, the difference was not significant. Other spinopelvic parameters, including C7 SVA, thoracic kyphosis, and lumbar lordosis, did not differ significantly between women who delivered vaginally than in those who delivered by CS.

In this study, we aimed to examine the relationship between the number of vaginal deliveries and spinopelvic parameters. We found that women who delivered vaginally three or more times had the highest PIs among the three groups. Similarly, we demonstrated that women who delivered vaginally had higher PIs than did those who delivered by CS.

Multiparity has been reported as a predisposing factor for increasing PI–lumbar lordosis mismatch. Bailey et al. [12] reported that ligament laxity and biomechanical demands on the sacrum during pregnancy could change the PI. However, they did not find a significant difference in PI among groups classified according to parity. A possible explanation for this is that they could not determine the total number of vaginal and cesarean births. In that respect, our study shows the importance of evaluating the difference in PI according to vaginal delivery.

Vaughan [22] reported that the pelvic joints, including the sacroiliac joints and pubic symphysis, soften during pregnancy, and the true conjugate length increases by 0.5–1 cm during vaginal delivery. Additionally, Borell and Fernstrom [23] reported that forward nutation of the sacrum occurred, causing pelvic expansion and enabling the passage of the fetus into the birth canal during vaginal delivery. The forward nutation of the sacrum results in increased sacral slope; therefore, increased PI could be attributed to the number of vaginal deliveries, during which fetus loading stress affects pelvis alignment, rather than the total parity (Fig. 3). Of interest is that we similarly observed that women who delivered vaginally had higher PIs than did those who delivered by CS (Table 2). Because bony birth canal movements are less likely to occur during CS, the PI differences according to whether subjects had delivered vaginally or not are valid measurements. The proportion of patients who delivered by CS in this study (8.45%) is lower than the global incidence of CS delivery (31.2%) [24]. Furthermore, sacral slope, in this study, was observed to be steeper than that of another cohort in Japan [25]; a possible explanation for this difference might be that a steeper sacral slope is associated with a broader birth canal, which allows women to deliver vaginally more easily.

This study had several limitations. First, we did not collect information about the ages of the women during each pregnancy. The expansion of the true conjugate line is greatest in younger pregnant women because aging increases the surface and number of ridges of the sacroiliac joint, which promotes interlocking and stability [5,26]. Therefore, it is uncertain whether women who give birth at a younger age eventually have a greater PI than those who deliver later in life. Second, we could not evaluate the effect of CS or at least one vaginal delivery on PI and sacral slope. Third, we could not determine whether a greater PI enables more efficient vaginal delivery. Thus, a longitudinal study would be needed to answer these questions adequately. However, our results provide information regarding the association between vaginal delivery and PI, since the participants were age-matched and none had any history of vertebral fracture or abdominal surgery.

Women who deliver vaginally have greater PIs than do women who deliver by CS. This is the first report demonstrating a correlation between vaginal delivery and PI.