Pre-pregnancy intrauterine device use is associated with a reduced risk of subsequent preterm birth: a large population-based cohort study.

The data for this study were derived from the NFPHEP in Yunnan province, China. This project was launched by the National Health and Family Planning Commission and the Ministry of Finance of China to provide free health examination and birth planning counseling for couples who planning to have a child within the next six months, with the aim of reducing the potential risk factors for birth defects and preventing adverse pregnancy outcomes [ 22 ]. This project was initiated in Yunnan Province in 2010 and expanded to all 129 counties/districts in the province in 2013. The detailed study design (including questionnaire/interview/survey), management, and implementation have been published elsewhere [ 22 – 24 ]. We extracted Yunnan data from all the 248,830 subjects who participated in the NFPHEP and gave a live birth between January 1, 2013, and December 31, 2019. At the initial enrolment and follow-up visits, participants underwent a preconception interview, physical examination, and laboratory tests. Standard questionnaires were used for data collection by trained local health workers in face-to-face interviews, including couples’ socio-demographic information, lifestyle, disease history, fertility history, and current contraceptive use or the methods used immediately before enrolment for current non-users. Early pregnancy follow-up was conducted within 12 weeks of pregnancy, and pregnancy information of the subjects was collected timely and accurately, including the date of last menstrual period, urine pregnancy test, and ultrasound to determine intrauterine pregnancy. Healthcare personnel provided eugenics-related guidance and advised participants to receive regular pregnancy care. Women were followed up for pregnancy outcomes within 6 weeks postpartum or 2 weeks following the conclusion of other pregnancy events. Since 6,821 (2.7%) women were missing information on contraceptive method before pregnancy and 1572 (0.6%) women had post-term pregnancies, they were excluded from this analysis, resulting in a final cohort of 240,437 women who had a live birth included in this analysis (Fig.  1 ). This study was reviewed and approved by the Ethics Committee of Yunnan Population and Family Planning Research Institute (Approval number: 2017101702). Written informed consent forms were obtained from all participants before data collection. Fig. 1 The flow chart of participants selection The flow chart of participants selection Contraceptive use before pregnancy is the main exposure factor we are interested in. We categorized all participants into three groups according to the methods they used immediately before pregnancy: (1) non-method users, including women who reported not using any contraceptive method before pregnancy; (2) IUDs users, including women who reported that the method used before pregnancy was IUDs. The type of IUD was almost exclusively copper-bearing ones, as levonorgestrel-releasing IUDs were rarely used in China. If used, it was usually not for contraception but mainly to treat conditions such as unexplained excessive menstrual bleeding; (3) other method users, including those who used a method other than an IUD before pregnancy, mostly condoms. If women used more than one contraceptive method before pregnancy, those who used another method (e.g. condom) in addition to an IUD were included in the IUD group. Otherwise, they were included in the other method group. The primary outcome of interest in this study was PTB occurring during the follow-up period after the preconception health examination. We defined PTB as all births delivered at a gestational age of less than 37 weeks or 259 days [ 25 ]. Gestational age was defined as the time between the woman’s last menstrual period and the date of delivery [ 25 ]. A considerable number of potential confounding factors of PTB have been investigated in previous studies [ 26 – 28 ]. We selected them using directed acyclic graphs (see Supplementary Figure S1 ), including age (< 20 years, 20–24 years, 25–29 years, 30–34 years, and ≥ 35 years), occupation (farmer, other), education level (primary school or below, junior high school, senior high school, college or above), ethnicity (Han, Yi, other), household register (urban, town), socioeconomic development level of the county where the women lived (county development status: normal, low, very low), pre-pregnancy body mass index (BMI) (underweight: <18.5, normal: 18.5–23.9, overweight or obese: ≥24), anemia (yes, no), parity (one, two or above), history of PTB (yes, no), chronic disease history (yes, no), adverse pregnancy history (yes, no), complications of pregnancy (yes, no), smoking status (yes, no), alcohol consumption (yes, no), neonate sex (male, female) and singleton (yes, no). County development status was measured according to the national poverty-stricken county standard [ 29 , 30 ]. Anemia was defined as hemoglobin (Hb) less than 110 g/L for pregnant women as recommended by the WHO [ 31 ]. Chronic disease history includes hypertension, heart disease, diabetes, epilepsy, thyroid disease, cancer, tuberculosis, and hepatitis B. Adverse pregnancy history includes spontaneous abortion, induced abortion, birth defect and stillbirth. Complications of pregnancy include gestational diabetes, gestational hypertension, and placenta previa. To deal with missing values, we used a fully conditional multiple imputation procedure to impute 10 replications (greater than the percentage of missing data). Supplementary Appendix 1 describes the covariates and multiple imputation methods in detail. The proportions of women’s baseline characteristics by pre-pregnancy contraceptive method were grouped and calculated. Pearson χ 2 tests were used to compare differences in the distributions of categorical characteristics between groups of pre-pregnancy contraceptive use. We performed univariate and multivariable Poisson regression models to explore the association between contraceptive method use and subsequent PTB, and calculated relative risk (RR) with corresponding 95% confidence intervals (CIs). In model 1, we used non-method users as the control, and in model 2, we used the other method users as the control. In both models, we controlled for women’s age, occupation, education level, ethnicity, household register, county development status, BMI, anemia, parity, history of PTB, chronic disease history, adverse pregnancy history, complications of pregnancy, smoking status, alcohol consumption, neonate sex and singleton to calculate adjusted relative risk (aRRs). We also calculated PTB rates, rate differences, and rate ratios between users of different contraceptive methods. In subgroup analyses, we performed the stratified Poisson regression on selected baseline characteristics and examined the aRRs and 95% CIs. In addition, we examined the interaction between pre-pregnancy IUD use and county development status on both additive and multiplicative scales. The product term was included in the Poisson regression model to assess the multiplicative interaction, which measures the relative change in risk [ 32 ]. Additive interactions were evaluated using three indicators: relative excess risk due to interaction (RERI), attributable proportion due to interaction (AP), and synergy index (S), which measures the absolute change in risk [ 33 ]. Pre-pregnancy IUD use was recorded in the additive interaction calculation, as it acts as a preventive factor [ 34 ]. In sensitivity analyses, we categorized PTB into late preterm (gestational age: 34 to < 37 weeks), moderate preterm (32 to < 34 weeks), very preterm (28 to < 32 weeks), and extremely preterm (< 28 weeks) groups [ 35 ]. We conducted multinomial logistic regression analyses for these categories. Our analysis focused on singleton births and excluded individuals who used multiple contraceptive methods prior to pregnancy. To avoid the influence of chronic disease on the association between pre-pregnancy contraceptive method use and subsequent PTB, further sensitivity analyses were conducted by excluding individuals with a history of chronic diseases. All statistical inferences were 2-sided, and P values < 0.05 were considered statistically significant. All analyses were performed using R, version 4.2.1.

Of the 240,437 women included in the analysis, 45,374 (18.9%) had used an IUD before pregnancy and were classified in the IUD group, 155,649 (64.7%) in the non-use group, and 39,414 (16.4%) in the other method group, with > 95% of the other methods being condoms. The mean age of all participants was 26.3 years (SD 4.8); 115,626 (48.1%) were primiparas and 1,365 (0.6%) were twins or multiple births. Compared to women in the non-use and other method groups, those in the IUD group were older (≥ 35 years: 12.8% vs. 5.1% and 6.1%, respectively), less educated (college or higher: 4.8% vs. 14.2% and 25.0%, respectively), and more likely to be farmers (93.3% vs. 88.4% and 75.4%, respectively) and multiparas (96.6% vs. 38.1% and 55.0%, respectively) (Table  1 ). Table 1 Baseline characteristics of participants, by pre-pregnancy contraceptive methods Characteristics Non-use IUDs Other methods Total P No. subjects 155,649 45,374 39,414 240,437 Age (years) < 0.001  < 20 8628 (5.5) 237 (0.5) 562 (1.4) 9427 (3.9)  20–24 64,954 (41.7) 7061 (15.6) 11,910 (30.2) 83,925 (34.9)  25–29 55,456 (35.6) 20,507 (45.2) 17,664 (44.8) 93,627 (38.9)  30–34 18,641 (12.0) 11,760 (25.9) 6861 (17.4) 37,262 (15.5)  >=35 7874 (5.1) 5804 (12.8) 2414 (6.1) 16,092 (6.7)  Unclear 96 (0.1) 5 (0.0) 3 (0.0) 104 (0.0) Ethnicity < 0.001  Han 95,510 (61.4) 26,857 (59.2) 26,639 (67.6) 149,006 (62.0)  Yi 23,080 (14.8) 7412 (16.3) 5568 (14.1) 36,060 (15.0)  Others 37,059 (23.8) 11,105 (24.5) 7207 (18.3) 55,371 (23.0) Occupation < 0.001  Farmer 137,609 (88.4) 42,318 (93.3) 29,716 (75.4) 209,643 (87.2)  Others 18,040 (11.6) 3056 (6.7) 9698 (24.6) 30,794 (12.8) Household register < 0.001  Urban 150,144 (96.5) 44,167 (97.3) 37,122 (94.2) 231,433 (96.3)  Town 5505 (3.5) 1207 (2.7) 2292 (5.8) 9004 (3.7) Education level < 0.001  Primary School or below 26,428 (17.0) 12,266 (27.0) 3860 (9.8) 42,554 (17.7)  Junior high school 79,800 (51.3) 26,511 (58.4) 17,189 (43.6) 123,500 (51.4)  Senior high school 24,631 (15.8) 3773 (8.3) 7767 (19.7) 36,171 (15.0)  College or above 22,100 (14.2) 2172 (4.8) 9869 (25.0) 34,141 (14.2)  Unclear 2690 (1.7) 652 (1.4) 729 (1.8) 4071 (1.7) County development status < 0.001  Normal 37,445 (24.1) 7266 (16.0) 3571 (9.1) 48,282 (20.1)  Low 58,203 (37.4) 21,425 (47.2) 13,493 (34.2) 93,121 (38.7)  Very low 60,001 (38.5) 16,683 (36.8) 22,350 (56.7) 99,034 (41.2) BMI before pregnancy < 0.001  =24 24,651 (15.8) 9562 (21.1) 6952 (17.6) 41,165 (17.1)  Unclear 148 (0.1) 22 (0.0) 20 (0.1) 190 (0.1) Parity < 0.001  One 96,355 (61.9) 1529 (3.4) 17,742 (45.0) 115,626 (48.1)  Two or above 59,294 (38.1) 43,845 (96.6) 21,672 (55.0) 124,811 (51.9) Anemia < 0.001  No 134,898 (86.7) 39,956 (88.1) 35,741 (90.7) 210,595 (87.6)  Yes 19,773 (12.7) 5105 (11.3) 3393 (8.6) 28,271 (11.8)  Unclear 978 (0.6) 313 (0.7) 280 (0.7) 1571 (0.7) Adverse pregnancy history < 0.001  No 50,019 (32.1) 27,301 (60.2) 15,255 (38.7) 92,575 (38.5)  Yes 32,721 (21.0) 17,137 (37.8) 15,408 (39.1) 65,266 (27.1)  Unclear 72,909 (46.8) 936 (2.1) 8751 (22.2) 82,596 (34.4) Alcohol consumption < 0.001  No 152,144 (97.7) 44,276 (97.6) 37,084 (94.1) 233,504 (97.1)  Yes 2786 (1.8) 871 (1.9) 2182 (5.5) 5839 (2.4)  Unclear 719 (0.5) 227 (0.5) 148 (0.4) 1094 (0.5) Smoking status < 0.001  No 154,571 (99.3) 45,030 (99.2) 39,013 (99.0) 238,614 (99.2)  Yes 539 (0.3) 193 (0.4) 275 (0.7) 1007 (0.4)  Unclear 539 (0.3) 151 (0.3) 126 (0.3) 816 (0.3) Neonate sex 0.214  Male 80,711 (51.9) 23,349 (51.5) 20,205 (51.3) 124,265 (51.7)  Female 74,249 (47.7) 21,818 (48.1) 19,038 (48.3) 115,105 (47.9)  Unclear 689 (0.4) 207 (0.5) 171 (0.4) 1067 (0.4) Singleton 0.621  No 154,780 (99.4) 45,103 (99.4) 39,189 (99.4) 239,072 (99.4)  Yes 869 (0.6) 271 (0.6) 225 (0.6) 1365 (0.6) Abbreviation: IUDs: intrauterine devices; BMI: Body Mass Index Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy Adverse pregnancy history: including preterm birth, overdue birth, spontaneous abortion, induced abortion, and stillbirth Baseline characteristics of participants, by pre-pregnancy contraceptive methods Abbreviation: IUDs: intrauterine devices; BMI: Body Mass Index Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy Adverse pregnancy history: including preterm birth, overdue birth, spontaneous abortion, induced abortion, and stillbirth Overall, 5.30% (12,741/240,437) of study participants had a PTB, of which 65.9% were late preterm, 12.7% were moderate preterm, 12.3% were very preterm, and 9.1% were extremely preterm. Among women who had used an IUD before pregnancy, the PTB rate was 4.86%, with 95% CI ranging from 4.66 to 5.06%. This rate was significantly lower compare to those who did not used any methods (5.42%, 95% CI 5.31–5.53%), with a rate difference of -0.56% (-0.79% to -0.33%) and a rate ratio of 0.90 (0.86 to 0.94). Similarly, the PTB rate for IUD users was significantly lower than for users of other methods (5.33%, 5.11–5.56%), with a rate difference of -0.47% (-0.78% to -0.18%) and a rate ratio of 0.91 (0.86 to 0.97). However, there was no statistically significant difference in PTB rates between the non-method users and those using other method (rate difference of 0.09%, -0.16–0.34%; rate ratio of 1.02, 0.97 to 1.07) as detailed in Table  2 . Table 2 The PTB rates among contraceptive users, by pre-pregnancy contraceptive methods Contraceptives No. of PTB PTB rate, % (95% CI) PTB rate comparisons (95% CI) PTB rate comparisons (95% CI) Rate difference, % Rate ratio Rate difference, % Rate ratio Total 12,741 5.30 (5.20 to 5.38) Non-use 8435 5.42 (5.31 to 5.53) (reference) 1.00 (reference) 0.09 (-0.16 to 0.34) 1.02 (0.97 to 1.07) IUDs 2204 4.86 (4.66 to 5.06) -0.56 (-0.79 to -0.33) 0.90 (0.86 to 0.94) -0.47 (-0.78 to -0.18) 0.91 (0.86 to 0.97) Other methods 2102 5.33 (5.11 to 5.56) -0.09 (-0.34 to 0.16) 0.98 (0.94 to 1.03) (reference) 1.00 (reference) Abbreviation: PTB: preterm birth; IUDs: intrauterine devices Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy The PTB rates among contraceptive users, by pre-pregnancy contraceptive methods Abbreviation: PTB: preterm birth; IUDs: intrauterine devices Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy We performed both the univariate and multivariate Poisson regression models to estimate the association between pre-pregnancy contraceptive use and the risk of subsequent PTB. The results of the analyses showed that the risk of subsequent PTB was significantly lower in women who used an IUD before pregnancy compared with their counterparts who did not use any method (cRR 0.90 [0.86–0.94], aRR 0.84 [0.80–0.89]) and those who used other methods before pregnancy (cRR 0.91 [0.86–0.97], aRR 0.84 [0.79–0.90]) (Table  3 ). We further stratified the data by women’s age, ethnicity, occupation, education level, county development status, pre-pregnancy BMI, parity, anemia, and adverse pregnancy history. In most of these subgroup analyses, pre-pregnancy IUD users were 10–31% less likely than their counterparts to have a subsequent PTB. In counties with a normal development level, the risk of subsequent PTB among IUD users was only 0.7 times that of non-users (aRR 0.70, 0.63–0.78) and other methods users (aRR 0.69, 0.62–0.77) (Fig.  2 ). Table 3 Association between pre-pregnancy contraceptive methods and PTB Contraceptives Model 1 (95%CI) Model 2 (95%CI) Crude RR Adjusted RR Crude RR Adjusted RR Non-use 1.00 (reference) 1.00 (reference) 1.02 (0.97 to 1.07) 1.00 (0.95 to 1.06) IUDs 0.90 (0.86 to 0.94) 0.84 (0.80 to 0.89) 0.91 (0.86 to 0.97) 0.84 (0.79 to 0.90) Other methods 0.98 (0.94 to 1.03) 0.99 (0.94 to 1.05) 1.00 (reference) 1.00 (reference) Abbreviation: PTB: preterm birth; IUDs: intrauterine devices Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy Model 1: Non-use is the control Model 2: Other methods is the control Association between pre-pregnancy contraceptive methods and PTB Abbreviation: PTB: preterm birth; IUDs: intrauterine devices Non-use: including women who reported not using any contraceptive method before pregnancy IUDs: including women who reported that the method used before pregnancy was IUDs Other methods: including women who used a method other than an IUD before pregnancy Model 1: Non-use is the control Model 2: Other methods is the control Fig. 2 Association between pre-pregnancy contraceptive methods and PTB: findings of stratification analyses. Abbreviation: PTB: preterm birth; IUDs: intrauterine devices. Non-use: including women who reported not using any contraceptive method before pregnancy. IUDs: including women who reported that the method used before pregnancy was IUDs. Other methods: including women who used a method other than an IUD before pregnancy. Model 1: Non-use is the control. Model 2: Other methods is the control Association between pre-pregnancy contraceptive methods and PTB: findings of stratification analyses. Abbreviation: PTB: preterm birth; IUDs: intrauterine devices. Non-use: including women who reported not using any contraceptive method before pregnancy. IUDs: including women who reported that the method used before pregnancy was IUDs. Other methods: including women who used a method other than an IUD before pregnancy. Model 1: Non-use is the control. Model 2: Other methods is the control Table  4 presents the results of the interaction between pre-pregnancy IUD use and county development status on both additive and multiplicative scales. The risk of subsequent PTB was significantly lower among IUD users living in counties categorized as normal (aRR 0.71, 0.62 to 0.82) and low (aRR 0.87, 0.76 to 0.99) development levels compared to non-users living in the least developed counties. Notably, a significant negative additive interaction was observed between pre-pregnancy IUD use and low county development status, with RERI, AP, and S being − 0.27 (-0.41 to -0.14), -0.20 (-0.29 to -0.11) and 0.58 (0.49 to 0.68), respectively. Table 4 The interaction of pre-pregnancy IUD use and county development status on both additive and multiplicative scales No. of PTB No. of Total PTB rate, % (95% CI) Estimate 95% confidence interval Lower limit Upper limit Risk ratios for joint effects Main effect  No IUDs and very low development 2203 41,016 5.4 (5.2–5.6) 1.00 (reference)  No IUDs and low development 3989 71,696 5.6 (5.4–5.7) 1.02 0.97 1.08  No IUDs and normal development 4345 82,351 5.3 (5.1–5.4) 1.03 0.97 1.09  IUDs and very low development 401 7266 5.5 (5.0-6.1) 1.01 0.90 1.13 Joint effect  IUDs and low development 1121 21,425 5.2 (4.9–5.5) 0.87 0.76 0.99  IUDs and normal development 682 16,683 4.1 (3.8–4.4) 0.71 0.62 0.82 Measures of additive interaction*  RERI -0.27 -0.41 -0.14  AP -0.20 -0.29 -0.11  S 0.58 0.49 0.68 Abbreviation: PTB: preterm birth; IUDs: intrauterine devices; RERI: relative excess risk due to interaction; AP: attributable proportion due to interaction; S: synergy index *Pre-pregnancy IUD use was recoded in the calculation of additive interaction and the population was categorized into IUD users and non-IUD users Bolding indicates P values < 0.05 The interaction of pre-pregnancy IUD use and county development status on both additive and multiplicative scales Abbreviation: PTB: preterm birth; IUDs: intrauterine devices; RERI: relative excess risk due to interaction; AP: attributable proportion due to interaction; S: synergy index *Pre-pregnancy IUD use was recoded in the calculation of additive interaction and the population was categorized into IUD users and non-IUD users Bolding indicates P values < 0.05 In multinomial logistic regression analysis, there was a negative association between pre-pregnancy use of IUDs and the subsequent late preterm (aRR 0.88, 0.82–0.94), moderate preterm (aRR 0.79, 0.65–0.95), very preterm (aRR 0.60, 0.50–0.71), and extremely preterm (aRR 0.51, 0.40–0.65). We performed a series of sensitive analyses by limiting our study to singleton births, excluding women with chronic diseases, and removing individuals who had used more than one type of contraceptive method before pregnancy. These analyses confirmed that IUD use before pregnancy was associated with a reduced risk of subsequent PTB (see Supplementary Tables 1 and 2 ).

Preterm birth (PTB) is one of the primary causes of death in children under five [ 1 , 2 ]. It was also one of the leading causes of most neonatal deaths in 2018 [ 3 ]. In addition, babies born preterm are at greater risk of developing a range of short- and long-term conditions, such as attention deficit disorder [ 4 ], emotional problems [ 5 ], and respiratory and gastrointestinal complications [ 6 ]. PTB affects nearly 15 million births, approximately 11.1% of all births worldwide, with significant regional variations: 12-13% in the USA, 5-9% in Europe, and 18% in Africa [ 7 ]. In China, the incidence of PTB is estimated to have risen from 5 to 10% in recent years, at around 1.5 million per year [ 8 ]. Copper-bearing intrauterine devices (IUDs) are the most widely used contraceptive method in China and the second most widely used method worldwide [ 9 ]. Approximately 120 million women (accounting for more than 50% of married contraceptive users) in China, and 14.3% of women of reproductive age in the world used this method [ 10 , 11 ]. IUDs are long-acting, easy to use, and highly effective - the Pearl Index is less than 1 in 100 women in the first year of use [ 12 ]. In addition to preventing unwanted pregnancy, copper IUDs have some non-contraceptive health benefits. Results from several case-control studies indicated that copper IUD use was associated with a decreased risk of endometrial cancer [ 13 – 15 ]. Systematic reviews also reported that prior copper IUD use was associated with a reduced risk of endometrial cancer, endometriosis, and cervical cancer [ 16 , 17 ]. The reason for these findings may be that the copper IUD may interfere with local hormone response and/or alter hormone production [ 16 ]. A clinical intervention study found that short-term copper IUD placement improved the embryo transfer placement and pregnancy rates [ 18 ]. However, there are a number of side effects associated with the use of copper IUDs, including prolonged and heavy monthly bleeding, irregular bleeding, more cramps and pain during monthly bleeding, pelvic inflammatory disease, and miscarriage, preterm birth, or infection in the rare case that the woman becomes pregnant with the IUD in place [ 19 ]. China stands out as one of the few countries where IUDs serve as the primary contraceptive method. Concerns surrounding the long-term benefits and potential side effects of IUDs are prevalent among both Chinese users and healthcare providers. A study carried out at a single center in Dongguan, China, revealed that women with a history of IUD use demonstrated a lower likelihood of PTB than the control group, which comprised users of other contraceptive methods and non-users [ 20 ]. Likewise, an ecological study conducted in the USA indicated that women using long-acting reversible contraceptives, including IUDs and implants, exhibited a notably reduced risk of PTB [ 21 ]. Despite these findings, there remains a scarcity of robust evidence establishing the association between IUD use and subsequent PTB. Notably lacking are studies based on large population data. Leveraging information from the National Free Preconception Health Examination Project (NFPHEP) in Yunnan Province, involving over 242,000 participants, we sought to explore the relationship between pre-pregnancy IUD utilization and subsequent PTB. Additionally, we investigated how IUD usage interacts with the level of socioeconomic development in the counties where women reside concerning PTB risk.

Among the 242,009 women who participated in the National Free Preconception Health Examination Project in Yunnan Province in 2013–2019, 5.30% (5.20-5.38%) had a PTB. The rates were 4.86% (4.66-5.06%) for IUD users, 5.33% (5.11-5.56%) for users of other methods, and 5.42% (5.31-5.53%) for non-users. Multivariate Poisson regression analysis, adjusting for potential confounding, revealed a 16% reduction in the risk of PTB in the offspring of women who used IUDs before pregnancy compared with other method users and non-users. Furthermore, this association remained unchanged in both subgroup and sensitivity analyses. Our results are consistent with a cohort study [ 20 ] by Jiang and colleagues in Dongguan, China, which recruited 12,508 multiparas aged 19–48 years. They reported a 26% (aRR 0.74, 95% CI 0.59–0.91) reduction in the risk of PTB in a subsequent pregnancy in women with a history of IUD use compared with their counterparts. However, the control group in the Jiang’s study included women who used other methods and non-users, precluding a direct comparison of the relative risk of IUD users and users of other methods and non-method users. In contrast, our study indicated that the risk of PTB in women using other methods was largely similar to that of non-method users (aOR 0.99, 95% CI 0.94–1.05). Using other contraceptive methods could be considered as a negative control exposure, and this result bolsters the validity of the association of pre-pregnancy IUD use with a reduced risk of PTB [ 36 ]. The underlying mechanisms of the effect of IUDs in reducing the risk of PTB in the offspring are unclear. Early studies have shown that PTB is influenced by a combination of socio-demographic, nutritional, medical, obstetric, and environmental factors, yet the pathogenesis remains incompletely understood [ 26 , 35 ]. In our multifactorial analysis, some covariates, including advanced age, low education attainment, low BMI, a history of PTB, and a history of chronic diseases, were associated with a higher risk of subsequent PTB, aligning with extant literature [ 26 , 35 ]. In this investigation, the predominant type of IUDs used were copper IUDs (> 99%). Mao and colleagues reported that copper IUDs might inflict localized mechanical trauma to the endometrium and provoke an immune-mediated inflammatory response, thereby enhancing endometrial receptivity [ 18 , 37 ]. Enhanced endometrial receptivity is conducive to successful embryo implantation [ 38 ], suggesting that copper IUDs might reduce the risk of PTB by promoting by facilitating more secure implantation. In addition, several studies have indicated that copper IUDs have a protective effect against endometrial hyperplasia [ 16 , 17 , 39 ], which may foster the development of uterine spiral arteries, stimulate the proliferation and differentiation of epithelial and stromal cells, regulate the vasomotor activity of capillaries, increase endometrial blood flow, and catalyze cellular edema and decidualization of stromal cells [ 18 , 40 , 41 ]. These effects collectively facilitate embryo implantation and could exert an influence on the occurrence of PTB. Besides, some studies suggest that IUD placement helps to separate the opposing uterine cavity surfaces, and subsequent IUD removal may ameliorate uterine adhesions [ 42 , 43 ], thereby promoting embryo implantation and potentially reducing the risk of PTB [ 18 ]. Conducting interaction analysis on both additive and multiplicative scales, we assessed the joint effect of IUD use and county-level socioeconomic development. Compared with non-IUD users in the least developed counties, IUD users in counties with a normal and low levels of development had a significantly reduced risk of subsequent PTB (normal level: aRR 0.72, 95% CI 0.62 to 0.83; low level: aRR 0.87, 95% CI 0.76 to 0.98). Additionally, a significant negative additive interaction was observed between pre-pregnancy IUD use and low county development status (RERI − 0.27, 95 CI -0.40 to -0.13; AP -0.20, 95% CI -0.29 to -0.11; S 0.55, 95% CI 0.46 to 0.65). The reasons for these findings are unclear. One plausible explanation could be attributed to the substandard hygienic environment and limited medical resources in the least developed counties [ 44 ], potentially leading to IUD users in these areas facing a higher risk of pelvic inflammatory disease, consequently increasing the risk of subsequent PTB [ 35 , 43 ]. However, further studies are warranted for confirmation. To our knowledge, this is the largest cohort study (> 200,000 women) to investigate the association of pre-pregnancy IUD use with the risk of subsequent PTB. We included all women who participated in the NFPHEP in Yunnan Province and had a live birth between 2013 and 2019 in this study, thereby minimizing the potential for selection bias. The consistency observed across our stratified and sensitivity analyses further bolsters the robustness of the conclusions drawn from our findings. However, this study has several limitations. First, because of constraints in data availability, certain pivotal confounding variables were not accounted for in our analytical model, including nutritional intake, maternal diet, lifestyle during pregnancy, environmental exposures, and partner’s condition. Second, potential misclassification bias may persist, as our estimation of gestational age relied on women’s last menstrual period rather than direct ultrasound-based measurements. Third, the absence of detailed information on specific IUD types, duration of use, and the interval between IUD removal and conception precluded a more nuanced analysis. Consequently, the generalizability of these findings should be approached with caution.

The present study suggests that IUD use may have a protective effect against subsequent PTB. Notably, women who were pre-conceptional IUD users in regions characterized by normal development levels exhibited a reduced risk of subsequent PTB compared to those in the least developed settings. Over the past decade, the utilization of IUDs in China has witnessed a marked decline. In light of these considerations, we advocate for the national family planning initiative and healthcare providers to endorse IUDs as a frontline contraceptive option, emphasizing their proven efficacy and the potential ancillary benefit of mitigating PTB risk.

Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1