Factors associated with resection weight and complications after massive-weight-loss abdominoplasty: A retrospective 10-year single-center study Authors: Sabrina Luettschwager, Patrick Mandal, Lars- Peter Kamolz, Hanna Luze, Maximilian Moshammer, Michael Schintler, Michael Kohlhauser Published: 2026 ABSTRACT HighlightsBackgroundAbdominoplasty removes redundant skin and subcutaneous tissue but poses perioperative challenges. Although higher resection weight has been linked to complications, variation in resection burden and its relationship to post-operative complications remain incompletely defined. This exploratory study assessed patient- and procedure-related factors associated with resection weight and evaluated whether resection weight was associated with post-operative complications.MethodsThis retrospective cohort included 217 adults undergoing abdominoplasty after massive weight loss (MWL; excess weight loss ≥ 50%) at the Medical University of Graz (2012–2023). Pregnancy-related and oncologic procedures, isolated revisions, and incomplete records were excluded. Variables included weight history, resection weight, post-operative complications (Clavien–Dindo), demographics, and procedure-related factors. Correlation, group comparison, multivariable linear and logistic regression analyses were applied.ResultsMean age was 45 years (SD 12); 76% were female. Mean pre-operative BMI was 26.3 kg/m2 (SD 3.5) and mean resection weight was 1,727 g (SD 1,132). Pre-operative BMI showed the strongest bivariate association with resection weight (Spearman r = 0.620, p < 0.001). In multivariable linear regression, pre-operative BMI (B = 209.9 g per kg/m², p < 0.001), female sex (+363.5 g, p = 0.004), metabolic disease (+346.0 g, p = 0.006), and concomitant belt lipectomy (+365.6 g, p < 0.001) were associated with higher resection weight, explaining 56% of variance. Post-operative complications occurred in 76 patients (35.0%). In adjusted logistic regression, younger age (OR 0.969 per year, p = 0.016) and pre-operative anemia (OR 2.153, p = 0.043) were associated with the composite complication endpoint; resection weight was not (OR 1.015 per 100 g, p = 0.245).ConclusionPre-operative BMI was the strongest factor associated with resection weight, with female sex, metabolic disease, and procedural extent adding explanatory value. Resection weight was not significantly associated with composite, major, or wound-specific complications; however, smaller effects could not be excluded.Keywords: Abdominoplasty, Resection weight, Complications, Clavien–Dindo, Post-bariatric surgery, Massive weight loss (MWL) Abbreviations American Society of Anesthesiologists (classification) Body mass index Confidence interval Excess weight loss Fleur-de-lis Massive weight loss Minimum detectable effect Odds ratio Coefficient of determination Standard deviation Variance inflation factor Beta coefficient Pre-operative body mass index Maximum body mass index Maximum body mass index – pre-operative body mass index Interquartile range Introduction Body contouring surgery, particularly abdominoplasty, is increasingly performed after massive weight loss (MWL) following bariatric surgery or lifestyle changes.1,2 Abdominoplasty removes redundant abdominal skin and subcutaneous tissue and is among the most common plastic surgery procedures worldwide.3, 4, 5 Approximately 1.2 million abdominoplasties were performed globally in 2023.6 This trend parallels the obesity burden: in 2022, more than 2.5 billion adults were overweight, including nearly 900 million with obesity (body mass index [BMI] ≥ 30 kg/m²).7 Redundant skin after MWL affects up to 96% of post-bariatric patients and imposes physical and psychosocial burden.8, 9, 10 Abdominoplasty addresses these issues but carries a higher complication rate than other body contouring procedures.11 Prior studies have linked resection weight, BMI, and method of weight loss to post-operative outcomes.12, 13, 14 However, reported risk estimates and proposed resection-weight threshold (commonly ≥ 2 kg) vary across studies, reflecting differences in case mix and study design.15, 16, 17 Although multiple predictors of complications have been described, a clinically applicable, multivariable estimate of expected resection weight, integrating pre-operative anthropometry, sex, comorbidity, and procedural extent, has not been systematically evaluated. The relative contribution of maximum BMI, pre-operative BMI, and delta BMI (maximum BMI – pre-operative BMI) remains uncertain. The aims of this exploratory retrospective cohort study were (i) to assess patient- and procedure-related factors associated with resection weight and (ii) to evaluate its adjusted association with post-operative complications. Additional analyses explored differences by weight loss method and procedural extent, potentially supporting surgical planning, counselling and benchmarking.18 Patients and methods Study design This single-center retrospective study was conducted in accordance with the Declaration of Helsinki and reported using the STROBE guidelines. Ethical approval was granted by the Ethics Board of the Medical University of Graz (EK-Nr.: 36-118 ex 23/24). Patients and data collection We identified 642 consecutive adults (≥18 years) who underwent abdominoplasty at the Clinical Department of Plastic, Aesthetic and Reconstructive Surgery, Graz, between 1 January 2012 and 31 December 2023. Inclusion required surgical treatment of a symptomatic abdominal panniculus after MWL, defined as excess weight loss (EWL) ≥50% (excess weight = maximum body weight – weight at BMI 25 kg/m2).19 Indications included functional impairment (e.g., recurrent intertrigo, hygiene difficulties, limitations in daily activities) or clinically relevant contour deformity after weight stabilization. Only patients with Austrian statutory insurance were included for documentation consistency. Exclusion criteria were pregnancy-related abdominoplasty, oncologic reconstruction (e.g., DIEP, TRAM), isolated secondary procedures (herniotomy, scar revision, contour refinement), incomplete documentation of key variables, and EWL <50%. Of 642 patients screened, 217 met all criteria (Fig. 1). Flowchart of patient selection. MWL = Massive weight loss. Variables Data were retrieved from medical and surgical reports, inpatient and outpatient records, and pre-operative anesthetic documentation. Patient-related variables comprised age (years), biological sex, maximum BMI, pre-operative BMI, and delta BMI, total weight loss (kg), weight-loss duration (years), and weight-loss method (post-bariatric vs. non-bariatric). Clinical covariates included ASA score, smoking status, metabolic disease (diabetes, hypertension, hypercholesterolemia, dyslipidemia, or hepatic steatosis), diabetes, cardiovascular disease, previous abdominal surgery, anticoagulation, steroid use, pre- and post-operative anemia (hemoglobin <12 g/dL in females and <13 g/dL in males), and operative time (minutes). Resection weight was defined as the specimen weight attributable exclusively to the abdominoplasty component, measured intra-operatively after en-bloc excision and before wound closure. In combined procedures, only the abdominoplasty specimen was recorded. Liposuction aspirate was excluded; residual tumescent fluid may have affected measurements to an undetermined but likely small degree. Complications were classified using Clavien–Dindo and captured during the inpatient stay and all documented follow-up at the study center (median follow-up 3 months).20 Complications managed elsewhere were not systematically captured. Wound-specific complications comprised wound healing disorder or dehiscence, seroma, hematoma or post-operative bleeding, skin or fat necrosis, wound infection, fistula or suture-related complications, and medically indicated wound revision, debridement, or necrosectomy. Complications were analyzed at the patient level using the highest applicable Clavien–Dindo grade per patient. Purely aesthetic revisions, dog-ear corrections, contour irregularities, scar revisions without wound pathology or functional impairment, and isolated hernia findings without wound-related intervention were not classified as post-operative complications. Complication documentation was reviewed from operative reports, wound documentation, and follow-up notes. The primary complication classification was performed by a single physician investigator based on the predefined criteria. An independent physician reviewed anonymized complication descriptions while blinded to the initial classification and patient- and procedure-related variables. Exact agreement was 213/217 cases (98.2%). Cohen’s κ was 0.960 for the binary endpoint, and linear weighted κ was 0.951 (95% bootstrap CI, 0.897–0.990) for ordinal grading. Four discrepant aesthetic or contour-related revisions were resolved by consensus as non-complications. Weight loss duration was missing in 103/217 (47%) and was excluded from multivariable models; all other variables had <5% missingness and were analyzed using complete case analysis. Surgical technique All procedures were performed under general anesthesia under board-certified plastic-surgeon supervision. The standard procedure was transverse abdominoplasty with umbilical transposition; a fleur-de-lis (FDL) variant was performed for pronounced vertical excess (n = 18; 8%). Rectus diastasis was corrected by midline plication when present (n = 44; 20%). Redundant skin and subcutaneous tissue were excised en bloc. Concomitant procedures included belt lipectomy (n = 115; 53%), thigh lift (n = 55; 25%), breast lift (n = 13; 6%), thoracic lift (n = 2; 0.9%), brachioplasty (n = 2; 0.9%), and lower leg lift (n = 1; 0.5%), with overlap. Procedural heterogeneity was addressed by covariate adjustment and an isolated abdominoplasty sensitivity analysis. Two closed-suction abdominal drains were used in 97% of cases. Statistics Analysis was performed in IBM® SPSS® (v29; IBM Corp., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation (SD) or median with interquartile range (IQR). Because several continuous variables were non-normally distributed, Spearman correlations and Mann–Whitney U tests were used as appropriate; chi-square tests were used for categorical comparisons. Factors associated with resection weight were assessed using multivariable linear regression. The primary model included pre-operative BMI, age, biological sex, metabolic disease, and belt lipectomy, selected based on clinical plausibility. Collinear anthropometric variables were not entered simultaneously; an alternative model substituted pre-operative weight for pre-operative BMI. Assumptions were assessed using residual and Q-Q plots, variance inflation factors (VIFs), Cook's distance, and leverage. Log-transformed resection weight was examined in sensitivity analysis. Estimates are reported with 95% confidence intervals (CI) and model fit with R2, adjusted R2, and the F-test. Candidate variables associated with post-operative complications were screened using crosstabs and univariable logistic regression. Resection weight was retained a priori as the predefined hypothesis variable. Age and pre-operative anemia were selected based on clinical plausibility and univariable signal strength; all screened variables are reported in Table 8a. Resection weight was scaled per 100 g. The composite and wound-specific models include resection weight, age and pre-operative anemia. The major complication model included resection weight, age and pre-operative BMI; major complications were defined as Clavien–Dindo grade ≥ IIIa and occurred in 47 patients. An isolated abdominoplasty sensitivity analysis was performed. Table 8aScreening of candidate factors for post-operative complications.Table 8a dummy alt textPredictorScreening methodEffect estimate / raw ratep-valueRetained for modelAge (per year)Univariable logistic regressionOR 0.9710.020YesPre-operative anemiaCrosstab / χ²51.4% vs. 31.9%0.026YesResection weight (per 100 g)Univariable logistic regressionOR 1.0070.594Yes*Cardiovascular diseaseCrosstab + univariable logistic regressionOR 0.515; 24.6% vs. 38.8%0.054 / 0.056NoPost-op anticoagulationCrosstab / χ²23.1% vs. 37.6%0.084NoSexCrosstab / χ²33.5% vs. 39.6%0.419NoWeight loss methodCrosstab + univariable logistic regressionOR 0.789; 31.8% vs. 37.1%0.419 / 0.420NoSmokingCrosstab / χ²32.8% vs. 37.6%0.454NoDiabetes mellitusCrosstab / χ²38.1% vs. 34.7%0.756NoLiposuctionCrosstab / χ²50.0% vs. 33.3%0.120NoBelt lipectomyCrosstab / χ²34.5% vs. 35.6%0.858NoBMI pre-op (per kg/m²)Univariable logistic regressionOR 0.9960.912NoMetabolic diseaseCrosstab / χ²30.8% vs. 37.4%0.325NoASA ScoreCrosstab / χ²No clear gradient0.690NoPrevious abdominal surgeryCrosstab / χ²34.5% vs. 35.6%0.870NoRectus plicationCrosstab / χ²25.6% vs. 37.4%0.147NoFleur-de-lis (FDL)Crosstab / χ²38.9% vs. 34.7%0.720NoQuilting suturesCrosstab / Fisher14.3% vs. 36.5%0.146NoSteroid useCrosstab / Fisher40.0% vs. 34.9%1.000NoAlcohol useCrosstab / χ²38.5% vs. 29.9%0.194NoThigh liftCrosstab / χ²29.6% vs. 36.8%0.338NoBreast liftCrosstab / Fisher30.8% vs. 35.3%1.000No⁎Retained as the predefined primary hypothesis variable despite the absence of a univariable signal.BMI pre-op = pre-operative body mass index. Screening of candidate factors for post-operative complications. Retained as the predefined primary hypothesis variable despite the absence of a univariable signal. BMI pre-op = pre-operative body mass index. Given the retrospective design and fixed cohort, analyses were interpreted as exploratory associations rather than confirmatory tests or definitive prediction models. Tests were two-sided, and p < 0.05 was considered statistically significant. No a priori sample size calculation was performed. Model precision was contextualized using minimum detectable effect (MDE) analyses. For the linear models, the smallest partial effect detectable with 80% power at α = 0.05 was f² = 0.037. For the logistic models, MDE odds ratios for resection weight per 100 g were estimated using the method of Hsieh, Bloch and Larsen.21 Calculations were based on the observed standard deviation of resection weight and an auxiliary R2 of 0.20. Results Cohort characteristics Of 642 patients screened, 217 met inclusion criteria (Fig. 1). Mean age was 45 ± 12 years; 164 (76%) were female. Mean pre-operative BMI was 26.3 ± 3.5 kg/m², maximum BMI 43.1 ± 6.8 kg/m², and delta BMI 16.8 ± 5.9 kg/m². Eighty-five patients (39%) had undergone bariatric surgery, and 132 patients (61%) achieved non-surgical weight loss. Mean weight loss was 48 ± 17 kg, and mean resection weight was 1,727 ± 1,132 g (range 87–6,980 g) (Table 1). FDL was performed in 18 (8%), rectus plication in 44 (20%), quilting/progressive-tension sutures in 14 (6%), and liposuction in 22 (10%). Belt lipectomy was performed in 115 (53%); 158 patients (73%) underwent at least one concomitant procedure (Table 2a). Among the listed concomitant body-contouring procedures excluding liposuction, 66 patients (30.4%) had none, 114 (52.5%) had one, and 37 (17.1%) had two; no patient had three or more (Table 2b). Table 1Patient cohort characteristics.Table 1 dummy alt textTotal (N = 217)Age (years)*45 ± 12 (21–79)Sex Female164 (76%) Male53 (24%)BMI max (kg/m²)*43.1 ± 6.8 (25.8–70.2)BMI delta (kg/m²)*16.8 ± 5.9 (5.3–40.9)BMI pre-op (kg/m²)*26.3 ± 3.5 (17.3–42.6)Weight loss (kg)*48 ± 17 (15–110)Weight loss method Non-bariatric132 (61%) Post-bariatric85 (39%)Smoking No101 (47%) Yes116 (53%)Metabolic disease No139 (64%) Yes78 (36%)Pre-operative anemia No182 (84%) Yes35 (16%)Post-operative complication No141 (65%) Yes76 (35%)Resection weight (g)*1727 ± 1132 (87–6980) Female1817 ± 1200 (87–6980) Male1452 ± 841 (262–5000)Total (N = 114)Time period of weight loss (years)⁎⁎3.0 (IQR 2.0–5.0)⁎Values are presented as mean ± SD, with range in parentheses⁎⁎Values are presented as median with interquartile range (IQR) in parentheses; BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative body mass index; BMI pre-op = pre-operative body mass index. Patient cohort characteristics. Values are presented as mean ± SD, with range in parentheses Values are presented as median with interquartile range (IQR) in parentheses; BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative body mass index; BMI pre-op = pre-operative body mass index. Table 2aAbdominoplasty procedure characteristics.Table 2a dummy alt textTotal (N = 217)Fleur-de-lis Yes18 (8%)Rectus plication Yes44 (20%)Quilting/progressive tension sutures Yes14 (6%)Liposuction Yes22 (10%)Belt lipectomy Yes115 (53%)Thoracic lift Yes2 (0.9%)Brachioplasty Yes2 (0.9%)Breast lift Yes13 (6%)Lower leg lift Yes1 (0.5%)Thigh lift Yes55 (25%)Concomitant procedures Yes158 (73%) Abdominoplasty procedure characteristics. Table 2bCombinations of concomitant body-contouring procedures excluding liposuction.Table 2b dummy alt textConcomitant procedure combination with abdominoplastyn (% of total cohort)No listed concomitant body-contouring procedure excluding liposuction66 (30.4%)Belt lipectomy only78 (35.9%)Belt lipectomy + thigh lift30 (13.8%)Thigh lift only25 (11.5%)Breast lift only8 (3.7%)Belt lipectomy + breast lift5 (2.3%)Thoracic lift only2 (0.9%)Brachioplasty only1 (0.5%)Belt lipectomy + brachioplasty1 (0.5%)Belt lipectomy + lower leg lift1 (0.5%)Total217 (100%)All patients underwent abdominoplasty. Liposuction was reported separately as an adjunctive contouring procedure in Table 2a and was not included in this combination table. Therefore, “no listed concomitant body-contouring procedure excluding liposuction” does not necessarily equal isolated abdominoplasty; the predefined isolated-abdominoplasty subgroup comprised 59 patients. Combinations of concomitant body-contouring procedures excluding liposuction. All patients underwent abdominoplasty. Liposuction was reported separately as an adjunctive contouring procedure in Table 2a and was not included in this combination table. Therefore, “no listed concomitant body-contouring procedure excluding liposuction” does not necessarily equal isolated abdominoplasty; the predefined isolated-abdominoplasty subgroup comprised 59 patients. Comparison between non-bariatric and post-bariatric patients Post-bariatric patients were older (48 ± 12 years [27–79] vs. 43 ± 12 years [21–75], p = 0.002), had longer weight loss duration (4.0 years [IQR 3.0–5.0] vs. 2.0 years [IQR 2.0–4.0], p < 0.001), higher maximum BMI (46.3 ± 6.5 kg/m2 [34.9–70.2] vs. 41.1 ± 6.2 kg/m² [25.8–66.9], p < 0.001), and greater delta BMI (19.6 ± 5.5 kg/m2 [8.9–36.8] vs. 15.0 ± 5.5 kg/m2 [5.3–40.9], p < 0.001). Pre-operative BMI was comparable (26.7 ± 4.1 kg/m2 [17.3–42.6] vs. 26.1 ± 3.1 kg/m2 [19.8–38.8], p = 0.456). Resection weight did not differ significantly between post-bariatric and non-bariatric patients (1,942 ± 1,288 g vs. 1,590 ± 1,000 g, p = 0.057) (Table 3). Table 3Comparison of characteristics between non-bariatric and post-bariatric patients.Table 3 dummy alt textNon-bariatric groupPost-bariatric groupp-valueTotal (N = 132)Total (N = 85)0.006Female91 (69%)73 (86%)Male41 (31%)12 (14%)Age (years)*43 ± 12 (21–75)48 ± 12 (27–79)0.002Weight loss (kg)*44 ± 17 (15–110)55 ± 16 (25–104)<0.001BMI max (kg/m2)*41.1 ± 6.2 (25.8–66.9)46.3 ± 6.5 (34.9–70.2)<0.001BMI delta (kg/m2)*15.0 ± 5.5 (5.3–40.9)19.6 ± 5.5 (8.9–36.8)<0.001BMI pre-op (kg/m2)*26.1 ± 3.1 (19.8–38.8)26.7 ± 4.1 (17.3–42.6)0.456Resection weight (g)*1590 ± 1000 (87–6980)1942 ± 1288 (323–6453)0.057Time period of weight loss (years)⁎⁎2.0 (IQR 2.0–4.0)4.0 (IQR 3.0–5.0)<0.001Smoking0.577 No59 (45%)42 (49%) Yes73 (55%)43 (51%)Metabolic disease0.001 No96 (73%)43 (51%) Yes36 (27%)42 (49%)Complication0.419 No83 (63%)58 (68%) Yes49 (37%)27 (32%)Anemia pre-op0.706 No112 (85%)70 (82%) Yes20 (15%)15 (18%)⁎Values are presented as mean ± SD, with range in parentheses⁎⁎Values are presented as median with interquartile range (IQR) in parentheses; BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative body mass index;BMI pre-op = pre-operative body mass index. Comparison of characteristics between non-bariatric and post-bariatric patients. Values are presented as mean ± SD, with range in parentheses Values are presented as median with interquartile range (IQR) in parentheses; BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative body mass index; BMI pre-op = pre-operative body mass index. Missing data Missingness was low across key variables. Weight-loss duration was available in 114/217 patients (53%), whereas operative time had <2% missing data (Table 4). Table 4Missingness across key study variables.Table 4 dummy alt textTotal (N =217)Valid n (%)Missing n (%)Weight loss duration (years)114 (53%)103 (47%)Operative time (min)214 (99%)3 (1%) Missingness across key study variables. Factors associated with resection weight Bivariate correlations with resection weight were strongest for pre-operative BMI (Spearman r = 0.620, p < 0.001), followed by pre-operative weight (r = 0.458, p < 0.001), maximum BMI (r = 0.320, p < 0.001), and age (r = 0.288, p < 0.001). Delta BMI (r = -0.03, p = 0.70) and duration of weight loss (r = 0.15, p = 0.12) were not associated (Table 5). Table 5Bivariate correlations between resection weight and selected anthropometric and demographic variables.Table 5 dummy alt textVariableCorrelation Coefficient r (Spearman)p (2-tailed)NAge (years)0.288⁎⁎<0.001217Sex0.137*0.044217Weight max (kg)0.238⁎⁎<0.001217Weight pre-op (kg)0.458⁎⁎<0.001217BMI max (kg/m2)0.320⁎⁎<0.001217BMI delta (kg/m2)-0.0260.704217BMI pre-op (kg/m2)0.620⁎⁎<0.001217Weight loss duration (years)0.1480.116114⁎⁎Correlation is significant at the 0.01 level (2-tailed);⁎Correlation is significant at the 0.05 level (2-tailed);BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative Body mass index; BMI pre-op = pre-operative body mass index.Because BMI delta is mathematically derived from BMI max and BMI pre-op, these variables were not entered simultaneously into the same multivariable regression model. Bivariate correlations between resection weight and selected anthropometric and demographic variables. Correlation is significant at the 0.01 level (2-tailed); Correlation is significant at the 0.05 level (2-tailed); BMI max = maximum body mass index; BMI delta = maximum body mass index–pre-operative Body mass index; BMI pre-op = pre-operative body mass index. Because BMI delta is mathematically derived from BMI max and BMI pre-op, these variables were not entered simultaneously into the same multivariable regression model. In the primary multivariable linear regression, pre-operative BMI showed the strongest adjusted association with resection weight (+209.9 g per kg/m2, 95% CI 179.9–239.9, p < 0.001). Female sex (+363.5 g, 95% CI 115.1–612.0, p = 0.004), metabolic disease (+346.0 g, 95% CI 101.0–591.1, p = 0.006), and concomitant belt lipectomy (+365.6 g, 95% CI 158.0–573.3, p < 0.001) also showed positive adjusted associations with resection weight. Age was not associated with resection weight after adjustment (p = 0.636). The model explained 56% of the variance (R2 = 0.557, adjusted R2 = 0.547, F (5,211) = 53.1, p < 0.001) (Table 6a). VIFs were ≤ 1.5. Diagnostics showed approximately normal residuals with minor upper-tail deviation, no major heteroscedasticity, and no influential observations (Cook's D ≤ 0.18). Log-transformation did not materially alter the results. Findings were comparable in the alternative weight-based model (Table 6b). The minimum detectable partial R² was approximately 3.6% for both linear models at 80% power. Table 6aPrimary multivariable linear regression model of factors associated with resection weight using BMIpre-op.Table 6a dummy alt textPredictorB (g)SEβ95% CIpConstant−4502.9420.0—−5330.8 to −3674.9<0.001Age (years)2.55.30.027−7.9 to 12.90.636Female sex363.5126.00.138115.1 to 612.00.004Metabolic disease346.0124.30.147101.0 to 591.10.006Belt lipectomy365.6105.40.162158.0 to 573.3<0.001BMI pre-op (kg/m²)209.915.20.656179.9 to 239.9<0.001Model fit: R² = 0.557, Adjusted R² = 0.547, F(5,211) = 53.10, p < 0.001, N = 217* No relevant multicollinearity was observed (all VIFs ≤ 1.50); BMI pre-op = pre-operative body mass index. Primary multivariable linear regression model of factors associated with resection weight using BMIpre-op. Model fit: R² = 0.557, Adjusted R² = 0.547, F(5,211) = 53.10, p < 0.001, N = 217 * No relevant multicollinearity was observed (all VIFs ≤ 1.50); BMI pre-op = pre-operative body mass index. Table 6bAlternative multivariable linear regression model of factors associated with resection weight using pre-operative weight.Table 6b dummy alt textPredictorB (g)SEβ95% CIpConstant−5055.6431.9—−5907.1 to −4204.2<0.001Age (years)10.95.00.1161.0 to 20.80.031Female sex1112.2137.00.423842.1 to 1382.2<0.001Metabolic disease363.1120.80.154125.1 to 601.20.003Belt lipectomy360.7102.40.159158.8 to 562.6<0.001Pre-operative weight (kg)68.14.70.73658.9 to 77.3<0.001Model fit: R² = 0.582, Adjusted R² = 0.572, F(5,211) = 58.66, p < 0.001, N = 217 Alternative multivariable linear regression model of factors associated with resection weight using pre-operative weight. Model fit: R² = 0.582, Adjusted R² = 0.572, F(5,211) = 58.66, p < 0.001, N = 217 Complication frequency and severity Post-operative complications occurred in 76 of 217 patients (35.0%): 37.1% in the non-bariatric and 31.8% in the post-bariatric group (p = 0.419). Minor complications (Clavien–Dindo grades I–II) occurred in 29 patients (13.4%) and major complications (grade ≥ IIIa) in 47 patients (21.7%); no grade IV or V events occurred. The distribution of no, minor, and major complications did not differ between non-bariatric and post-bariatric patients (χ²(2) = 0.67, p = 0.714) (Table 7a). Wound-specific complications occurred in 71 (32.7%) (Table 7b). Table 7aPost-operative complications according to Clavien–Dindo classification by weight loss procedure.Table 7a dummy alt textNon-bariatric group (N = 132)Post-bariatric group (N = 85)No complications*83 (63%)58 (68%)Complications49 (37%)27 (32%)Minor complications*19 (14%)10 (12%) Degree I16 (12%)9 (11%) Degree II3 (2%)1 (1%)Major complications*30 (23%)17 (20%) Degree IIIa14 (11%)7 (8%) Degree IIIb16 (12%)10 (12%) Degree IV0 (0%)0 (0%) Degree V0 (0%)0 (0%)⁎p = 0.714, Chi-square test comparing the overall distribution of no, minor (Clavien–Dindo I–II), and major (≥ IIIa) complications between non-bariatric and post-bariatric patients. Post-operative complications according to Clavien–Dindo classification by weight loss procedure. p = 0.714, Chi-square test comparing the overall distribution of no, minor (Clavien–Dindo I–II), and major (≥ IIIa) complications between non-bariatric and post-bariatric patients. Table 7bWound-specific complication types.Table 7b dummy alt textWound-specific complication categoryn% of total cohortAny wound-specific complication7132.7%Wound healing disorder / dehiscence3516.1%Seroma188.3%Hematoma / postoperative bleeding156.9%Skin or fat necrosis52.3%Wound infection41.8%Fistula / suture-related complications31.4%Medically indicated wound revision / debridement / necrosectomy73.2%Categories were not mutually exclusive; one patient could have more than one wound-specific complication type. Repeated occurrences of the same subtype within one patient, such as multiple seroma punctures, were counted once. Percentages refer to the total cohort of 217 patients. Wound-specific complication types. Categories were not mutually exclusive; one patient could have more than one wound-specific complication type. Repeated occurrences of the same subtype within one patient, such as multiple seroma punctures, were counted once. Percentages refer to the total cohort of 217 patients. Predictors of complications Screening identified age and pre-operative anemia as the strongest candidate associations; resection weight was retained as the predefined hypothesis variable despite no univariable association (Table 8a). In the adjusted composite endpoint model, younger age and pre-operative anemia were associated with post-operative complications (age: OR 0.969, 95% CI 0.944–0.994, p = 0.016; pre-operative anemia: OR 2.153, 95% CI 1.024–4.528, p = 0.043). Resection weight was not significantly associated with the composite endpoint (OR 1.015 per 100 g, 95% CI 0.990–1.042, p = 0.245) (Table 8b). Table 8bMultivariable logistic regression model for post-operative complications.Table 8b dummy alt textPredictorOR95% CIpAge0.9690.944 to 0.9940.016Pre-operative anemia*2.1531.024 to 4.5280.043Resection weight (per 100 g)1.0150.990 to 1.0420.245Based on 217 complete cases; pre-operative anemia was present in 35 patients (16.1%).⁎Reference category: no pre-operative anemia Multivariable logistic regression model for post-operative complications. Based on 217 complete cases; pre-operative anemia was present in 35 patients (16.1%). Reference category: no pre-operative anemia Resection weight was not associated with major complications (OR 1.014 per 100 g, 95% CI 0.976–1.054, p = 0.471) or wound-specific complications (OR 1.018 per 100 g, 95% CI 0.992–1.045, p = 0.179) (Table 8c, Table 8d). At 80% power, the MDE odds ratios were 1.040, 1.047, and 1.041 per 100 g for the composite, major, and wound-specific endpoints, respectively. The observed odds ratios were below these thresholds in all three models. Table 8cMultivariable logistic regression model for major post-operative complications.Table 8c dummy alt textPredictorOR95% CIpAge0.9770.949 to 1.0060.115BMI pre-op (kg/m2)1.0320.908 to 1.1730.626Resection weight (per 100 g)1.0140.976 to 1.0540.471Major complications were defined as Clavien–Dindo grade ≥ IIIa. Based on 217 complete cases and 47 major events.BMI pre-op = pre-operative body mass index. Multivariable logistic regression model for major post-operative complications. Major complications were defined as Clavien–Dindo grade ≥ IIIa. Based on 217 complete cases and 47 major events. BMI pre-op = pre-operative body mass index. Table 8dMultivariable logistic regression model for wound-specific complications.Table 8d dummy alt textPredictorOR95% CIpAge0.9720.947 to 0.9980.035Pre-operative anemia*2.1391.015 to 4.5060.046Resection weight (per 100 g)1.0180.992 to 1.0450.179Based on 217 complete cases; pre-operative anemia was present in 35 patients (16.1%).Wound-specific complications included wound healing disorder/dehiscence, seroma, hematoma/postoperative bleeding, skin or fat necrosis, wound infection, fistula/suture-related complications, and medically indicated wound revision/debridement.⁎Reference category: no pre-operative anemia. Multivariable logistic regression model for wound-specific complications. Based on 217 complete cases; pre-operative anemia was present in 35 patients (16.1%). Wound-specific complications included wound healing disorder/dehiscence, seroma, hematoma/postoperative bleeding, skin or fat necrosis, wound infection, fistula/suture-related complications, and medically indicated wound revision/debridement. Reference category: no pre-operative anemia. Sensitivity analyses Restricting analyses to isolated abdominoplasty (n = 59) confirmed the pre-operative BMI (β = 177.7 g per 1 kg/m², standardized β = 0.72, p < 0.001) and metabolic disease (β = 549.0 g, standardized β = 0.26, p = 0.011) associations, whereas the female sex effect was attenuated (β = 30.9 g, standardized β = 0.01, p = 0.887). No association between resection weight and complications emerged in this subgroup (OR = 0.975 per 100 g, Nagelkerke R2 = 0.018, p = 0.409) (Table 9). Table 9Sensitivity analyses comparing the primary study findings with the subgroup restricted to isolated abdominoplasty procedures (n=59).Table 9 dummy alt textOutcomeMain analysisIsolated ABPInterpretationBMI pre-op → resection weightβ = 209.9 g per 1 kg/m²; p < 0.001β = 177.7 g per 1 kg/m²; standardized β = 0.72; p < 0.001robustFemale sex → resection weightβ = +363.5 g; p = 0.004β = +30.9 g; standardized β = 0.01; p = 0.887attenuatedMetabolic disease → resection weightβ = +346.0 g; p = 0.006β = +549.0 g; standardized β = 0.26; p = 0.011robustResection weight → complicationOR = 1.015 per 100 g; p = 0.245OR = 0.975 per 100 g; p = 0.409; Nagelkerke R2 = 0.018No evidence of associationBMI pre-op = pre-operative body mass index. Sensitivity analyses comparing the primary study findings with the subgroup restricted to isolated abdominoplasty procedures (n=59). BMI pre-op = pre-operative body mass index. Discussion In this exploratory cohort, pre-operative BMI, female sex, metabolic disease, and concomitant belt lipectomy were associated with resection weight after adjustment (R2 = 0.56), whereas resection weight was not significantly associated with composite, major, or wound-specific complications. Given the retrospective design and limited number of severity-stratified events, these findings should be interpreted as exploratory associations rather than confirmatory prediction results. The strong association of pre-operative BMI indicates that the patient’s pre-operative anthropometric profile was more closely related to resection weight than historical weight trajectory.22,23 BMI delta and weight-loss duration were not associated with resection weight. However, BMI delta does not capture skin quality, tissue laxity, elasticity, or adipose distribution. No mechanistic conclusion can be drawn regarding residual adiposity versus skin properties. Female sex was associated with higher resection weight in the primary model, consistent with reported differences in fat distribution, skin structure, and soft-tissue biomechanics.24,25 This association attenuated in the isolated-abdominoplasty subgroup, suggesting that possible procedural selection may partly explain the finding. Metabolic disease may reflect obesity-related tissue burden, while the association with belt lipectomy indicates that resection weight also depends on surgical strategy. No significant association was observed across all three complication endpoints. This contrasts with prior studies linking higher resection weight or operative burden, particularly resection weights ≥ 2 kg, to increased complication rates.16,17 Resection weight may be an imperfect proxy for operative burden in a cohort with frequent concomitant procedures. In addition, institutional BMI thresholds and pre-operative optimization within the Austrian statutory insurance framework may have reduced the representation of patients at particularly high operative risk. Accordingly, smaller clinically relevant effects cannot be excluded. Contrary to some prior reports,15 overall and severity-stratified complication rates did not differ significantly between post-bariatric and non-bariatric patients. This may reflect selection within the Austrian statutory insurance framework, including requirements for weight stability and pre-operative optimization. The primary model yielded the following multivariable equation: Female sex, metabolic disease, and belt lipectomy are binary variables (1 = yes, 0 = no). For example, a female patient with a pre-operative BMI of 30 kg/m², metabolic disease, and planned belt lipectomy would have an estimated resection weight of approximately 2.9 kg. This point estimate should not be interpreted as an individual prediction interval. The equation requires external validation and recalibration. Given the substantial residual variance and absence of external validation, it should be considered exploratory rather than a validated clinical prediction tool. Clinical implications After external validation, the equation may support pre-operative counselling, surgical planning, and institutional benchmarking by providing an estimate of expected resection weight. Such estimates could assist with case preparation and communication of the anticipated extent of resection. Future prospective studies should evaluate additional factors, including nutritional markers,26 body composition, smoking exposure,27 duration of weight stability, physical activity, and patient-reported symptom burden. Strengths and limitations Strengths include the relatively large consecutive single-center cohort comprising both post-bariatric and non-bariatric patients. Additional strengths include clinically motivated multivariable modelling with diagnostics and sensitivity analyses, and standardized Clavien–Dindo grading with independent blinded review. Limitations include the retrospective design and associated documentation bias, particularly for minor complications managed outside the center. Despite high inter-rater agreement, retrospective Clavien–Dindo grading remains subject to documentation-dependent misclassification. Median follow-up of 3 months limited the detection of late complications. The complication analyses were exploratory, and MDE analyses indicated that odds ratios below approximately 1.04–1.05 per 100 g could not be excluded, particularly for severity-specific endpoints. The wound-specific endpoint was more clinically homogeneous than the overall composite endpoint but remained mechanistically heterogeneous. The study was not powered for subtype-specific analyses. Nutritional markers were inconsistently recorded. The reported diabetes prevalence (n = 17, 8%) is likely an underestimate attributable to underdocumentation.28 Weight-loss duration showed substantial missingness and was excluded from multivariable models. Combined procedures were frequent (n = 158, 73%), limiting transferability to isolated aesthetic abdominoplasty. Belt lipectomy was included in the primary model, and isolated-abdominoplasty sensitivity analyses showed broadly consistent patterns. This small subgroup (n = 59) limited the precision of complication estimates. Less frequent concomitant procedures were not individually modelled. Residual tumescent fluid may have slightly inflated measured resection weight. Restriction to patients covered by Austrian statutory insurance further limits generalizability to reconstructive post-MWL populations. Conclusion In this exploratory cohort of 217 post-MWL abdominoplasty patients, pre-operative BMI was most strongly associated with resection weight, with female sex, metabolic disease, and concomitant belt lipectomy providing additional explanatory value. BMI delta and weight-loss duration were not associated with resection weight, suggesting that current anthropometric status may be more informative than weight-loss history. Resection weight was not significantly associated with composite, major, or wound-specific complications; however, MDE analyses indicate that smaller effects cannot be excluded. The proposed equation may serve as a basis for further evaluation of resection-weight estimation. External validation is required before it can be considered for clinical use in surgical planning or patient counselling. Prospective multicenter studies are needed to confirm the resection-weight model and complication-risk findings. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used ChatGPT and Claude solely to improve the language and readability of the manuscript. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Ethical approval Ethical approval was granted by the Ethics Board of the Medical University of Graz (EK-Nr.: 36-118 ex 23/24). The study was conducted in accordance with the Declaration of Helsinki. Declaration of competing interest None. Acknowledgements This manuscript forms part of the diploma thesis of Sabrina Luettschwager at the Medical University of Graz. References