Association between first-trimester intrauterine hematoma and twin pregnancy 2 outcomes: A retrospective cohort study

: Background: In recent years, we have found that first-trimester intrauterine hematoma in twin pregnancy has become increasingly common. T he majority of studies on intrauterine hematoma have 20 excluded twin pregnancies, while others did not differentiate between singleton and twin pregnancies. The associations in twin pregnancy are not clear. Therefore, the primary objective of our study was to examine the associations between first-trimester intrauterine hematoma and pregnancy outcomes in twin pregnancy. Material and methods: 1020 twin pregnancies in women who underwent a routine examination from 25 January 2014 to December 2018 were enrolled. According to the presence or absence of intrauterine 26 hematoma, we compared the baseline data and pregnancy outcomes between two groups. Multivariable logistic 3.26, 95% CI 1.11-4.61). However, It did not have increased odds of adverse pregnancy outcomes 32 after 20 Weeks of Gestation .In the final regression model analysis, the associations of hematoma with 33 previous miscarriage history, accepted assisted conception, accompanying vaginal bleeding and 34 miscarriage and vanishing twin syndrome were no longer significant. No association was found 35 between hematoma size or the presence of vaginal bleeding and the risk of pregnancy loss or the 36 vanishing twin syndrome before 20 weeks of gestation (P>0.05). 37 Conclusion: In women with twin pregnancies, the presence of intrauterine hematoma in the first 38 trimester was associated with one or both fetal losses before 20 weeks of gestation. However, 39 chorionicity in twins, the conception method, the intrauterine hematoma size and the presence of 40 vaginal bleeding were not independently associated with pregnancy loss. 41


21
The associations in twin pregnancy are not clear. Therefore, the primary objective of our study was to 22 examine the associations between first-trimester intrauterine hematoma and pregnancy outcomes in 23 twin pregnancy.

49
With the development and application of assisted reproductive techniques, the incidence of twin 50 pregnancy is increasing. It has been reported that assisted reproductive techniques contributed to 16.4% 51 of all multiple-birth infants, and approximately 30.4% of assisted reproductive techniques conceived 52 infants were twins in 2016 [5]. In recent years, we have found that first-trimester intrauterine hematoma 53 in twin pregnancy has become increasingly common. Twin pregnancy is associated with a higher 54 incidence of maternal-fetal complications and more adverse pregnancy outcomes, such as early 55 miscarriage, premature birth, preeclampsia, prenatal bleeding, postpartum bleeding, intrauterine growth 56 restriction and stillbirth, than singleton pregnancy[6]. However, it has not been reported whether the 57 effect of intrauterine hematoma in early pregnancy on the outcome in twin pregnancy is the same as 58 that in singleton pregnancy. At present, the majority of studies on intrauterine hematoma have excluded 59 twin pregnancies, while others did not differentiate between singleton and twin pregnancies. The 60 associations in twin pregnancy are not clear. Therefore, the primary objective of our study was to 61 examine the associations between first-trimester intrauterine hematoma and pregnancy outcomes in 62 twin pregnancy. We also assessed the risk factors in women with twin pregnancy.  Quantitative characteristics are described as the mean ± standard deviation. A t-test was used for 93 comparison between the AP and NP groups. Qualitative characteristics were described by number 94 (percentage), and the chi-square test was used for comparison between the AP and NP groups.We 95 correlated three groups of intrauterine hematoma size using the Kruskal-Wallis test or chi-square 96 test .The associations between intrauterine hematoma and pregnancy loss were estimated using logistic 97 regression analyses.Initially,unadjusted analyses estimated crude odds ratios and95% confidence 98 intervals (CI) (model 1). Multivariable logistic regression analysis was used to adjust for possible 99 confounding. We carried out data analyses using SAS version 9.4 (SAS Institute Inc., Cary, NC); P 100 levels were significant at less than 0.05.

102
Of 1200 consecutive women with twins gestation seen during the study period, 1020 were recruited.  (Table 1). Women with a previous miscarriage history and 111 those who underwent assisted conception were likely to develop intrauterine hematoma. Pregnant 112 women with intrauterine hematoma were more likely to experience vaginal bleeding (p<0.001) (Table   113 1). However, in the final regression model analysis, the associations with previous miscarriage history, 114 assisted conception, accompanying vaginal bleeding and miscarriage and the vanishing twin syndrome 115 were no longer significant ( Table 2, Table 3).

116
We compared the pregnancy outcomes between the two groups. In the AP group, 63 patients (30.1%) 117 had miscarriages, and twenty-six patients (12.4%) had vanishing twin syndrome. In the NP group, there 118 were 50 cases (6.5%) of miscarriage, and 36 cases (4.7%) of vanishing twin syndrome. There were 119 significant differences between the two groups (P<0.001) (Table 4). Our logistic regression analyses 120 showed that first-trimester intrauterine hematoma was significantly associated with increased odds of 121 extreme miscarriage (adjusted odds ratio 14.27, 95% CI 8.25-24.70) and vanishing twin syndrome 122 (adjusted odds ratio 3.26, 95% CI 1.11-4.61) (Tables II and III). However, unlike the NP group, the AP 123 group did not have increased odds of stillbirth, preeclampsia, preterm labor (<34 weeks), low birth 124 weight , postpartum hemorrhage or fetal distress (Table 4).

125
We performed a subanalysis of the 209 women with intrauterine hematoma. We compared the 126 intrauterine hematoma features between women who did and did not ultimately experience pregnancy 127 loss (one or two embryos) at less than 20 weeks of gestation (Table 5). We found no associations 128 between intrauterine hematoma volume, intrauterine hematoma diameter or vaginal bleeding and 129 pregnancy loss (one or two embryos) before 20 weeks of gestation.

131
In this retrospective cohort study, the incidence of hematoma in twin pregnancies reached 21.3%, 132 which was similar to the incidence of hematoma in singleton pregnancies[9-11]. Moreover, we found 133 that the fetal loss rate in pregnant women with early intrauterine hemorrhage (IUH) was significantly 134 higher, the abortion rate was 13 times higher, and the vanishing twin syndrome rate was 2 times higher 135 than those in women without early IUH. However, first-trimester intrauterine hematoma was not 136 significantly associated with an increased risk of stillbirth, preeclampsia, preterm labor, low birth 137 weight , postpartum hemorrhage or fetal distress in twin gestation.

138
The effect of hematoma on first trimester pregnancy outcomes has been debated for many years. Many 139 studies have specifically examined the relationship between first-trimester IUH and pregnancy 140 outcomes in singleton pregnancies. Sandor Nagy reported that women presenting with first-trimester 141 IUH had higher risks of pregnancy-induced hypertension (RR 2.1, 95% CI 1.5-2.9), preeclampsia (RR 142 4.0, 95% CI 2.4-6.7), placental abruption (RR 5.6, 95% CI 2.8-11.1) and small for gestation age . It was not associated with adverse pregnancy outcomes in women at more than 20 weeks 148 of gestation [15]. However, the majority of these studies excluded twin pregnancies. In our study, 149 logistic regression analyses were performed to explore associations between intrauterine hematoma and 150 pregnancy outcomes. To our knowledge, our study is the first to specifically examine the associations 151 between first-trimester intrauterine hematoma and pregnancy outcomes in twin gestations.

152
We did not find any specific characteristics of twin gestation to be predictive of pregnancy loss or the 153 vanishing twin syndrome when intrauterine hematoma was present. Multivariable logistic regression 154 analysis was used to adjust for possible confounding factors. Maternal age and chorionicity in twins 155 were not predictive factors. This finding was similar to that of McLennan's study, which found that 156 maternal age was not the main factor affecting adverse outcomes in twin pregnancies [16]. Although the 157 intrauterine hematoma group had higher rates of maternal miscarriage history, assisted conception, and 158 vaginal bleeding, we found that these factors were not major contributors to fetal loss in early

181
It has been reported that in women with normal twin pregnancies, approximately 30% will become 182 singleton pregnancies, and 10% will result in no fetuses [20][21][22]. The disappearance of gestational sacs 183 or embryos after documented fetal heart activity in multiple pregnancies is known as the vanishing twin 184 phenomenon [23]. And the vanishing twin phenomenon was thought likely to have the association with 185 a chromosomal abnormality [24].We found that the risk of total pregnancy loss was notably higher than 186 the disappearance of one twin if the women presented IUH (30.1% vs 12.4%). The exact reason is not 187 clear. Perhaps the effect of intrauterine hematoma on early gestation is "all or nothing". It was reported 188 that the vanishing twin phenomenon is associated with preterm delivery , very preterm delivery and 189 small for gestation age neonates and low birth weight infants [25]. However, we did not explore the 190 differences in pregnancy outcomes of pregnancies with a vanishing twin caused by first-trimester 191 intrauterine hematoma, which can be further analyzed in the future.

192
Our study has some limitations due to its retrospective design. The population was from a single 193 obstetric practice, so the data may be subject to regional limitations. Another limitation of this study is 194 that the sizes of the hematomas may have changed since the ultrasound examinations. In addition, 195 persistent intrauterine hematoma may have an impact on pregnancy outcomes, but the duration of 196 intrauterine hematoma was not specifically evaluated in this study.

198
In women with twin pregnancy, the presence of intrauterine hematoma in the first trimester is 199 associated with one or both fetal losses before 20 weeks of gestation. In addition, chorionicity in twins,