Is there a difference in the risk of serious maternal complications during pregnancy and the postpartum in twin pregnancies according to mode of conception: natural conception, non-IVF fertility treatment, IVF, ICSI or oocyte donation? Women with twin pregnancies after medically assisted reproduction (MAR) had an overall risk of serious maternal complications 30% higher compared with women with natural twin pregnancies, and this association varied according to the MAR procedure; the risk was increased by 50% with IVF using autologous oocytes and by 270% with oocyte donation. IVF has been reported as a risk factor for serious maternal complications in several concordant studies of singleton pregnancies. For twin pregnancies, this association is less well documented with imprecise categorisation of the mode of conception, and results are contradictory. This is a secondary analysis of the national, observational, prospective, population-based cohort study of twin pregnancies (JUmeaux Mode d'Accouchement), which took place in France from 10 February 2014 through 1 March 2015. All French maternity units performing more than 1500 annual deliveries were invited to participate, regardless of their academic, public or private status or level of care. Of the 191 eligible units, 176 (92%) participated. Women with a twin pregnancy who gave birth at or after 22 weeks of gestation were eligible (N = 8823 women included). We excluded women whose mode of conception was unknown (n = 75). Serious maternal complications were regrouped within the recently emerged concept of severe acute maternal morbidity (SAMM), as a binary composite outcome. The exposure of interest was the mode of conception, studied in five classes: natural conception (reference group), non-IVF fertility treatment including insemination and ovarian stimulation, IVF with autologous oocyte, ICSI with autologous oocyte and oocyte donation. To assess the association between the mode of conception and SAMM, we used multivariate logistic regression to adjust for confounders. Structural equation modelling (SEM) was used to explore the contribution to this association of potential intermediate factors, i.e. factors possibly caused by the mode of conception and responsible for SAMM: non-severe pre-eclampsia, placenta praevia and planned mode of delivery. Among the 8748 women of the study population, 5890 (67.3%) conceived naturally, 854 (9.8%) had non-IVF fertility treatment, 1307 (14.9%) had IVF with autologous oocytes, 368 (4.2%) had ICSI with autologous oocytes and 329 (3.8%) used oocyte donation. Overall, 538 (6.1%) developed SAMM. Women with twin pregnancy after any type of MAR had a higher risk of SAMM than those with a natural twin pregnancy, after adjustment for confounders (7.9% (227/2858) compared to 5.3% (311/5890), adjusted odds ratio (aOR) 1.3, 95% CI 1.1-1.6). This association varied according to the MAR procedure. The risk of SAMM was higher among women with IVF using either autologous oocytes (8.3%; 108/1307) or oocyte donation (14.0%; 46/329) compared with the reference group (respectively aOR 1.5, 95% CI 1.1-1.9 and aOR 2.7, 95% CI 1.8-4.1) and higher after oocyte donation compared with autologous oocytes (aOR 1.7, 95% CI 1.1-2.6). Conversely, the risk of SAMM for women with non-IVF fertility treatment (6.2%; 53/854) and with ICSI using autologous oocytes (5.4%; 20/368) did not differ from that of the reference group (5.3%; 311/5890) (respectively aOR 1.1, 95% CI 0.8-1.5 and aOR 0.9, 95% CI 0.6-1.5). The tested intermediate factors poorly explained these increased risks. Beyond the confounders and intermediate factors considered in our analysis, specific causes of infertility and specific aspects of infertility treatments may explain the differences in the risk of SAMM by mode of conception. However, these data were not available. Our study showed an increased risk of SAMM in women with twin pregnancies after MAR, notably after IVF using autologous oocytes and particularly after oocyte donation. To avoid unnecessary exposure to the high-risk combination of MAR and multiple pregnancies, transfer of a single embryo should be encouraged whenever possible. Knowledge of these differential risks may inform discussions between clinicians and women about the mode of conception and help to optimise obstetric care for women in subgroups at higher risk. This work was supported by a grant from the French Ministry of Health (Programme Hospitalier de Recherche Clinique, AOM2012). There are no competing interests. Not applicable.