Cost-effectiveness of levonorgestrel intrauterine system versus hysteroscopic niche resection for caesarean scar-related spotting in China: an economic evaluation alongside a randomised controlled trial

A niche is a myometrial defect at the site of a previous caesarean section (CS) uterine scar and represents the most common long-term complication after CS. 1 For example, transvaginal ultrasound detects niches in approximately 24% of cases, whereas gel instillation sonohysterography identifies them in up to 56%. 2 8 Niches are associated with a range of symptoms, including abnormal uterine bleeding—particularly postmenstrual spotting, dysmenorrhoea, chronic pelvic pain and infertility. 4 9 These symptoms can significantly impair women’s physical and mental health. 10 To address these problems, several surgical interventions have been developed to reduce prolonged bleeding, postmenstrual spotting and dysmenorrhoea, thereby improving quality of life. 11 14 Hysteroscopic niche resection (HNR) is a minimally invasive procedure suitable for women with a residual myometrium of at least 2–3 mm. In cases of thinner residual myometrium (<2.5 mm), HNR carries a higher risk of uterine perforation. 15 Although it can be performed in daycare settings, the procedure usually requires cervical dilation to 8–9 mm, and in most centres it is carried out under general or spinal anaesthesia. HNR improves menstrual outflow by resecting the distal part of the niche and reduces bleeding from small vessels through coagulation of the niche surface. The levonorgestrel intrauterine system (LNG-IUS) has also been investigated as a treatment for niche-related postmenstrual spotting. In a randomised controlled trial (RCT) of 103 women, LNG-IUS was associated with a 4 day greater reduction in spotting compared with expectant management at 6 months. 15 Similarly, a prospective cohort study of 78 patients reported that LNG-IUS was more effective than HNR, with median reductions in postmenstrual spotting of 7 days versus 5 days, respectively. In this study, 83.3% of LNG-IUS users achieved a reduction of more than 50% in spotting, with low complication rates. 16 To directly compare the effectiveness of LNG-IUS and HNR, we conducted an RCT involving 208 patients. Clinical outcomes, including postmenstrual spotting, dysmenorrhoea and patient satisfaction, were reported previously in the MIHYS NICHE trial (ChiCTR1900025677). 17 At 6 months, 78.4% (80/102) of women in the LNG-IUS group reported improvement in spotting, with a median of 0.0 spotting days compared with 2.0 days in the HNR group. While these clinical findings provide important insights, cost-effectiveness is also critical for informing healthcare policy and treatment recommendations. Building on the MIHYS NICHE trial, the present study evaluates the cost-effectiveness of LNG-IUS compared with HNR from a healthcare perspective.

This economic evaluation was conducted alongside a RCT comparing HNR with LNG-IUS for the treatment of niche-related postmenstrual spotting. Trained investigators carried out data collection and monitoring. Eligible participants were women aged 18–48 years presenting with postmenstrual spotting after CS. Inclusion criteria were: (1) MRI showing niche depth ≥2 mm and a residual myometrium ≥2.2 mm, (2) No fertility intentions within 1 year and (3) Willingness to provide informed consent. Exclusion criteria included: contraindications to HNR or LNG-IUS insertion (such as contraindications to anaesthesia, suspected malignancy, atypical endometrial cells or cervical dysplasia), presence of adenomyosis or submucosal leiomyoma causing uterine cavity ≥9 cm, unwillingness to undergo either intervention, positive pregnancy test, existing intrauterine device, use of anticoagulants, endocrine disorders affecting menstrual cycle or irregular cycles (>35 days or inter-cycle variability ≥2 weeks). Participants were recruited at IPMCH affiliated to Shanghai Jiao Tong University between October 2019 and January 2021. Randomisation was performed centrally using a web-based Research Electronic Data Capture system with a block size of 4. Women were assigned to either the LNG-IUS group or the HNR group. In the LNG-IUS group, participants received the 52 mg LNG-IUS (Mirena, Bayer AG, Germany), providing therapeutic hormone release for up to 5 years. Insertions were performed during the non-menstrual phase under ultrasound guidance, without anaesthesia, by trained gynaecologists. 18 An endometrial biopsy was performed prior to insertion. Per protocol, no additional HNR was permitted within the 6-month period. In the HNR group, patients underwent hysteroscopic resection under general anaesthesia in a daycare setting. A resectoscope (Ch. 26 model WA22061 with a 12° optic 22001A) with a 3-mm-deep, 5-mm-wide loop (Olympus, Germany) was used following standard procedures described by Vervoort et al 19 The distal niche outflow tract was resected, and the niche cavity surface was electrocoagulated. Two experienced gynaecologists performed all procedures. Participants were advised to use contraception for 1 year following hysteroscopy. The clinical outcome was effectiveness in reducing postmenstrual spotting at 6 months, defined as ≥50% reduction in spotting days compared with baseline. The effective rate was calculated as the proportion of women achieving this endpoint. Secondary outcomes included changes in menstrual patterns, assessed through patient-maintained menstrual diaries. Follow-up assessments were scheduled at 3, 6, 9 and 12 months after randomisation. Patients completed a menstrual diary to record the intensity of bleeding/spotting according to their subjective impression of the heaviest flow for that day ( online supplemental figure S1 ). The economic evaluation was performed from a healthcare perspective and included only direct medical costs. Data were obtained from hospital billing systems and patient case records, while reference values were taken from previously published studies and price standards published by the National Health Commission. 20 All costs were expressed in RenMinBi (RMB,￥) and adjusted to 2019 values using the Consumer Price Index. The analysis included both surgical costs, such as procedure and anaesthesia fees, as well as other treatment-related expenses, including diagnostic imaging, preoperative examinations and outpatient or inpatient care. A detailed overview of the cost parameters and unit prices is presented in table 1 . CEA from a healthcare perspective for spotting days and effective rate (total=206; levonorgestrel-releasing intrauterine system, n=102; hysteroscopic niche resection, n=104). ICERs were calculated as Δtotal costs/Δeffect. CE, cost-effectiveness; CEA, cost-effectiveness analysis; ICER, incremental cost-effectiveness ratio. Cost-effectiveness was assessed by calculating incremental cost-effectiveness ratios (ICERs). These were determined by dividing the difference in mean costs between the LNG-IUS and HNR groups by the difference in treatment outcomes, measured both as the proportion of women achieving at least a 50% reduction in spotting days and as the actual reduction in spotting days. To account for uncertainty in the estimates, a non-parametric bootstrap with 1000 replications was applied. The results were displayed graphically using cost-effectiveness planes, which illustrate the variation in incremental costs and outcomes, and cost-effectiveness acceptability curves (CEACs), which show the probability that LNG-IUS is more cost-effective than HNR at different willingness-to-pay (WTP) thresholds. All analyses followed the intention-to-treat (ITT) principle. Continuous variables with a normal distribution were compared between groups using Student’s t-test, and results were presented as mean differences with 95% CIs. For continuous variables that did not follow a normal distribution, values were reported as medians with IQRs, and group differences were assessed using the Mann-Whitney U test. For outcomes that were not normally distributed, the between-group effect size was quantified as the median difference, estimated using the Hodges–Lehmann estimator. The corresponding 95% CIs were derived via non-parametric bootstrapping. Specifically, 1000 bootstrap replicates were generated by resampling participants with replacement within each treatment group; the Hodges–Lehmann estimator of the median difference was computed for each replicate, and the 2.5th and 97.5th percentiles of the resulting empirical bootstrap distribution were used to define the two-sided 95% CI. Categorical outcomes were analysed using χ 2 tests, with treatment effects expressed as relative risks and corresponding 95% CIs. Economic outcomes, including mean treatment costs and differences in spotting days, were compared between groups using Student’s t-test. Cost differences were presented with 95% CIs, and uncertainty in cost-effectiveness was further examined using bootstrap analysis, as described above. All statistical analyses were performed using SPSS V.26 (IBM Statistics, Armonk, New York, USA) and Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) p<0.05 represents statistically significant.

Between October 2019 and January 2021, 208 eligible women were recruited and randomised into two groups: 104 in the LNG-IUS group and 104 in the HNR group ( figure 1 ). In the LNG-IUS group, one participant was excluded due to a malignant endometrial pathology result, and another was lost to follow-up. After 6 months, 102 women in the LNG-IUS group and 104 in the HNR group had complete outcome data and were included in the ITT analysis. Baseline demographic and clinical characteristics were well balanced between the two groups ( table 2 ). The median number of baseline spotting days was eight in both groups (IQR 6–10 days). Data are presented as numbers (percentages), as mean (± SD) or as median (IQR). Spotting days/cycle=spotting at the end of menstruation+intermenstrual spotting. Total bleeding days/cycle=menstrual bleeding days+spotting days. BMI, body mass index; HNR, hysteroscopic niche resection; LNG-IUS, levonorgestrel - releasing intrauterine system. At 6 months, 78.4% (80/102) of women in the LNG-IUS group and 73.1% (76/104) in the HNR group achieved at least a 50% reduction in spotting days relative to baseline (RR 1.07, 95% CI 0.92 to 1.25; p=0.370) ( table 3 and online supplemental table S1 ). The reduction in spotting days was significantly greater in the LNG-IUS group, with a median of 0.0 days (IQR 0.0 to 2.8) compared with 2.0 days (IQR 0.8 to 4.3) in the HNR group (p<0.001). Complications and side effects are summarised in online supplemental table S2 . At the 12th-month follow-up, the number of women reporting acne reduced to three, one woman reported breast tenderness, and two still reported weight gain in the LNG-IUS group. Data are presented as means (SE) or medians (IQRs). The outcome measure is the median difference. The outcome measure is relative risk. All cost data were obtained from the hospital medical insurance billing system, which strictly follows the standardised pricing guidelines issued by China’s National Health Commission (2019 Edition). 20 The outcome measure is the mean difference. Other treatment-related costs include diagnostic procedures, preoperative examinations and perioperative care. HNR, hysteroscopic niche resection; LNG-IUS, levonorgestrel - releasing intrauterine system. The mean cost of treatment was substantially lower in the LNG-IUS group (￥2273.8 RMB) compared with the HNR group (￥6318.0 RMB), with a mean cost difference of −￥4044.2 RMB (95% CI −4367.3 to −3721.1). The higher costs in the HNR group were mainly due to preoperative examinations and anaesthesia fees. In contrast, costs in the LNG-IUS group were dominated by the device itself (￥1153 RMB), with no additional anaesthesia or hospitalisation expenses. Cost-effectiveness analysis showed that LNG-IUS was highly cost-effective compared with HNR. The ICERs were estimated at −￥763.2 RMB per one percentage point increase in effectiveness and ￥1833.87 RMB per 1 day reduction in spotting. Most bootstrapped cost-effectiveness pairs (92.8%) fell within the southeast quadrant of the cost-effectiveness plane, indicating that LNG-IUS was dominant over HNR. For effective rate comparisons, 18.1% of samples fell in the southeast quadrant, but the overall pattern still showed greater certainty about cost savings than about differences in effect size ( figure 2 , table 1 ). The CEAC further supported the dominance of LNG-IUS. At a WTP threshold of ￥0 RMB per spotting day reduced, the probability of LNG-IUS being cost-effective compared with HNR was 0.99. At a WTP of ￥10 000 RMB per spotting day reduced, this probability remained 0.60. For effective rate outcomes, there was a 97% probability that LNG-IUS would be considered cost-effective at a WTP ￥10 000 RMB per one percentage point increase in effectiveness.

The economic evaluation, conducted alongside a RCT, compared the cost-effectiveness of LNG-IUS or HNR for the treatment of niche-related bleeding. After 6 months, 78.4% of women in the LNG-IUS group and 73.1% in the HNR group achieved at least a 50% reduction in spotting days. While clinical effectiveness was comparable, the mean cost per woman was significantly lower in the LNG-IUS group, making LNG-IUS the dominant strategy for reducing niche-related postmenstrual spotting. Embedding the economic evaluation alongside a RCT enabled prospective, patient-level measurement of both resource use and clinical outcomes under a standardised protocol, reducing recall bias and improving internal validity. 21 The trial achieved a low loss to follow-up, which strengthens the reliability of estimated costs and effects. Costs were assessed using consistent unit prices (standardised to 2019 levels), and uncertainty in the cost-effectiveness results was quantified using non-parametric bootstrapping and CEACs. Several methodological limitations should be considered. First, this was a single-centre study and costs were valued using Chinese pricing and reimbursement structures. Therefore, transferability of the absolute cost estimates to other settings may be limited, particularly for anaesthesia, diagnostic imaging and procedure-related charges. Second, the analysis was conducted from a healthcare perspective and included only direct medical costs. Finally, some relevant cost components were not measured within the analytic framework, such as contraception costs for participants undergoing HNR, which may affect comparative costs. Previous studies have demonstrated the effectiveness of HNR in improving niche-related bleeding. For example, Vervoort et al reported a 4 day reduction in spotting at 6 months after HNR, 15 which is consistent with the 5 day reduction observed in our trial. However, our study differed in defining treatment effectiveness based on the percentage reduction in spotting, allowing for direct comparison between groups. Evidence for LNG-IUS is more limited, with only two earlier prospective cohort studies (22 and 78 patients, respectively) suggesting benefit. Our trial strengthens this evidence by confirming the effectiveness of LNG-IUS in a larger randomised setting. In addition to comparing the effective rate between the two groups, we assessed the number of spotting days at the sixth month. These two outcome measures complement each other by describing the treatment effect from different statistical perspectives. 22 23 In our trial, the effective rates of LNG-IUS and HNR were not statistically different at 6 months, but LNG-IUS provided a greater overall reduction in spotting days. Previous analyses of this same cohort showed that the therapeutic effect of LNG-IUS improved over time, with superior outcomes at 1 year. 17 These findings suggest that its cost-effectiveness advantage will persist and likely increase with longer follow-up. LNG-IUS also has particular advantages: it avoids surgery, provides contraceptive protection and can be inserted in an outpatient setting by a wider range of healthcare providers, making it more accessible than HNR. In summary, mean costs were significantly lower in the LNG-IUS group (￥2273.8 RMB) compared with the HNR group (￥6318.0 RMB). The ICER of LNG-IUS over HNR was −￥763.21 per one percentage point improvement in effectiveness. Each day of spotting avoided with LNG-IUS was associated with cost savings of ￥1833.87 RMB. Together, these results suggest that LNG-IUS is both less costly and at least as effective as HNR, supporting its use as a first-line treatment for women with niche-related spotting and residual myometrium >2.2 mm who do not desire pregnancy within the following year.