A total, 27 hospitals with outpatient induced abortion services were selected from 16 provinces in the eastern (Fujian, Shandong, Shanghai, Tianjin, Zhejiang), central (Guangxi, Jilin, Jiangxi, Shanxi) and western (Inner Mongolia, Guizhou, Shaanxi, Xinjiang, Yunnan, Sichuan, Chongqing) regions of China. Each institution recruited women who met the inclusion criteria for induced abortion on a continuous basis for two consecutive months in 2019.

We conducted this prospective observational cohort study, which was approved by The Ethics Committee of the Shanghai Institute of Planned Parenthood research (IRB number: PJ2019-22). Women seeking an induced abortion at the study sites were invited to participate in this study. Inclusion criteria for participants were as follows: women aged 18 years or older, intrauterine pregnancy confirmed by ultrasound, voluntary acceptance of induced abortion In early pregnancy (<10 gestational weeks), no contraindications to surgical abortion, being able to understand and answer survey questions, being willing to participate in this study and sign an informed consent form, and being willing to undergo follow-up when returning for a medical examination two weeks after surgery. Patients were excluded if they had undergone an in-patient abortion, medical abortion, spontaneous abortion, or other pathological abortion, if they had a serious chronic disease, or if the physician deemed them inappropriate to participate in this study.

Participants underwent routine pre-and post-operative examinations, and laboratory tests at the hospital. Data on the examinations and tests were extracted from the hospital’s electronic medical record, including temperature, pulse, and relevant gynecological examinations such as the cervix, uterus, and adnexa; detection of vaginal microbiota morphology such as vaginal cleanliness, discharge, and microbial flora; the metabolic products of aerobic bacteria, anaerobic bacteria, fungi, trichomonas, and pH value, etc. Information on antibiotic use was collected at the follow-up interviews two weeks after surgery, including the name of the antibiotics, use dosage, medication regimen, and duration of use. The diagnosis of LGTI was extracted from participants’ medical records, which was diagnosed mainly based on associated symptoms (itching or burning of the vulva), and lab examination of vaginal discharge. We added up the abnormal results of all the laboratory tests and defined a new variable to measure the number of abnormal vaginal discharge tests.

We used EpiData 3.1 for data entry and R 4.0.2 software for data analysis. We used frequencies and percentages to describe the distribution of categorical data, and the chi-squared test to compare the difference in the distribution between categories of baseline characteristics. In addition, a two-level multivariate logistic regression model was used to control for potential confounders and to examine the effectiveness of perioperative antibiotic use on LGTI at two weeks postabortion. A p-value of less than 0.05 was considered statistically significant.

A total of 8,190 women who underwent induced abortion were recruited for this study. Most participants were from tertiary hospitals (84.7%) and maternal and child hospitals (MCHs) (56.4%). The average age of the participants was 29.69 ± 5.98 years, with the majority aged 26–35 years (55.7%). Most participants (72.1%) were married, 40.4% had a bachelor’s degree or higher education level, and 22.5 and 38.4% were employed in manual and intellectual labor, respectively. The majority (73.9%) had their first sexual experience between the ages of 18 and 25 years. The majority (91.9%) reported no sexual intercourse during menstruation, and 59.8% had a history of gynecological and/or obstetric infectious diseases. (Table 1).

As presented in Table 2, the majority of participants had routine gynecological examinations. For example, uterine mobility and tenderness, cervical tenderness, lower abdominal tenderness, pelvic examination, and perianal examination were reported in over 95% of women, and body temperature, cervical smear, and external genitalia were reported more than 80% of participants. Abnormal findings are generally low, most ranging from 0.2 to 0.6%, with a few over 2%, and cervical erosion (21.3%) and cervical smear color (8.4%) being the most common abnormalities.

Testing for vaginal microecological morphology varied between normal flora and potential pathogens. A quarter or less had been tested for gram-positive cocci (25.5%) and rod (17.1%), with positive rates of 17.1 and 38.6%, respectively. A similar proportion of women had been tested for gram-negative cocci (25.0%) and rods (22.2%) and Chlamydia trachomatis (16.3%), with positive rates of 16.3, 6.6 and 18.6%, respectively. More than 90% had fungal spore and hyphae testing, of which 2.4 and 1.8% were positive, respectively.

Trimethylamine, vaginal secretion odor and pH value are indicators of vaginal microecological function. One third of the women had a trimethylamine test, and about 94% had vaginal secretion and PH value tests, with positive rates of 1.2, 2.6 and 22.6%, respectively.

Of all the study participants, 78.5% (6,427) used antibiotics in the perioperative period, and all the antibiotics were included in the Chinese guideline. The use of antibiotics was significantly higher in second-level MCHs (96.0%, 1,899/2,318) than in tertiary general hospitals (81.9%, 1,899/2,318), tertiary MCHs (75.0%, 3,462/4,618), and second-level general hospitals (74.2%, 470/633) (p < 0.001). The most commonly used antibiotics were second-generation cephalosporins (36.8%, 2,367), followed by nitroimidazoles (12.3%, 788), macrolides (7.5%, 484), tetracyclines (4.9%, 312), and lincomycin (3.4%, 220). Only 51 women (0.8%) used quinolones. In addition, 126 women (2.2%) used antifungals. Of all the antibiotic users, 4,297 (66.9%) used only one type of antibiotic, most commonly the second-generation cephalosporins (37.1%, 1,593); 964 (15.0%) used two or three types of antibiotics, with more than half (482) combining nitroimidazoles and second-generation cephalosporins (Table 3).

Table 4 shows the incidence of LGTI by women’s characteristics. As can be seen in the table, the incidence of LGTI in women who used antibiotics and those who did not was 2.7 and 3.1%, respectively. The result of the chi-square test shows no significant difference in the rate between the two groups (p = 0.513). In addition, we used a two-level logistic regression model adjusting for hospital factors, women’s demographics, and perioperative variables, and found no significant difference in the risk of LGTI two weeks post-induced abortion between those using antibiotics and those not using them (OR = 1.01, 95% CI 0.59–1.74). However, the risk increased with the number of abnormal discharge tests (OR = 1.39, 95% CI 1.04–1.86). No significant associations were observed with hospital factors, women’s demographics, and age at first sexual intercourse (Table 4).

The surgical procedure of induced abortion is considered to be a clean-contaminated procedure, as it causes a degree of trauma to the female uterine cavity, which can disrupt the normal barrier function of the cervix (10). For this reason, routine screening for LGTI is usually required before induced abortion. This study found that the majority (>90%) of patients in China underwent routine vaginal secretion tests and gynecological examinations before abortion. Although these tests and examinations were inexpensive and easily accessible for general screening purposes, they may not detect specific pathogenic microorganisms. More than 70% of women undergoing induced abortion in this study were not tested for potentially pathogenic microorganisms in the vagina, such as gram-negative/positive cocci/bacteria, mycoplasma, and chlamydia. This posed a higher risk of missed diagnoses, as pathogenic microorganisms may enter the uterine cavity through the cervix during the invasive procedure, leading to complications such as uterine and pelvic infections, and, in the long term, tubal obstruction (11). We recommend an increase in testing for vaginal pathogenic microorganisms prior to an induced abortion surgery to facilitate timely intervention and to reduce the risk of LGTI following the procedure.

Findings of this study showed that abnormal amount of vaginal discharge and abnormal vaginal cleanliness were most frequent in Chinese women seeking abortion services, accounting for about one-fifth of participants who had a vaginal examination. These results were similar to those reported by Sun and colleagues in China (19.6%) (12). The results of the vaginal microbiota testing showed that the prevalence of gram-positive cocci and bacilli, and gram-negative cocci and bacilli was 38.6, 17.1, 16.3 and 6.6%, respectively, and the prevalence of fungi and trichomonas was 2.4 and 0.3%, respectively. These findings suggest that a significant number of Chinese women undergoing induced abortion were positive for conditional vaginal pathogenic microorganisms. Providers should pay due attention to screening for vaginal pathogens before performing abortion.

Approximately four out of five (78.5%) of women undergoing induced abortion received prophylactic antibiotic treatment in the perioperative period. Either the international guidelines [i.e., the World Health Organization’s “Safe Abortion: Technical and Policy Guidance for Health Systems” (8) and the American College of Obstetricians and Gynecologists’ “SFP guideline 20,102” (6)] or the national expert consensus [i.e., the “Chinese Expert Consensus on the Prophylactic Use of Antibiotics in Induced Abortion (2019)” (10)] recommend the prophylactic use of antibiotics before induced abortion surgery. However, a significant proportion of service provides in China did not follow these recommendations. Unexpectedly, the antibiotic use rate was lower in general hospitals and tertiary hospitals than in MCHs and secondary hospitals. This may be due to the differences in evaluation regulations between hospitals in China regarding the antibiotic use. According to the “Circular on Further Strengthening the Management of Antimicrobial Drugs to Curb Drug Resistance “issued by the Ministry of Health (13), the use of antimicrobial drugs is included as an indicator in the evaluation of hospital level and clinical performance. The top-to-bottom hospital evaluation regulation may need be revised to allow the use of antibiotics in abortion services in all levels and types of hospitals to avoid conflicts between clinical need and hospital evaluation regulations.

Previous studies have shown that the pathogens of LGTI after induced abortion are mainly Escherichia coli, Staphylococcus aureus, and sexually transmitted pathogens such as Neisseria gonorrhoeae and Chlamydia trachomatis (6). The “Chinese Expert Consensus on the Prophylactic Use of Antibiotics in Induced Abortion (2019)” (10) suggests that prophylactic antibiotics should be effective against aerobic bacteria, anaerobic bacteria, and sexually transmitted disease pathogens. Second-generation cephalosporins, metronidazole, doxycycline, and azithromycin are recommended for first-line use. In this study, we found that the most commonly used antibiotics during the perioperative period of induced abortion in China were second-generation cephalosporins (36.8%), which are effective against gram-positive cocci and gram-negative bacilli. Nitroimidazoles was the second most used antibiotics (12.3%). Previous literature reports that anaerobic spore bacteria are common in postoperative infections after gynecological surgery, and nitroimidazoles are the preferred drugs for the treatment and prevention of anaerobic bacterial infections and trichomoniasis (14). Macrolides accounted for 7.5%, tetracyclines 4.8% and quinolones 0.8% of the antibiotics used. Azithromycin, tetracycline, and quinolones (e.g., levofloxacin) are mainly used to treat Chlamydia trachomatis and Mycoplasma dentalium (14). In the group of women who received two or more types of antibiotics, more than half of them used both nitroimidazoles and second-generation cephalosporins. This combination is one of the drug regimens recommended by the Chinese expert consensus (10).

Furthermore, with an increase in the number of induced abortion procedures, the risk of genital tract infection increases correspondingly (11). This study found that the incidence of LGTI was lower in women who used antibiotics in the perioperative period (2.7%) than that in non-users (3.1%), but the difference was not statistically significant. This conclusion remained after controlling for potential confounders using a two-level multivariate logistic regression model. The main reason for this may be that this study was conducted in a real-world setting and did not establish standardized protocols for antibiotic use and diagnosis of LGTI. The lack of standardization may have led to a certain degree of misclassification and under-diagnosis of LGTI, thereby reducing the effectiveness of antibiotic prophylaxis. Nevertheless, we found that the more positive tests in the preoperative vaginal discharge, the higher the risk of postoperative LGTI (OR = 1.39, 95% CI 1.04–1.86). For those with more than one abnormal vaginal discharge test, we recommend testing for vaginal pathogenic microorganisms and administering antibiotics accordingly.

In summary, about four out of five women who had an induced abortion in China took antibiotics in the perioperative period. However, the use of antibiotics did not significantly reduce the risk of LGTI two weeks after abortion. Instead, the risk was significantly associated with the number of abnormal preoperative vaginal discharge tests. We recommend that the use of antibiotics in the perioperative period of induced abortion should be further standardized. In addition, the effectiveness of prophylactic treatment in the abortion service in China needs further study.

This study was conducted in 27 hospitals in 16 out of 31 provinces in mainland China with a sample size of 8,190. To our knowledge, this is the largest study compared with previous similar studies in China. However, it should be noted that this is an observational study conducted in a real-world setting. The diagnostic of LGTI and laboratory testing methods may vary between hospitals, leading to potential information bias in the results.