114 5 CHAPTER 5 6 | GOODARZI et al. with advancing gestational week. NM contributed more to the increased TM rate compared with FM (Figure S6). In the whole population, non-Dutch women, women with ≥3 previous births, women aged ≥40 years, and low-SES women had the highest rate of PM. Although the TM rate increased with advancing gestational week, the degree of increase differed between subcategories. The trend analyses specified by FM and NM showed a similar trend and are therefore not shown. 3.3.2 | Association between multiple determinants and total mortality Table 2 shows the absolute number and rate of TM for the interaction between two maternal characteristics in the whole study population (model 1). The TM rate associated with single maternal characteristic (Table 1) differed for subcategories when taking a second maternal characteristic into account. The absolute number of TM in the 8- year study period was very low for many subcategories. The TM rate was higher among low-SES women but only for non-Dutch and not for Dutch women. Among Dutch women, the TM rate increased from 40 years and from ≥3 previous births and among non-Dutch women from 35 to 39 years and from ≥2 previous births. Among women aged ≥40 years, primiparous women and high-SES women had the highest rate of PM. 3.4 | Logistic regression analyses 3.4.1 | Association between single determinants and total mortality The results of the univariable and multivariable logistic regression analyses for TM are shown in Figure S7. In the whole population, the OR for TM was the highest for non-Dutch women and women aged ≥40 years compared with reference categories, and remained significant after adjustment for other maternal characteristics. Low-SES women had the highest OR for TM compared with other SES categories, although not significantly different from the reference category after adjustment for other maternal characteristics. Among parity categories, the OR for TM was the lowest for primiparous women compared with nulliparous women, and remained significantly different after adjustment for other maternal characteristics. Women with ≥3 previous births had higher OR than primiparous women compared with nulliparous women, but this difference was not significant after adjustment for other maternal characteristics. After stratification by gestational week, the OR for TM remained significant compared with reference categories after adjustment for other maternal characteristics for non-Dutch women pregnant at 40 + 0 weeks, primiparous women pregnant at 39 + 0 weeks, 40 + 0 weeks and 41 + 0 weeks, women with ≥3 previous birth and women aged <25 years pregnant at 37 + 0 weeks, and women with ≤2 previous births and women aged ≥40 years pregnant at 41 + 0 weeks. The aOR for TM for non-Dutch women pregnant at 37 + 0 weeks and women aged ≥40 years pregnant at ≥39 weeks were not significant compared with the reference categories. The differences in aOR between the parity subcategories were the largest among women pregnant at 37 + 0 weeks and the smallest among women pregnant at 41 + 0 weeks. The difference in aOR between women aged <25 years and 25-29 years was largest for women pregnant at 37 + 0 weeks and the smallest for women pregnant at 41 + 0 weeks. 3.4.2 | Association between multiple determinants and total mortality Interaction between two maternal characteristics in the whole study population was examined by entering a second maternal characteristic as interaction term in the multivariable regression analysis of model 1. When a second maternal characteristic was taken into account, the aOR for TM associated with a single maternal characteristic differed for subcategories. First, we tested differences between subcategories. The interaction between ethnicity and SES and between parity and age showed significant differences in TM between subcategories after adjustment for other maternal characteristics (Figure S7). To examine the specific direction of the interaction effect, different interactions between ethnicity and SES, and between parity and age were tested (Figures S8–S11). Significant aOR are listed in Table 3. Among low-SES women compared with high- SES women, the aOR for TM increased from 1.37 [CI 1.16- 1.62] in the whole study population to 1.76 [1.33-2.33] for non-Dutch women compared with Dutch women. Among women aged ≥35-39 years compared with women aged 25- 29 years, the aOR for TM was increased for nulliparous women, respectively, from 1.15 [0.95-1.51] to 1.47 [1.04- 2.07] for women aged 35-39 years, and from 1.82 [1.23- 2.72] to 3.99 [CI 2.37-6.72] for women aged ≥40 years. Three-way interaction with ethnicity and SES were tested and found not significant for parity and not feasible due to insufficient power for age. Three-way interactions with parity and age were tested and found not feasible due to insufficient power.
RkJQdWJsaXNoZXIy MjY0ODMw