Metabolic Improvements and Weight Trajectories After Metabolic-Bariatric Surgery in Women with Polyendocrine Metabolic Ovarian Syndrome

Abstract

Introduction

Obesity in women with Polyendocrine Metabolic Ovarian Syndrome (PMOS) is associated with substantial metabolic and reproductive burden. metabolic-bariatric surgery (MBS) is effective for weight loss, yet postoperative responses vary considerably. Identifying distinct weight loss trajectories and their associated metabolic outcomes may enable individualized postoperative management.

Methods

In this prospective study, women with obesity and PMOS who underwent sleeve gastrectomy were followed at 1, 3, 6, and 12 months postoperatively. Group-based trajectory modeling (GBTM) based on percentage total weight loss (%TWL) was used to identify distinct patterns of weight loss. Longitudinal changes in anthropometric and metabolic parameters were compared between the identified trajectory groups.

Results

A total of 111 women were included. Significant improvements in anthropometric and metabolic parameters were observed over 12 months, with the most pronounced changes occurring within the first three months. Median BMI decreased from 38.5 kg/m.

² preoperatively to 26.7 kg/m² at 12 months (median %TWL 30.1%; %EWL 83.9%), and HbA1c improved from 5.8% to 5.2%. GBTM identified two distinct %TWL trajectories: an accelerated-decline group (57.7% of patients) and a normal-decline group (42.3%). Despite comparable baseline characteristics between the two groups, patients in the accelerated-decline group demonstrated consistently superior metabolic outcomes, including lower fasting insulin, C-peptide, and LDL-cholesterol at various postoperative time points.

Conclusion

MBS yields substantial metabolic improvements in women with obesity and PMOS, with two distinct weight loss trajectories. Greater weight loss after MBS is associated with more pronounced metabolic benefits, yet preoperative characteristics do not distinguish between trajectory groups, underscoring the heterogeneity in surgical response and the need for further research into the determinants of differential outcomes.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11695-026-08916-7.

Key Points

Significant improvements in HbA1c, fasting glucose, and HDL-C were observed within 3 months after MBS and sustained at 12 months.

Two distinct %TWL trajectories were identified: an accelerated-decline group (57.7% of patients) and a normal-decline group (42.3%).

Despite comparable baseline characteristics between the two groups, the accelerated-decline group demonstrated superior metabolic outcomes, including lower fasting insulin, C-peptide, and LDL-cholesterol.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11695-026-08916-7.

Introduction

Obesity is a growing global public health concern, particularly among women with Polyendocrine Metabolic Ovarian Syndrome (PMOS). PMOS, one of the most common endocrine disorders affecting reproductive-aged women, affects 6% ~ 10% of this population [1]. The prevalence of obesity in women with PMOS is disproportionately high, affecting 50% ~ 80% of patients, which means women with PMOS are three to five times more likely to experience obesity than their healthy counterparts [2]. Obesity not only exacerbates metabolic disturbances in PMOS but also significantly increases health risks. It is reported that women with PMOS have a 2.87-fold increased risk of developing type 2 diabetes, a 1.66-fold increased risk of cardiovascular disease, and a 2.7-fold increased risk of endometrial cancer [3]. Furthermore, obesity adversely affects reproductive function, contributing to ovulatory dysfunction, infertility, and increased fertility-related stress [4]. Therefore, weight management is of paramount importance in the treatment of PMOS.

Metabolic-bariatric surgery (MBS) has emerged as an effective intervention for sustainable weight loss and metabolic improvement in individuals with severe obesity [5, 6]. Accumulating evidence, including recent randomized controlled trials and prospective cohort studies, demonstrates that MBS significantly improves not only body mass index (BMI) and metabolic profiles but also ovulation, menstrual regularity, and hormonal disturbances in women with PMOS [5, 7–9]. Studies have shown that 79.03% of PMOS patients achieved remission of irregular menstruation after surgery, and fasting insulin levels decreased from 34.2 mIU/mL preoperatively to 9.0 mIU/mL postoperatively [5, 7].

However, most existing studies report average treatment effects, overlooking the substantial inter-individual variability in postoperative weight loss and metabolic response. It is estimated that 20 ~ 30% of patients experience suboptimal weight loss or substantial weight recurrence early in the postoperative course [10]. Under this aggregated approach, patients with accelerated postoperative weight loss and those with slower decline are indiscriminately combined, masking potentially meaningful clinical distinctions. Consequently, whether baseline characteristics differ between those who experience accelerated versus slow weight loss remains inadequately investigated. More importantly, how these distinct early weight-loss trajectories correlate with postoperative improvements in metabolic parameters, including glycemic control, insulin resistance, and lipid profiles, has been largely unexplored. Bridging this knowledge gap is essential, as understanding the temporal relationship between weight-loss kinetics and metabolic remission could shift postoperative management from a one-size-fits-all follow-up schedule toward an early response-stratified strategy.

Recognizing this, Group-based trajectory modeling (GBTM) offers a powerful analytical framework to identify distinct subpopulations that follow similar longitudinal patterns [11]. This technique uses finite mixture modeling to group individuals based on their longitudinal trajectories rather than assuming a single average trend, thereby enabling a more nuanced characterization of heterogeneity in postoperative weight loss trajectories.

To date, few studies have applied trajectory modeling to post-MBS weight changes specifically in women with PMOS. Moreover, the relationship between distinct weight-loss trajectory phenotypes and the corresponding longitudinal improvements in metabolic parameters remains largely unexplored in this population. This study therefore aims to identify distinct BMI decline trajectories over the 12 months following MBS in a cohort of Chinese women with obesity and PMOS using GBTM, and to characterize the temporal patterns of anthropometric and metabolic improvements associated with each trajectory. By uncovering these patterns of response, we hope to generate hypotheses that may eventually inform more individualized postoperative care and identify potential windows for targeted intervention.

Methods

Study Design and Participants

This exploratory prospective cohort study was conducted at China-Japan Friendship Hospital in Beijing, between January 2017 and December 2023. The study protocol was approved by the Institutional Ethics Review Committee of China-Japan Friendship Hospital (Approval Number: 2021-112-K70), and written informed consent was obtained from all participants prior to enrollment. The sample size was determined by consecutive enrollment of eligible patients during the study period.

A total of 201 women with obesity and PMOS who underwent MBS were initially screened for eligibility. Participants were eligible if they: (1) were aged 18 ~ 50 years; (2) had a body mass index (BMI) ≥ 28 kg/m²; (3) met the international diagnostic criteria for PMOS (Rotterdam criteria); and (4) underwent sleeve gastrectomy (SG) as MBS procedure. Exclusion criteria included: (1) presence of severe cardiovascular disease or other endocrine disorders (e.g., Cushing’s syndrome, congenital adrenal hyperplasia, androgen-secreting tumors); (2) pregnancy or breastfeeding; (3) previous MBS; and (4) inability or unwillingness to adhere to scheduled follow-up assessments.

After excluding participants with missing baseline data or follow-up records, 111 patients who completed at least one follow-up visit were included in the final analysis (Figure S1). Participants were followed at 1, 3, 6, and 12 months post-surgery.

Surgical Procedures

All included patients underwent SG using a standardized institutional protocol. Experienced bariatric surgeons performed all procedures. A symmetric three‑port technique was used with a 36 Fr bougie. Sequential stapling started 2–6 cm proximal to the pylorus and ended 0.5–1 cm from the His angle, and the staple line was routinely oversewn with absorbable barbed suture.

Data Collection and Measurements

Data collection was performed at baseline (preoperative) and at each follow-up visit (1, 3, 6, and 12 months post-surgery).

Demographic information and medical history, including hypertension and type 2 diabetes mellitus (T2DM), was obtained from medical records.

Anthropometric measurements were performed by trained researchers in the morning with participants in a fasting state and wearing light undergarments. Height was measured to the nearest 1 cm using a wall-mounted stadiometer. Body weight was measured to the nearest 1 kg using a calibrated electronic scale. BMI was calculated as weight in kilograms divided by height in meters squared (kg/m²). Waist circumference was measured at the midpoint between the lower rib margin and the iliac crest using a non-elastic measuring tape, and hip circumference was measured at the widest level over the greater trochanters; both were recorded to the nearest 0.5 cm. Waist-to-hip ratio (WHR) was calculated as waist circumference divided by hip circumference. Percentage total weight loss (%TWL) was calculated as (baseline weight – follow-up weight)/baseline weight × 100. Percentage excess weight loss (%EWL) was calculated as (weight loss/excess weight) × 100, where excess weight was defined as baseline weight minus ideal body weight corresponding to a BMI of 25 kg/m².

Biochemical Measurements

Blood samples were collected after at least 12 h of overnight fasting. Fasting blood glucose, lipids, and uric acid were measured using an AU5800 chemistry analyzer (Beckman Coulter). Fasting insulin, C‑peptide, and reproductive hormones (total testosterone, estradiol, prolactin, FSH, LH) were measured by electrochemiluminescence immunoassay (Cobas e601, Roche Diagnostics). HbA1c was determined by high‑performance liquid chromatography (HLC‑723G11, Tosoh). Postprandial blood glucose, C‑peptide, and insulin were measured at 1 and 2 h after a 75 g glucose challenge.

Statistical Analysis

Continuous variables were tested for normality using the Shapiro–Wilk test. Normally distributed data were presented as mean ± standard deviation (SD), while non-normally distributed data were presented as median with interquartile range (P25 ~ P75). Categorical variables were expressed as frequencies and percentages.

GBTM was applied to identify distinct patterns of %TWL change following MBS [11]. GBTM is a specialized form of finite mixture modeling that identifies latent groups of individuals following similar trajectories over time. The optimal number of trajectory groups and the polynomial order of each trajectory (linear, quadratic, or cubic) were selected based on the Bayesian Information Criterion (BIC), with a preference for parsimony and clinical interpretability. Models with 2 ~ 5 trajectory groups were tested, and the final model was selected based on the lowest BIC value, adequate group size (≥ 5% of the sample), and distinct trajectory shapes.

Linear mixed-effects models were employed to assess changes in metabolic parameters over the follow-up period. Time was treated as a fixed effect, and participants were included as random intercepts to account for within-subject correlation. Post-hoc comparisons between consecutive time points were performed using Tukey’s adjustment for multiple comparisons. A two-tailed P-value < 0.05 was considered statistically significant. To compare the longitudinal trajectories between the identified GBTM groups, we extended the linear mixed-effects models by adding group as a fixed effect, along with the group × time interaction term.

All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA); GBTM was implemented using the PROC TRAJ procedure.

Results

Baseline Characteristics of Participants

Of the 201 women initially screened, 85 were excluded from the final analysis: 6 had missing baseline data, 5 underwent surgical procedures other than sleeve gastrectomy, and 79 were lost to follow‑up. Thus, a total of 111 women were included (Figure S1). The baseline characteristics of the included participants are presented in Table 1. At baseline, participants had a median BMI of 38.5 kg/m², median HbA1c of 5.8%, and median HOMA-IR of 7.16, indicating marked obesity, dysglycemia, and insulin resistance. Detailed baseline characteristics are presented in Table 1.

Table 1.

Basic characteristics of the participants

	Mean ± SD or Median (p25-p75) or N(%)
N	111
Age,year	32.3 ± 7.3
Hypertension, yes(%)	19(18.10)
T2DM, yes(%)	21(20.00)
Height, cm	164.3 ± 5.9
Preoperative characteristics
Body weight, kg	107.0 ± 17.6
BMI,kg/m2	38.5(35.3 ~ 43.7)
Waist circumference, cm	115.0(109.0 ~ 125.0)
Hip circumference, cm	123.0(115.0 ~ 130.0)
WHR	0.95 ± 0.07
Testosterone,nmol/L	2.04 ± 0.84
Estradiol,pmol/L	167.82(138.82 ~ 250.41)
Prolactin,mIU/L	391.82(285.29 ~ 505.80)
FSH,mIU/ml	5.25(3.98 ~ 6.54)
LH,mIU/ml	6.10(4.24 ~ 8.69)
LH/FSH ratio	1.18(0.86 ~ 1.71)
FBG, mmol/L	5.78(5.07 ~ 7.21)
1-h postprandial blood glucose, mmol/L	11.94(9.50 ~ 14.99)
2-h postprandial blood glucose, mmol/L	9.55(7.62 ~ 13.17)
Fasting insulin,μU/mL	25.40(18.59 ~ 39.50)
1-h postprandial insulin,μU/mL	137.50(96.96 ~ 218.50)
2-h postprandial insulin,μU/mL	124.95(84.84 ~ 206.00)
HOMA-IR	7.50(4.86 ~ 11.15)
Fasting C-peptide, ng/mL	3.84(3.33 ~ 5.14)
1-h postprandial C-peptide, ng/mL	11.50(8.40 ~ 14.20)
2-h postprandial C-peptide, ng/mL	11.90(9.19 ~ 15.03)
HbA1c	5.8(5.5 ~ 7.0)
Uric acid,μmol/L	403(352 ~ 487)
TC, mmol/L	4.84(4.29 ~ 5.25)
TG, mmol/L	1.62(1.27 ~ 2.14)
d
LDL-C, mmol/L	2.99(2.61 ~ 3.44)
TG/HDL-C ratio	1.55(1.09 ~ 2.23)

Data are presented as mean ± SD for normally distributed continuous variables, median (P25 ~ P75) for non-normally distributed continuous variables, and N (%) for categorical variables

T2DM type 2 diabetes mellitus; BMI body mass index; WHR waist-to-hip ratio; FSH follicle-stimulating hormone; LH luteinizing hormone; FBG fasting blood glucose; HOMA-IR homeostatic model assessment of insulin resistance; HbA1c glycated hemoglobin; TC total cholesterol; TG triglycerides; HDL-C high-density lipoprotein cholesterol; LDL-C low-density lipoprotein cholesterol

Postoperative Changes in Anthropometric and Metabolic Parameters

As shown in Table 2, significant improvements were observed in most anthropometric and metabolic parameters over the 12‑month follow‑up (all P < 0.001, except 1‑hour postprandial insulin and C‑peptide).

Table 2.

Postoperative changes in metabolic parameters

	Preoperative	1-month Postoperative	3-month Postoperative	6-month Postoperative	12-month Postoperative	P-value for overall change#
BMI,kg/m2	38.51(35.25 ~ 43.69)	34.06(30.82 ~ 39.06)*	30.66(27.69 ~ 35.26)*	28.04(25.41 ~ 31.72)*	26.72(24.13 ~ 30.35)	< 0.0001
%TWL	-	11.9(9.8 ~ 13.9)*	20.7(18.5 ~ 23.3)*	27.3(24.3 ~ 30.5)*	30.1(22.1 ~ 34.8)	< 0.0001
%EWL	-	32.6(23.1 ~ 43.8)*	58.5(44.7 ~ 74.0)*	77.7(64.0 ~ 96.0)*	83.9(66.5 ~ 107.4)	< 0.0001
FBG, mmol/L	5.78(5.07–7.21)	4.97(4.36 ~ 5.68)*	4.84(4.35 ~ 5.31)	4.79(4.42 ~ 5.10)	4.56(4.20 ~ 4.86)	< 0.0001
1-h postprandial blood glucose, mmol/L	11.94(9.50 ~ 14.99)	-	10.25(7.87 ~ 11.06)*	9.44(6.93 ~ 11.38)	8.74(7.27 ~ 10.94)	< 0.0001
2-h postprandial blood glucose, mmol/L	9.55(7.62 ~ 13.17)	-	5.32(4.58 ~ 7.30)*	4.91(4.12 ~ 7.21)	5.76(4.50 ~ 6.70)	< 0.0001
Fasting insulin,μU/mL	25.4(18.59 ~ 39.50)	11.45(7.05 ~ 17.70)*	9.25(7.22 ~ 14.15)	8.11(5.79 ~ 11.30)	8.7(5.61 ~ 11.80)	< 0.0001
1-h postprandial insulin,μU/mL	137.50(96.96 ~ 218.50)	-	79.85(38.20 ~ 156.00)	102.50(90.73 ~ 256.30)	95.70(58.85 ~ 175.00)	1.0
2-h postprandial insulin,μU/mL	124.95(84.84 ~ 206.00)	-	20.80(12.20 ~ 61.68)*	62.00(17.16 ~ 91.37)	38.26(20.70 ~ 89.35)	< 0.0001
Fasting C-peptide, ng/mL	3.84(3.33 ~ 5.14)	2.68(1.63 ~ 3.31)*	2.52(1.86 ~ 3.09)	2.12(1.68 ~ 2.55)	1.93(1.56 ~ 2.49)	0.0004
1-h postprandial C-peptide, ng/mL	11.50(8.40 ~ 14.20)		10.20(6.71 ~ 13.00)	11.70(10.01 ~ 15.50)	10.40(8.32 ~ 13.45)	0.21
2-h postprandial C-peptide, ng/mL	11.90(9.19 ~ 15.03)		5.48(3.71 ~ 8.82)*	7.22(5.39 ~ 12.73)	7.17(5.57 ~ 10.60)	0.0004
HbA1c	5.8(5.5 ~ 7.0)	5.5(5.1 ~ 5.8)*	5.2(5.0 ~ 5.5)	5.2(4.9 ~ 5.4)	5.2(5.0 ~ 5.4)	< 0.0001
Uric acid,μmol/L	403.0(352.0 ~ 487.0)	422.0(359.0 ~ 509.0)	347.5(301.0 ~ 403.0)*	319.0(274.0 ~ 370.0)	319.0(280.2 ~ 359.0)	< 0.0001
TC, mmol/L	4.84(4.29 ~ 5.25)	4.25(3.80 ~ 4.77)	4.84(4.36 ~ 5.70)	4.87(4.40 ~ 5.02)	4.55(4.03 ~ 4.97)	< 0.0001
TG, mmol/L	1.62(1.27 ~ 2.14)	1.22(1.07 ~ 1.48)*	1.14(0.94 ~ 1.45)	1.07(0.82 ~ 1.20)	0.76(0.58 ~ 1.10)	< 0.0001
HDL-C, mmol/L	1.06(0.94 ~ 1.17)	0.92(0.76 ~ 1.03)*	1.10(0.98 ~ 1.19)*	1.20(1.03 ~ 1.28)	1.32(1.16 ~ 1.57)*	< 0.0001
LDL-C, mmol/L	2.99(2.61 ~ 3.44)	2.70(2.29 ~ 3.15)*	3.05(2.76 ~ 3.51)*	2.87(2.69 ~ 3.16)	2.64(2.27 ~ 3.06)	0.0002

Data are presented as median (P25 ~ P75). #P-value for overall change over time assessed by linear mixed-effects models. *P < 0.05 vs. previous time point (Tukey’s post-hoc comparison). BMI body mass index; WHR waist‑to‑hip ratio; %TWL percentage total weight loss; %EWL percentage excess weight loss; FBG fasting blood glucose; 1 h-PG 1-h postprandial glucose; 2 h-PG 2-h postprandial glucose; HbA1c glycated hemoglobin; HDL-C high-density lipoprotein cholesterol; LDL-C low-density lipoprotein cholesterol

The most pronounced changes occurred within the first 3 months after surgery. Median BMI decreased from 38.5 kg/m² preoperatively to 30.7 kg/m² at 3 months, corresponding to a median %TWL of 20.7% and %EWL of 58.5% by 3 months. Median HbA1c improved from 5.8% to 5.2% over the same period (both P < 0.001). Fasting glucose, fasting C‑peptide, and 2‑hour postprandial glucose also showed accelerated reductions within the first month, with further modest decreases by 3 months.

Beyond 3 months, most parameters remained stable or continued to improve at a slower rate. Median %TWL reached 30.1% by 12 months, and median %EWL reached 83.9%. Median BMI gradually declined from 30.7 kg/m² at 3 months to 26.7 kg/m² at 12 months. HDL‑C showed a unique pattern: after an initial decrease at 1 month (from 1.06 to 0.92 mmol/L), it progressively increased from 3 months onward, reaching 1.32 mmol/L at 12 months. LDL‑C and uric acid decreased significantly over the entire follow‑up period.

Identification of BMI Trajectories

Group-based trajectory modeling was applied to identify distinct patterns of %TWL over the 12-month follow-up period following MBS. Based on the BIC, adequate group size (≥ 5% of the sample), and clinical interpretability, a two-trajectory model was selected as the optimal solution.

As shown in Fig. 1, the two trajectory groups were characterized as follows: the accelerated-decline group (n = 64, 57.7%) and the normal-decline group (n = 47, 42.3%). In the accelerated-decline group, patients experienced a more pronounced and accelerated %TWL increase during the early postoperative period, reaching 11.9% at 1 month, 20.7% at 3 months, and 30.1% at 12 months. In contrast, the normal-decline group exhibited a relatively slower rate of weight loss, with %TWL reaching 6.6% at 1 month, 13.8% at 3 months, and 19.7% at 12 months. Although both groups achieved substantial weight loss by 12 months, the accelerated-decline group consistently demonstrated significantly greater %TWL at each follow-up time point compared to the normal-decline group (all P < 0.05). The mean posterior probability for the accelerated-decline group was 0.74, and for the normal-decline group was 0.80, indicating high classification certainty.

Fig. 1.

BMI trajectories after metabolic-bariatric surgery. Group-based trajectory modeling identified two distinct trajectories of percentage total weight loss (%TWL) over the 12-month postoperative period: the accelerated-decline group (n = 64, 57.7%) and the normal-decline group (n = 47, 42.3%). A %TWL trajectories identified by GBTM. B–L Longitudinal changes in anthropometric and metabolic parameters, including (B) BMI, (C) %EWL, (D) FBG, (E) fasting insulin, (F) fasting C-peptide, (G) HbA1c, (H) uric acid, (I) TC, (J) TG, (K) HDL-C, and (L) LDL-C, stratified by trajectory group. Data are presented as means with STD. *P < 0.05 between groups at the same time point

Baseline Characteristics and Postoperative Differences Between Trajectory Groups

The baseline characteristics of the two trajectory groups were comparable across all measured demographic, anthropometric, hormonal, and metabolic parameters (all P > 0.05; Table 3), indicating that preoperative phenotyping did not distinguish patients who would subsequently experience accelerated versus normal weight loss.

Table 3.

Univariate logistic regression analysis of preoperative factors associated with accelerated BMI decline trajectory

	Accelerated-decline group	Normal-decline group	P-value
N	64	47
Age,year	31.30 ± 6.98	33.64 ± 7.63	0.101
Hypertension, yes(%)	8(12.50)	11(23.40)	0.132
T2DM, yes(%)	13(20.31)	8(17.02)	0.662
Height, cm	164.62 ± 5.75	163.85 ± 6.23	0.509
Preoperative characteristics
Body weight, kg	106.76 ± 18.38	107.34 ± 16.67	0.863
BMI,kg/m2	37.64(35.12 ~ 44.19)	38.51(35.94 ~ 43.06)	0.527
Waist circumference, cm	116.5(109.0 ~ 125.5)	115.0(107.0 ~ 125.0)	0.571
Hip circumference, cm	124.0(115.0 ~ 134.0)	121.0(115.0 ~ 128.0)	0.301
Testosterone,nmol/L	2.12 ± 0.88	1.95 ± 0.79	0.301
Estradiol,pmol/L	170.65(143.11 ~ 256.49)	155.13(136.46 ~ 222.30)	0.502
Prolactin,mIU/L	350.22(273.23 ~ 482.01)	408.9(309.08 ~ 552.59)	0.174
FSH,mIU/ml	5.35(4.09 ~ 7.17)	5.14(3.85 ~ 6.15)	0.232
LH,mIU/ml	5.86(4.24 ~ 8.27)	6.24(4.29 ~ 8.69)	0.694
LH/FSH ratio	1.13(0.81 ~ 1.71)	1.37(0.91 ~ 1.67)	0.259
FBG, mmol/L	5.79(5.15 ~ 6.88)	5.72(4.97 ~ 7.45)	0.986
1-h postprandial blood glucose, mmol/L	11.86(9.365 ~ 15.06)	12.14(9.54 ~ 14.23)	0.948
2-h postprandial blood glucose, mmol/L	9.26(7.52 ~ 13.03)	9.97(7.62 ~ 14.38)	0.732
Fasting insulin, μU/mL	26.9(18.75 ~ 36.4)	24.8(18.22 ~ 41.7)	0.895
1-h postprandial insulin, μU/mL	140.2(85.3 ~ 231.4)	135(98.93 ~ 199.9)	0.731
2-h postprandial insulin, μU/mL	127(84.84 ~ 209.8)	116.7(84.4 ~ 205.0)	0.696
HOMA-IR	7.55(4.67 ~ 11.18)	7.36(5.01 ~ 11.15)	0.718
Fasting C-peptide, ng/mL	3.82(3.26 ~ 5.01)	3.86(3.38 ~ 5.29)	0.743
1-h postprandial C-peptide, ng/mL	11.5(8.12 ~ 14.2)	11.50(8.45 ~ 14.24)	0.937
2-h postprandial C-peptide, ng/mL	12.20(9.06 ~ 15.03)	11.70(9.24 ~ 15.10)	0.942
HbA1c, %	5.9(5.5 ~ 6.8)	6.0(5.5 ~ 7.6)	0.479
Uric acid, μmol/L	393.5(364.0 ~ 473.0)	413.0(341.0 ~ 496.0)	0.876
TC, mmol/L	4.67(4.23 ~ 5.18)	5.00(4.44 ~ 5.38)	0.051
TG, mmol/L	1.51(1.11 ~ 2.12)	1.71(1.31 ~ 2.17)	0.295
HDL-C, mmol/L	1.04(0.90 ~ 1.13)	1.07(0.98 ~ 1.18)	0.158
LDL-C, mmol/L	2.87(2.56 ~ 3.32)	3.07(2.72 ~ 3.51)	0.134

Data are presented as mean ± standard deviation (SD) for normally distributed continuous variables, or as median (interquartile range, IQR) for non‑normally distributed continuous variables. Categorical variables are presented as number (percentage). Univariate logistic regression analysis was performed to identify preoperative factors associated with the accelerated BMI decline trajectory. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported

BMI body mass index; T2DM type 2 diabetes mellitus; FSH follicle‑stimulating hormone; LH luteinizing hormone; FBG fasting blood glucose; HOMA‑IR homeostatic model assessment for insulin resistance; HbA1c glycated hemoglobin; TC total cholesterol; TG triglyceride; HDL‑C high‑density lipoprotein cholesterol; LDL‑C low‑density lipoprotein cholesterol

Despite the absence of baseline differences, the two groups exhibited distinct postoperative metabolic trajectories. As shown in Fig. 1, significant between-group differences emerged as early as 1 month postoperatively and persisted at various time points throughout the 12-month follow-up. Specifically, the accelerated-decline group demonstrated significantly lower fasting insulin levels at 1 month (11.77 ± 6.65 vs. 14.66 ± 7.69 μU/mL, P < 0.05), 6 months (7.98 ± 4.70 vs. 11.82 ± 8.75 μU/mL, P < 0.05), and 12 months (10.48 ± 13.99 vs. 12.34 ± 7.86 μU/mL, P < 0.05). Fasting C-peptide levels were significantly lower in the accelerated-decline group at 6 months (1.94 ± 0.52 vs. 2.41 ± 0.70 ng/mL, P < 0.05) and 12 months (1.76 ± 0.53 vs. 2.46 ± 0.76 ng/mL, P < 0.05). Uric acid levels differed significantly at 1 month (485.10 ± 129.96 vs. 406.37 ± 150.78 μmol/L, P < 0.05) and 3 months (383.87 ± 89.35 vs. 315.64 ± 87.86 μmol/L, P < 0.05). Total cholesterol was significantly lower in the accelerated-decline group at 1 month (4.16 ± 0.88 vs. 4.52 ± 0.90 mmol/L, P < 0.05) and 3 months (4.65 ± 0.92 vs. 5.32 ± 0.80 mmol/L, P < 0.05). LDL-cholesterol was significantly lower in the accelerated-decline group at 3 months (2.89 ± 0.72 vs. 3.58 ± 0.92 mmol/L, P < 0.05) and 12 months (2.41 ± 0.57 vs. 3.02 ± 0.60 mmol/L, P < 0.05).

Discussion

In this prospective cohort study of Chinese women with obesity and PMOS who underwent MBS, we observed significant improvements in anthropometric and metabolic parameters over the 12-month postoperative period, with the most pronounced changes occurring within the first three months. Notably, we identified two distinct %TWL trajectories: an accelerated‑decline group and a normal‑decline group using GBTM. Importantly, despite comparable baseline characteristics across all measured demographic, anthropometric, hormonal, and metabolic parameters between the two groups, significant metabolic divergence emerged as early as 1 month postoperatively and persisted throughout the 12-month follow-up. The accelerated-decline group demonstrated consistently superior metabolic profiles, including lower fasting insulin, C-peptide, uric acid, and LDL-cholesterol at various postoperative time points, suggesting that early weight-loss kinetics may serve as a functional indicator of metabolic plasticity.

The significant improvements observed in our study are consistent with prior research. A recent systematic review and meta-analysis by Benham et al., demonstrated that women with PMOS experience similar improvements in anthropometric, hormonal, and metabolic outcomes after MBS compared to those without PMOS [12]. Similarly, a meta-analysis by Hu et al., confirmed that MBS is superior to medical therapy in improving BMI, insulin resistance, and menstrual regularity in women with PMOS and obesity [13].

In terms of specific metabolic improvements, our findings align with a large prospective cohort study by Li et al. [14], which reported significant reductions in BMI, HbA1c, and fasting C-peptide by 3 months post-surgery, followed by a plateau phase, which mirrored our observations. Regarding HDL-C, our study demonstrated progressive increases beginning at 3 months and continuing through 12 months, consistent with Li et al. [14] and the BAMBINI trial [15], but contrasting with an earlier study by Turkmen et al., which showed no significant change at 6 months post-RYGB) [16]. This discrepancy may be due to differences in sample size, surgical technique (RYGB vs. SG in our cohort), population characteristics, or follow-up duration (12 vs. 6 months).

A key contribution of our study is the application of GBTM to characterize heterogeneity in postoperative weight loss. GBTM empirically identified two distinct %TWL trajectories, revealing that approximately half of the patients achieved accelerated weight loss, while the other half exhibited a slower but still substantial decline. Notably, despite virtually identical preoperative characteristics across all measured parameters, these two groups diverged significantly in their postoperative metabolic outcomes. Patients in the accelerated-decline group consistently demonstrated superior metabolic profiles, including lower fasting insulin, C-peptide, uric acid, and LDL-cholesterol at various time points, as well as more favorable lipid profiles. These findings suggest that greater weight loss after MBS is associated with more pronounced metabolic improvements, reinforcing the central role of weight reduction as a driver of metabolic remission in PMOS. The emergence of significant metabolic divergence as early as 1 month postoperatively highlights a critical window for early risk stratification; this early time point could serve as a clinical checkpoint to identify patients who may benefit from intensified lifestyle or pharmacological interventions to optimize long‑term outcomes. The fact that preoperative phenotyping could not predict trajectory membership further implies that the determinants of differential weight loss responses may lie in postoperative physiological adaptations, such as variations in appetite regulation, gut hormone secretion, or energy expenditure [17]. Indeed, good responders to bariatric surgery have been shown to exhibit significantly greater postprandial increases in GLP-1 and PYY compared to poor responders, and adaptive thermogenesis following weight loss may further contribute to inter-individual differences in weight loss outcomes [16, 18].

Several limitations should be acknowledged. First, the high loss to follow‑up rate may introduce selection bias, although baseline characteristics did not differ significantly between included and excluded patients. Second, the follow‑up period was limited to 12 months, which is insufficient to assess long‑term weight maintenance or trajectory stability. Third, this was a single-center study conducted at a specialized bariatric center, which may limit the generalizability of our findings to other populations or healthcare settings. Fourth, the study lacked data on several important outcomes, including body composition (e.g., fat mass, lean mass), which could have provided a more comprehensive assessment of weight loss quality, and reproductive outcomes such as menstrual regularity, ovulation, and hirsutism, which are central to PMOS management. Fifth, we did not collect detailed information on potential confounders that may influence postoperative weight loss, including postoperative dietary adherence, physical activity levels, and medication use (e.g., metformin, GLP-1 receptor agonists), which could partially account for the inter-individual variability observed. Notably, although the two trajectory groups were virtually identical in all measured baseline characteristics, we cannot exclude the possibility that unmeasured factors—such as genetic predisposition, psychosocial factors, or differential postoperative behavioral adaptations—may have contributed to the divergent weight loss responses. Finally, given the observational and exploratory nature of this study, causal inferences cannot be drawn, and our findings should be considered hypothesis-generating. These limitations should be addressed in future multicenter prospective studies with longer follow-up, larger sample sizes, and comprehensive assessments of both metabolic and reproductive outcomes.

In summary, this exploratory study provides preliminary evidence that MBS yields substantial and sustained metabolic improvements in Chinese women with obesity and PMOS. Using GBTM, we identified two distinct %TWL trajectories over the 12-month postoperative period: an accelerated-decline group and a normal-decline group, with the accelerated-decline group consistently demonstrating superior insulin sensitivity and lipid profiles. These findings underscore the heterogeneity in surgical response among patients with PMOS and suggest that greater weight loss after MBS is associated with more pronounced metabolic benefits. Given the exploratory and observational nature of this study, these results require confirmation in future multicenter prospective studies with longer follow-up and comprehensive assessments of both metabolic and reproductive outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contribution

L.Z. and X.L. contributed equally to this work. L.Z. and X.L. conceptualized the study, conducted the prospective cohort study, including participant recruitment, data collection, curation and drafted the manuscript, performed statistical analyses, and contributed to the result interpretation, and draft the manuscript. S.Z., B.Z., B.H., Y.G., and C.W. contributed to the conduction of the prospective cohort study, including participant recruitment, data collection, and curation. H.M. led the project, supervised the project and provided critical revisions to the manuscript. All authors contributed to data interpretation, manuscript editing, and approved the final version.

Funding

This work was supported by Capital's funds for health improvement and research (2026-2-4063), General Program of National Natural Science Foundation of China (82470857), Beijing Natural Science Foundation (7242125), China-Japan Friendship Hospital National High Level Hospital Clinical Research Funding (2025-NHLHCRF-DLYJ-RCT-05), “Kunlun Talents · High-end Innovation and Entrepreneurial Talents” project of Qinghai Province (202308210043).

Data Availability

The data of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare that they have no conflict of interest.