The impact of morbid obesity and metabolic syndrome in elective spine surgery outcomes: A systematic review Authors: Vinicius Ricieri Ferraz, Guilherme Santos Piedade, Bashar Dawoud, Joacir Graciolli Cordeiro, Carlos R. Goulart, Philippe A. Mercier, Tobias A. Mattei Published: 2026 ABSTRACT BackgroundMorbid obesity is increasingly common among patients undergoing elective spine surgery and is associated with substantial technical and physiological risk. By 2030, nearly 1 in 2 US adults is projected to have obesity, and severe obesity is expected to affect almost 1 in 4 adults. This systematic review synthesizes the available evidence on morbid obesity (BMI ≥ 40 kg per m2) and metabolic syndrome in elective spine surgery and summarizes complication patterns, resource utilization, and patient-reported outcomes (PROs), with attention to whether surgical approach potentially modifies risk.MethodsMEDLINE, EMBASE, Scopus, Web of Science, and the Cochrane Library were searched from January 1, 2000 through September 30, 2025 for comparative adult studies of elective spine surgery in morbidly obese versus non-obese patients. Two reviewers screened studies and extracted data. Risk of bias was assessed with ROBINS I, and reporting followed PRISMA guidance. Because outcome definitions, procedures, and data sources were heterogeneous, we primarily performed a structured qualitative synthesis and, where available, reported pooled estimates from the most comprehensive available meta-analyses as quantitative benchmarks.ResultsThirty-six studies were eligible. In the largest published meta-analysis from Jiang et al., 2014, morbid obesity was associated with higher odds of surgical site infection (Odds ratio [OR 2.33]), venous thromboembolism (OR 3.15), revision surgery (OR 1.43), and mortality (OR 2.6). A contemporary anchor-validation pool of 2 post-2014 cohorts with extractable surgical-site-infection data yielded a concordant estimate (pooled OR 3.03, 95% Confidence Interval [95% CI 2.20–4.17]; I² = 0%). Metabolic syndrome remained an independent predictor of wound complications (OR 1.47) and readmissions (OR 1.48) after adjustment for obesity. Although obese patients improve after surgery, the likelihood of achieving minimal clinically important difference (MCID) threshold in patient reported outcomes is often reduced. Evidence on preoperative bariatric surgery was mixed, and emerging pharmacologic optimization with GLP-1 receptor agonists warrants further study.ConclusionsMorbid obesity and metabolic syndrome are strong predictors of adverse perioperative outcomes and increased resource utilization after elective spine surgery. A value-based approach should incorporate metabolic health into risk stratification and emphasize scalable, equitable optimization pathways. Introduction The prevalence of morbid obesity is rising alongside an aging population and increasing rates of degenerative spinal disease. In the United States, projections suggest that by 2030 nearly 1 in 2 adults will have obesity and severe obesity will be widespread [1,2]. Higher body mass is linked to lumbar stenosis and degenerative disc disease, creating a cycle in which pain and limited mobility promote sedentary behavior and further weight gain [3]. Spine surgeons are therefore increasingly asked to treat patients with World Health Organization (WHO) Class III obesity (BMI ≥ 40 kg per m2) and high-risk WHO Class II obesity (BMI ≥ 35 kg per m2) with a major comorbidity burden [4]. Morbid obesity may lead to increased risks of complications after spine surgery through systemic and local mechanisms. Chronic inflammation, insulin resistance, limited cardiopulmonary reserve, and altered pharmacokinetics increase perioperative vulnerability, while body habitus can limit imaging quality, restrict exposure, and complicate positioning and instrumentation, often increasing operative time and blood loss [5,6]. Yet BMI alone is an imperfect proxy for operative difficulty because tissue distribution varies and correlates only modestly with measures such as posterior cervical subcutaneous fat thickness. Metabolic health and procedure specific factors may therefore be as important as BMI thresholds when counseling patients and selecting a surgical strategy [7]. Although obesity per se has been shown to be associated with higher surgical site infection, venous thromboembolism, longer length of stay, and higher costs, there are still few sufficiently actionable strategies to mitigate them [3]. Key questions remain regarding which procedures and approaches offer the best risk-benefit balance in morbidly obese patients, whether minimally invasive techniques mitigate complications compared with open surgery, and which optimization pathways meaningfully improve outcomes. This systematic review synthesizes comparative evidence on elective spine surgery outcomes in morbidly obese versus non-obese adults, focusing on complications, resource utilization, patient-reported outcomes, and the impact of surgical approach upon complication rates. Material and method Study design This systematic review was designed and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) [8]. The review protocol, outlining the PICOS framework, search strategy, and analytical plan, was established a priori to ensure methodological rigor and minimize bias. Search strategy A comprehensive literature search was executed across 5 major electronic databases: MEDLINE (via PubMed), EMBASE (Excerpta Medica dataBASE), Scopus, Web of Science, and the Cochrane Central Register of Controlled Trials (CENTRAL). The search covered the period from January 1, 2000 through September 30, 2025, with the final database query executed on the latter date. The search strategy combined Medical Subject Headings (MeSH) and free text keywords, employing a 3-concept logical AND operator scheme. The search string was structured as follows:(``morbid obesity'' OR ``severe obesity'' OR ``class III obesity'' OR ``BMI ≥ 40” OR bariatric)AND(``spine surgery'' OR ``spinal fusion'' OR ``laminectomy'' OR ``discectomy'' OR ``arthrodesis'' OR ``TLIF'' OR ``ACDF'' OR ``spinal deformity'')AND(``complications'' OR ``outcomes'' OR ``SSI'' OR ``pseudarthrosis'' OR ``readmission'' OR ``cost'' OR ``minimally invasive surgery''). (``morbid obesity'' OR ``severe obesity'' OR ``class III obesity'' OR ``BMI ≥ 40” OR bariatric) AND (``spine surgery'' OR ``spinal fusion'' OR ``laminectomy'' OR ``discectomy'' OR ``arthrodesis'' OR ``TLIF'' OR ``ACDF'' OR ``spinal deformity'') AND (``complications'' OR ``outcomes'' OR ``SSI'' OR ``pseudarthrosis'' OR ``readmission'' OR ``cost'' OR ``minimally invasive surgery''). Eligibility criteria Studies were included if they met the following criteria, based on the Population, Intervention(s), Comparator(s), Outcomes, and Study Types (PICOS) framework: Population: The review included studies of adult patients (age ≥ 18 years) with morbid obesity, defined as a BMI ≥ 40 kg/m2, undergoing elective surgery of the cervical, thoracic, or lumbar spine. Studies that stratified outcomes by obesity class, allowing for specific analysis of the Class III group, were prioritized. Intervention(s): Any elective spinal procedure for decompression, fusion, or deformity correction was considered. This included, but was not limited to, anterior cervical discectomy and fusion (ACDF), cervical disc arthroplasty, laminectomy, transforaminal lumbar interbody fusion (TLIF), and posterior spinal fusion (PSF). Surgical approaches were categorized as open, minimally invasive (MIS), or hybrid. Studies evaluating specific perioperative co-interventions, such as preoperative weight loss programs or enhanced nutritional support, were included for narrative synthesis. Comparator(s): The primary comparator group consisted of non-obese patients (defined as BMI < 30 kg/m2) undergoing equivalent surgical procedures. A secondary comparator group included patients with Class I to II obesity (BMI 30–39.9 kg/m2). Studies without a comparator group (for example, single-arm case series) were considered for descriptive purposes in the context of the discussion of complication rates but were not included in the comparative synthesis. Outcomes Primary Outcomes: (1) Major postoperative complications occurring within 90 days of the index procedure, including surgical site infection (SSI, both deep and superficial), venous thromboembolism (VTE, including deep vein thrombosis and pulmonary embolism [PE]), new or worsening neurological deficit, implant-related complications (eg, failure, malposition), pseudarthrosis, and all-cause mortality. (2) Patient Reported Functional Outcomes (PROs) at a minimum of 1-year follow-up, including the Oswestry Disability Index (ODI) for lumbar pathology, the Neck Disability Index (NDI) for cervical pathology, and the Visual Analog Scale (VAS) for back and/or leg pain. Secondary Outcomes: (1) Healthcare resource utilization metrics, including total operating room time, estimated blood loss (EBL), length of hospital stay (LOS), and direct hospital costs. (2) All cause readmission and reoperation rates at 1 year follow-up. Study Types: Randomized controlled trials (RCTs), prospective comparative cohort studies, and retrospective comparative cohort studies were eligible for inclusion. Large volume case series (n > 50) from national or institutional databases were also considered for the narrative synthesis. Only studies published between January 1, 2000 and September 30, 2025 were included to reflect contemporary surgical techniques and perioperative care. Selection process The study selection was conducted in a 2-stage process by 2 independent reviewers. In the first stage, all titles and abstracts retrieved from the search were screened for potential relevance. Studies that clearly did not meet the PICOS eligibility criteria were excluded. In the second stage, the full texts of all potentially relevant articles were retrieved and assessed for final inclusion. Any disagreements between the 2 reviewers at either stage were resolved through discussion and consensus, with adjudication by a third, senior reviewer if necessary. Data collection A standardized data extraction form, developed in Microsoft Excel, was used to systematically collect information from each included study. Two reviewers independently extracted the data, with discrepancies resolved by consensus. The following data points were extracted: Study Identifiers: Author(s), year of publication, and country of origin. Study Characteristics: Study design, study period, and mean follow-up duration. Patient Characteristics: Number of patients in the morbidly obese group and comparator group(s), mean age, sex distribution, specific BMI definitions and mean values, and key comorbidities (eg, diabetes, hypertension). Surgical Details: Type of spinal procedure(s), anatomical region (cervical, thoracic, lumbar), number of levels operated, and surgical approach (open, MIS, other). Outcome Data: For dichotomous outcomes (eg, complications), raw numbers of events and total patients in each group were extracted to construct 2 × 2 contingency tables. For continuous outcomes (eg, OR time, PRO scores), means and standard deviations (or equivalent measures of central tendency and dispersion) were extracted for each group. Equity and Economic Data: Where reported, data on patient race/ethnicity, insurance status, hospital type (eg, academic vs. community), and direct/indirect costs were extracted for the narrative synthesis. Risk of bias assessment The methodological quality and risk of bias of each included study were independently assessed by 2 reviewers. For any RCTs identified, the Cochrane Risk of Bias 2 (RoB 2) tool was used. For all non-randomized studies of interventions (NRSIs), which constituted the majority of the studies, the Risk of Bias in Nonrandomized Studies of Interventions (ROBINS I) tool was employed. The ROBINS I tool assesses bias across 7 domains: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of the reported result. A final judgment of low, moderate, serious, or critical risk of bias was assigned to each study. Particular attention was given to confounding domains highly relevant to this topic, such as the potential for confounding by indication and the influence of unmeasured variables like surgeon experience and institutional case volume. Effect measures and synthesis methods A mixed methods approach to synthesis was utilized. A structured narrative synthesis summarized findings across spinal region, procedure type, and outcome domain, with separate synthesis for health economics, access, and disparities. Because outcome definitions, procedures, and data sources were heterogeneous and several analyses relied on overlapping administrative databases, we did not perform a de novo quantitative pooling across all included studies [9]. Published pooled estimates from the most comprehensive prior meta-analyses are reported throughout as historical benchmarks, with their source explicitly attributed. As an exception to this approach, we conducted a contemporary anchor-validation pool restricted to post-2014 cohorts judged at moderate ROBINS I risk that reported extractable 2 × 2 surgical site infection data by BMI category. Two studies contributed to this pool: de la Garza-Ramos et al. [10] and Buerba et al. ([11], Class III subset). Crude odds ratios were combined using the DerSimonian–Laird random-effects estimator with the generic inverse-variance method on the log-odds ratio scale. Cochran’s Q, I², and between-study variance (τ²) were computed for heterogeneity assessment. Where available, we extracted adjusted effect estimates and highlighted the most clinically informative analyses. For quantitative context, we reported pooled estimates from the most comprehensive published meta-analyses and compared these benchmarks with findings from newer large cohort studies in our study set [6,[12], [13], [14]]. Effect measures are reported as odds ratios for dichotomous outcomes and as mean differences for continuous outcomes when provided by the source studies or meta-analyses. Planned subgroup considerations included surgical approach (minimally invasive versus open) and spinal region (cervical versus lumbar), but in most domains the evidence was not sufficiently homogeneous to support consistent pooled subgroup estimates. Risk of bias considerations focused on confounding, selection, and outcome measurement given the predominance of retrospective designs. Publication bias was not formally assessed because the review emphasized a heterogeneity-aware synthesis, and relied on published pooled benchmarks where relevant. Certainty of evidence Certainty of evidence was assessed using the GRADE framework for the 4 primary outcomes with available pooled estimates: surgical site infection, venous thromboembolism, mortality, and revision surgery. Each outcome was rated across the 5 GRADE domains (risk of bias, inconsistency, indirectness, imprecision, and publication bias). Because the body of evidence consisted almost entirely of non-randomized comparative studies, all outcomes started at low certainty per the GRADE approach to observational evidence, with further downgrading or upgrading applied as indicated by the domain assessments. Results Study selection and characteristics The comprehensive literature search yielded a total of 3,482 records. After removing duplicates, 2,157 unique titles and abstracts were screened. After title and abstract screening, 2,026 records were excluded, leaving 131 reports assessed for eligibility at the full-text level. Following the full-text assessment, 95 articles were excluded for the following reasons: no segmentation between BMI groups (n = 61); outcomes of interest not reported (n = 15); no access to the full text (n = 12); and conference proceedings (n = 7). This process resulted in the inclusion of 36 unique studies in the final qualitative and quantitative synthesis. The PRISMA 2020 flow diagram detailing this selection process is presented in Fig. 1.Fig. 1PRISMA flowchart diagram detailing this selection process.Fig 1 dummy alt text PRISMA flowchart diagram detailing this selection process. The 36 included studies were published between 2007 and 2025. The mean follow-up period varied from in-hospital outcomes only to 2 years. The risk of bias, as assessed by the ROBINS I tool, was judged to be serious for 7 studies [[14], [15], [16], [17], [18], [19], [20]] and moderate for the remaining 29 studies, primarily due to potential for confounding and bias in the selection of participants inherent in retrospective designs. The characteristics of all 36 included studies are detailed in Table 1.Table 1Summary of data for included studies.Table 1 dummy alt textStudyStudy designPopulation (Obese/Comparator)BMI cutoffs (kg/m²)Surgical procedure(s)Mean follow-upROBINS-IReports BMI ≥ 40 stratumPatel et al. (2007) [21]Retrospective cohort39/50≥30 vs. <30 (stratified)Thoracic/Lumbar fusionIn-hospitalModerateNoAdogwa et al. (2015) [22]Prospective cohort40/108≥30 (obese/morbidly obese)MIS-TLIF vs. Open TLIF2 yearsModerateNoBuerba et al. (2014) [11]Retrospective database716/9,671≥40 vs. <30Lumbar surgery (Fusion, Decomp)30 daysModerateYesKalanithi et al. (2012) [14]Retrospective database1,455/83,152Morbid obesity vs. NormalCervical/Lumbar fusionIn-hospitalSeriousNode la Garza-Ramos et al. (2015) [10]Retrospective cohort70/662>30 vs. <30Lumbar fusionLong-termModerateNoKrüger et al. (2019) [23]Retrospective cohort14/14≥40 vs. 20–25MIS-TLIF>1 yearModerateYesBerry et al. (2023) [24]Retrospective database582/5,927>30 vs. <30Cervical disc arthroplasty (CDA)In-hospitalModerateNoClaus et al (2021) [25]Retrospective database58/234≥40 vs. <30MIS-TLIF1 yearModerateYesJain et al. (2018) [17]Retrospective database5,791/150,136>40 vs. <25Posterior lumbar fusionIn-hospitalSeriousYesRavindra et al. (2015) [26]Cross-sectional16/218>40 vs. other BMI classesElective spine fusionPreoperativeModerateYesCha et al (2022). [27]Retrospective database12/162≥40 vs. other BMI classesMIS-TLIF2 yearsModerateYesChotai et al. (2016) [28]Prospective cohort18/114≥35 vs. <35 (stratified ≥ 40)ACDF2 yearsModerateYesBrown et al. (2020) [19]Retrospective cohort32/322≥40 vs. other BMI classesAdult spinal deformity1 yearSeriousYesSingh et al. (2025) [20]Retrospective database1,821/10,633>40 vs. other BMI classesCDA vs. ACDF2 yearsSeriousYesGrotle et al. (2019) [29]Longitudinal observational57,081 patientsNot specifiedLumbar surgery (Simple/Complex)15 yearsModerateNoTabatabai et al. (2021) [30]Retrospective cohort461/801≥30 vs. <30 (stratified)Spine, Knee, hip surgeryIn-hospitalModerateNoKhan JM et al. (2020) [31]Retrospective cohort279/290≥30 vs. <30Posterior lumbar fusion23 monthsModerateNoBono et al. (2018) [15]Retrospective database14,252/17,511≥30 vs. <30 (stratified)Lumbar surgery30 daysSeriousNoMarquez-Lara et al. (2014) [16]Retrospective database19,195/5,001≥25 vs. <25 (stratified)Lumbar surgery30 daysSeriousNoKatsevman et al. (2022) [32]Retrospective cohort63/60≥50 vs. <50Spine surgery (All)30 daysModerateYes (≥50)Lange et al. (2022) [33]Prospective observational526 patientsNot specifiedSpine surgery (All)1 yearModerateNoKhalooeifard R et al. (2022) [34]Prospective observational80/80≥30 vs. <30Posterior lumbar fusion>1 yearModerateNoHammer A et al. (2025) [35]Prospective cohort227 patientsContinuous + ≥35 stratumPosterior thoracolumbar fusion>1 yearModerateYes (continuous)Lovecchio et al. (2018) [36]Retrospective cohort1,903/16,702≥30 with MetS vs. No MetSPosterior lumbar fusion30 daysModerateNoThomas et al. (2023) [18]Retrospective database5,893/49,087≥30 with MetS vs. No MetSTLIF30 daysSeriousNoManson NA et al. (2016) [37]Retrospective cohort313/187≥25 vs. <25Thoracolumbar surgery>1 yearModerateNoSiccoli et al. (2021) [38]Prospective registry154 patientsBMI changeLumbar surgery1–2 yearsModerateNoHe X et al. (2020) [39]Retrospective cohort360/2,520≥30 with MetS vs. No MetSPLIF30 daysModerateNoNakajima et al. (2023) [40]Prospective cohort401/5,027≥30 vs. 18.5–30Lumbar surgery1 yearModerateNoHareni N et al. (2022) [41]National registry3,377/4,069≥30 vs. <25Decompression for CLSS1 yearModerateNoLeRoy TE et al. (2021) [42]Retrospective review131/53≥30 vs. <25Posterior lumbar interbody fusionIn-hospitalModerateNoCouch et al. (2023) [43]Retrospective cohort25/59Prior Bariatric vs. ControlInstrumented lumbar fusion2.4 yearsModerateNoMulvaney G et al. (2021) [44]Prospective registry4,349/3,952≥30 vs. <30Lumbar fusion1 yearModerateNoJoseph JR et al. (2019) [45]Retrospective review56 patients≥40 onlyThoracolumbar fusion1 yearModerateYesShahbandi A et al. (2025) [46]Retrospective database2,261 patientsStratified by BMI and MetSCervical laminoplasty30 daysModeratePartialQuraishi DA et al. (2025) [47]Retrospective cohort279 patientsStratified by BMIOpen TLIFIntraoperativeModeratePartialROBINS-I, Risk Of Bias In Nonrandomized Studies of Interventions; TLIF, transforaminal lumbar interbody fusion; PLIF, posterior lumbar interbody fusion; ACDF, anterior cervical discectomy and fusion; CDA, cervical disc arthroplasty; MIS, minimally invasive surgery; MetS, metabolic syndrome; CLSS, central lumbar spinal stenosis. Table 1Summary of data for included studies.Table 1 dummy alt textStudyStudy designPopulation (Obese/Comparator)BMI cutoffs (kg/m²)Surgical procedure(s)Mean follow-upROBINS-IReports BMI ≥ 40 stratumPatel et al. (2007) [21]Retrospective cohort39/50≥30 vs. <30 (stratified)Thoracic/Lumbar fusionIn-hospitalModerateNoAdogwa et al. (2015) [22]Prospective cohort40/108≥30 (obese/morbidly obese)MIS-TLIF vs. Open TLIF2 yearsModerateNoBuerba et al. (2014) [11]Retrospective database716/9,671≥40 vs. <30Lumbar surgery (Fusion, Decomp)30 daysModerateYesKalanithi et al. (2012) [14]Retrospective database1,455/83,152Morbid obesity vs. NormalCervical/Lumbar fusionIn-hospitalSeriousNode la Garza-Ramos et al. (2015) [10]Retrospective cohort70/662>30 vs. <30Lumbar fusionLong-termModerateNoKrüger et al. (2019) [23]Retrospective cohort14/14≥40 vs. 20–25MIS-TLIF>1 yearModerateYesBerry et al. (2023) [24]Retrospective database582/5,927>30 vs. <30Cervical disc arthroplasty (CDA)In-hospitalModerateNoClaus et al (2021) [25]Retrospective database58/234≥40 vs. <30MIS-TLIF1 yearModerateYesJain et al. (2018) [17]Retrospective database5,791/150,136>40 vs. <25Posterior lumbar fusionIn-hospitalSeriousYesRavindra et al. (2015) [26]Cross-sectional16/218>40 vs. other BMI classesElective spine fusionPreoperativeModerateYesCha et al (2022). [27]Retrospective database12/162≥40 vs. other BMI classesMIS-TLIF2 yearsModerateYesChotai et al. (2016) [28]Prospective cohort18/114≥35 vs. <35 (stratified ≥ 40)ACDF2 yearsModerateYesBrown et al. (2020) [19]Retrospective cohort32/322≥40 vs. other BMI classesAdult spinal deformity1 yearSeriousYesSingh et al. (2025) [20]Retrospective database1,821/10,633>40 vs. other BMI classesCDA vs. ACDF2 yearsSeriousYesGrotle et al. (2019) [29]Longitudinal observational57,081 patientsNot specifiedLumbar surgery (Simple/Complex)15 yearsModerateNoTabatabai et al. (2021) [30]Retrospective cohort461/801≥30 vs. <30 (stratified)Spine, Knee, hip surgeryIn-hospitalModerateNoKhan JM et al. (2020) [31]Retrospective cohort279/290≥30 vs. <30Posterior lumbar fusion23 monthsModerateNoBono et al. (2018) [15]Retrospective database14,252/17,511≥30 vs. <30 (stratified)Lumbar surgery30 daysSeriousNoMarquez-Lara et al. (2014) [16]Retrospective database19,195/5,001≥25 vs. <25 (stratified)Lumbar surgery30 daysSeriousNoKatsevman et al. (2022) [32]Retrospective cohort63/60≥50 vs. <50Spine surgery (All)30 daysModerateYes (≥50)Lange et al. (2022) [33]Prospective observational526 patientsNot specifiedSpine surgery (All)1 yearModerateNoKhalooeifard R et al. (2022) [34]Prospective observational80/80≥30 vs. <30Posterior lumbar fusion>1 yearModerateNoHammer A et al. (2025) [35]Prospective cohort227 patientsContinuous + ≥35 stratumPosterior thoracolumbar fusion>1 yearModerateYes (continuous)Lovecchio et al. (2018) [36]Retrospective cohort1,903/16,702≥30 with MetS vs. No MetSPosterior lumbar fusion30 daysModerateNoThomas et al. (2023) [18]Retrospective database5,893/49,087≥30 with MetS vs. No MetSTLIF30 daysSeriousNoManson NA et al. (2016) [37]Retrospective cohort313/187≥25 vs. <25Thoracolumbar surgery>1 yearModerateNoSiccoli et al. (2021) [38]Prospective registry154 patientsBMI changeLumbar surgery1–2 yearsModerateNoHe X et al. (2020) [39]Retrospective cohort360/2,520≥30 with MetS vs. No MetSPLIF30 daysModerateNoNakajima et al. (2023) [40]Prospective cohort401/5,027≥30 vs. 18.5–30Lumbar surgery1 yearModerateNoHareni N et al. (2022) [41]National registry3,377/4,069≥30 vs. <25Decompression for CLSS1 yearModerateNoLeRoy TE et al. (2021) [42]Retrospective review131/53≥30 vs. <25Posterior lumbar interbody fusionIn-hospitalModerateNoCouch et al. (2023) [43]Retrospective cohort25/59Prior Bariatric vs. ControlInstrumented lumbar fusion2.4 yearsModerateNoMulvaney G et al. (2021) [44]Prospective registry4,349/3,952≥30 vs. <30Lumbar fusion1 yearModerateNoJoseph JR et al. (2019) [45]Retrospective review56 patients≥40 onlyThoracolumbar fusion1 yearModerateYesShahbandi A et al. (2025) [46]Retrospective database2,261 patientsStratified by BMI and MetSCervical laminoplasty30 daysModeratePartialQuraishi DA et al. (2025) [47]Retrospective cohort279 patientsStratified by BMIOpen TLIFIntraoperativeModeratePartialROBINS-I, Risk Of Bias In Nonrandomized Studies of Interventions; TLIF, transforaminal lumbar interbody fusion; PLIF, posterior lumbar interbody fusion; ACDF, anterior cervical discectomy and fusion; CDA, cervical disc arthroplasty; MIS, minimally invasive surgery; MetS, metabolic syndrome; CLSS, central lumbar spinal stenosis. Summary of data for included studies. ROBINS-I, Risk Of Bias In Nonrandomized Studies of Interventions; TLIF, transforaminal lumbar interbody fusion; PLIF, posterior lumbar interbody fusion; ACDF, anterior cervical discectomy and fusion; CDA, cervical disc arthroplasty; MIS, minimally invasive surgery; MetS, metabolic syndrome; CLSS, central lumbar spinal stenosis. Synthesis of perioperative complications Across included comparative studies, morbid obesity was consistently associated with higher rates of postoperative complications. For a quantitative benchmark, pooled estimates from the landmark 2014 meta-analysis by Jiang et al. are summarized in Table 2 [2].Table 2Pooled estimates of postoperative complications in morbidly obese versus nonobese patients undergoing elective spine surgery CI, confidence interval; OR, odds ratio.Table 2 dummy alt textOutcomeSourceNo. studiesPooled OR (95% CI)p valueI²GRADE certaintySurgical Site InfectionJiang et al. [2]242.33 (1.94, 2.79)<.00149%LowSurgical Site Infection (anchor-validation)This review23.03 (2.20, 4.17)<.0010%LowVenous ThromboembolismJiang et al. [2]63.15 (1.92, 5.17)<.0010%Very LowMortalityJiang et al. [2]62.60 (1.50, 4.49)<.0010%Very LowRevision SurgeryJiang et al. [2]91.43 (1.05, 1.93).0228%Very LowPooled estimates derived from the meta-analysis of Jiang et al. [2] are presented as historical benchmarks. The anchor-validation pool (gray-shaded row) combines de la Garza-Ramos et al. [10] and Buerba et al. (Class III subset) [11] using DerSimonian–Laird random-effects estimation on log-odds ratios (Cochran's Q = 0.33, p = .57; τ² = 0). Table 2Pooled estimates of postoperative complications in morbidly obese versus nonobese patients undergoing elective spine surgery CI, confidence interval; OR, odds ratio.Table 2 dummy alt textOutcomeSourceNo. studiesPooled OR (95% CI)p valueI²GRADE certaintySurgical Site InfectionJiang et al. [2]242.33 (1.94, 2.79)<.00149%LowSurgical Site Infection (anchor-validation)This review23.03 (2.20, 4.17)<.0010%LowVenous ThromboembolismJiang et al. [2]63.15 (1.92, 5.17)<.0010%Very LowMortalityJiang et al. [2]62.60 (1.50, 4.49)<.0010%Very LowRevision SurgeryJiang et al. [2]91.43 (1.05, 1.93).0228%Very LowPooled estimates derived from the meta-analysis of Jiang et al. [2] are presented as historical benchmarks. The anchor-validation pool (gray-shaded row) combines de la Garza-Ramos et al. [10] and Buerba et al. (Class III subset) [11] using DerSimonian–Laird random-effects estimation on log-odds ratios (Cochran's Q = 0.33, p = .57; τ² = 0). Pooled estimates of postoperative complications in morbidly obese versus nonobese patients undergoing elective spine surgery CI, confidence interval; OR, odds ratio. Pooled estimates derived from the meta-analysis of Jiang et al. [2] are presented as historical benchmarks. The anchor-validation pool (gray-shaded row) combines de la Garza-Ramos et al. [10] and Buerba et al. (Class III subset) [11] using DerSimonian–Laird random-effects estimation on log-odds ratios (Cochran's Q = 0.33, p = .57; τ² = 0). To assess whether the Jiang et al. [2] surgical site infection estimate has been corroborated, contradicted, or attenuated by subsequent evidence, we performed a contemporary anchor-validation pool restricted to post-2014 cohorts judged at moderate ROBINS I risk that reported extractable 2 × 2 surgical site infection data by BMI category. Two studies contributed: de la Garza-Ramos et al. [10] (n = 732; obese vs. nonobese SSI 12.86% vs. 3.78%) and Buerba et al. [11] (NSQIP Class III subset, n = 10,387; wound complication rate 5.6% vs. 2.0%). The pooled crude odds ratio was 3.03 (95% CI 2.20–4.17; Cochran’s Q = 0.33, p = .57; I² = 0%; τ² = 0). This estimate is consistent in direction and approximate magnitude with the Jiang et al. [2] anchor of 2.33 (95% CI 1.94–2.79) and supports the robustness of the obesity-SSI association across study designs and time periods. We did not extend this approach to other primary outcomes because the available studies did not consistently report extractable 2 × 2 data by BMI category. To assess whether the principal complication signals were disproportionately driven by studies at serious risk of bias, we re-examined each outcome with the 7 ROBINS I serious-rated studies removed, these were largely administrative-database analyses: The direction and approximate magnitude of association for surgical site infection, venous thromboembolism, mortality, and revision surgery were preserved when restricted to moderate-risk studies, supporting the robustness of these signals. The risk of complications appears to be dose dependent, escalating significantly in the highest BMI categories. A retrospective study by Katsevman et al. [32] focusing specifically on ``super obese'' patients (BMI ≥ 50 kg/m²) found a threefold higher overall complication rate compared to controls with a BMI < 50 kg/m² (30% vs. 10%; OR 3.9). The risk was even more pronounced when comparing super obese patients to those with Class I through III obesity (BMI 30–50 kg/m²), with an odds ratio of 8.6 for complications. The most common complication in the super obese cohort was wound dehiscence or infection (13%). This elevated risk was particularly acute in nonelective procedures, where the complication rate for super obese patients reached 40%, compared to 14% in controls [32]. Metabolic syndrome as an independent risk factor Beyond BMI alone, the presence of metabolic syndrome (MetS) significantly increases risk. A propensity score matched analysis by Lovecchio et al. [36] found that even after controlling for obesity, patients with MetS had higher rates of wound complications (3.8% vs. 2.7%), readmissions (7.4% vs. 2.2%), and extended length of stay [36]. Similarly, Thomas et al. found that MetS in TLIF patients was associated with greater odds of wound infections (aOR 1.59), pulmonary events (aOR 1.55), renal events (aOR 2.87), and sepsis (aOR 1.68) [18]. This risk extends to the cervical spine, where a large database study on cervical laminoplasty found that while high grade obesity (Class II–III) was independently associated with longer operative times, it was the presence of MetS that independently predicted deep incisional surgical site infections (OR 3.25), postoperative pneumonia (OR 4.17), and prolonged hospitalization (OR 1.68) [46]. Reoperation and fusion failure In the Jiang meta-analysis, obesity was associated with 43% higher odds of requiring revision surgery (Table 2) [2]. This pattern was supported by large contemporary cohorts, including a prospective analysis by Lange et al. identifying BMI as an independent predictor of revision (OR 1.07) [33]. Hammer et al. similarly reported that revision risk increased with each additional BMI point (OR 1.10), with revision rates rising from 5.7% in normal weight patients to 31.0% in extremely obese patients (BMI ≥ 35) [35]. Regarding fusion success, Khalooeifard et al. [34] found higher nonfusion rates in the obese group and identified obesity as an independent predictor of nonfusion. Sensitivity synthesis restricted to studies reporting outcomes for the BMI ≥ 40 stratum Because the title and PICOS framework of this review specify Class III obesity (BMI ≥ 40 kg/m²) as the population of interest, while a number of included studies pooled all classes of obesity (BMI ≥ 30), we performed a sensitivity synthesis restricted to studies that specifically reported outcomes for the BMI ≥ 40 stratum. Twelve studies in our included set met this criterion: Buerba et al. [11], Krüger et al. [23], Claus et al. [25], Jain et al. [17], Cha et al. [27], Chotai et al. [28], Brown et al. [19], Singh et al. [20], Katsevman et al. [32], Joseph et al. [45], Hammer et al. [35] (with continuous BMI and a ≥35 stratum), and Ravindra et al. [26]. Within this BMI ≥ 40 subset, the principal qualitative signals of the broader review were preserved: morbid obesity remained associated with higher surgical site infection, longer operative time, higher blood loss, longer length of stay, and higher revision rates compared with nonobese controls. The magnitude of effect, where directly reported, was generally larger than in studies using a BMI ≥ 30 cutoff (consistent with the dose-response pattern observed by Katsevman et al. [32] and Hammer et al. [35]). Notably, the patient-reported outcome signal was less consistent in this restricted subset, with several BMI ≥ 40 studies reporting equivalent improvements in disability and pain scores compared with nonobese controls. This sensitivity analysis supports the central conclusions of the review while acknowledging that the strongest dose-response signals are concentrated in studies that explicitly stratified at BMI ≥ 40 or higher. Synthesis of healthcare resource utilization and patient reported outcomes Healthcare resource utilization was consistently higher in morbidly obese patients across procedures and study designs. For a quantitative benchmark, pooled continuous outcomes from Jiang et al. are presented. Resource Utilization: Studies consistently reported that surgery in obese patients was more resource intensive. In the Jiang meta-analysis, operative time was longer and estimated blood loss was higher [2]. These differences have direct financial consequences. A time driven activity-based costing analysis of TLIF procedures found that each unit increase in BMI added 1.90 minutes to total operating room time, corresponding to a $25.72 increase in personnel costs, although total intraoperative costs were not significantly impacted [47]. LeRoy et al. also found that patients with Class II to III obesity were prescribed more than double the number of postoperative opioids and remained on them for a significantly longer duration [42]. Patient Reported Outcomes (PROs): The synthesis of PROs revealed a nuanced picture. Several studies, including a large meta-analysis by Goyal et al. and a cohort study by Manson et al., found that obese patients achieve statistically significant improvements in pain and function, with the magnitude of improvement being comparable to nonobese patients [6,37]. However, other high-quality evidence suggests these gains may be attenuated. A large prospective study by Nakajima et al. found that obese patients reported significantly less improvement in leg pain [40]. A large registry study by Hareni et al. found that obese patients had lower satisfaction rates and less improvement in leg pain and disability scores compared to normal weight patients [41]. Furthermore, a large registry study by Mulvaney et al. which specifically excluded patients with surgical complications, found that increasing BMI was independently associated with worse 1 year disability scores and a decreased likelihood of achieving a minimal clinically important difference (MCID) [44]. Adding another layer of complexity, a study by Siccoli et al. found that dynamic preoperative weight change (either loss or gain) had no significant influence on long term PROs, suggesting that simple weight fluctuation may not be a key determinant of functional outcome [38]. Health Economics and Preoperative Optimization: The synthesis of economic studies indicates that spine surgery in the morbidly obese is a costly endeavor with questionable value from a traditional cost utility perspective. Chotai et al. [28] analyzed ACDF and found that while it was moderately cost effective in Class II obese patients, the cost per quality adjusted life year (QALY) gained for morbidly obese patients (BMI ≥ 40) was significantly higher at $138,094 [28]. Brown et al. [19] found that total surgical costs for adult spinal deformity correction were highest in the Obese Class III group, at $53,856, compared to $49,689 for normal weight patients. Critically, a meta-analysis by Gupta et al. found that undergoing bariatric surgery before spine surgery was associated with a significantly lower rate of all cause medical complications (RR: 0.54), though it did not significantly impact surgical complications or readmission rates [13]. This finding provides direct evidence for the clinical benefit of preoperative weight loss and supports the economic modeling by Paranjape et al. [53], which demonstrated that a combined protocol (bariatric surgery followed by lumbar fusion) was the dominant strategy over a 10-year horizon, being both less costly and more effective. However, conflicting evidence exists, as a single center study by Couch et al. found that patients with a history of bariatric surgery had a significantly increased risk of surgical site infection, readmission, and revision surgery, highlighting the potential for malnutrition and other post bariatric sequelae to negatively impact outcomes [43]. Subgroup analysis: the role of surgical approach (MIS vs. Open) A key objective was to determine if MIS techniques could mitigate the heightened risks in the obese population. The evidence was conflicting. A meta-analysis by Wang et al. focusing on MIS spine surgery found no significant difference in postoperative complications between obese and nonobese patients, although operative times were longer for the obese group [12]. Similarly, a subgroup analysis within a larger meta-analysis by Goyal et al. found that the increased risk of complications and reoperations seen in open surgery was not significant for patients undergoing MIS procedures [6]. Conversely, other studies suggest that the benefits of MIS may not fully materialize in this population. Krüger et al. found that while short term complication rates were comparable between morbidly obese and normal weight patients undergoing MIS TLIF, long term follow up suggested higher complication rates and worse pain outcomes in the obese group [23]. Across multiple reports, operative times for MIS procedures were consistently longer in obese patients [13]. Care Disparities: The available literature paints a stark picture of racial and socioeconomic disparities in spine surgery, which are likely amplified in the context of morbid obesity. A study by Sanford et al. [48] found that Black and Native American patients had a disproportionately high comorbidity burden and experienced more 30-day complications, with race itself being an independent risk factor for adverse events. Mohanty et al. demonstrated that these disparities begin preoperatively; Black patients and those from low socioeconomic backgrounds present for surgery with significantly worse baseline physical health and higher pain scores [49]. This is compounded by evidence showing that non-Hispanic Black patients also experience worse outcomes and higher complication rates after bariatric surgery itself [50]. This creates a cycle of disadvantage where the very populations most in need of preoperative optimization face the greatest systemic barriers to accessing it. Discussion Summary of principal findings This systematic review synthesizes contemporary comparative evidence showing that morbid obesity is associated with higher perioperative complication rates and greater resource utilization after elective spine surgery [3,6,11,14,15]. Published meta-analyses and large registry-based cohorts consistently demonstrate higher risks of surgical site infection, venous thromboembolism, and reoperation, with additional signal for higher mortality in the highest BMI strata [6,11,12,15,16,[21], [22],32,33,35]. At the same time, functional improvement after surgery remains achievable for many patients, although the probability of reaching patient reported outcome thresholds may be lower [6,37,40,41,44]. Taken together, these findings support risk stratification that extends beyond BMI alone and incorporates metabolic health and procedure specific factors [36,39,51]. The MIS paradox: a solution with caveats One of the most pressing clinical questions for surgeons is whether the adoption of minimally invasive surgery can abrogate the risks associated with obesity [6,13,22,25]. The theoretical premise is sound: smaller incisions and less muscle dissection should, in principle, lead to lower infection rates, less blood loss, and faster recovery [3,6,13]. However, the evidence synthesized in this review reveals a ``MIS paradox,'' where the expected benefits are not consistently realized [6,13,23]. The logical chain leading to this paradox begins with the observation that MIS procedures are technically more demanding in patients with a large body habitus [13,22,23]. This is reflected in the consistent finding of significantly longer operative times [12,47]. Prolonged surgical duration is, in itself, an independent and well-established risk factor for numerous complications, including SSI and VTE [11,15,16]. Therefore, a situation arises where the potential benefit of a smaller incision (reduced tissue trauma) is in direct competition with the harm of a longer procedure (increased exposure time, greater anesthetic load). The net effect on patient outcomes becomes ambiguous. This is borne out in the conflicting results of comparative studies: some show equivalent outcomes to open surgery, while others suggest worse long-term results for the morbidly obese undergoing MIS [22,23,25]. A key finding from 2 separate meta-analyses is that the increased risk of complications and reoperations associated with obesity in open surgery was not statistically significant in the MIS subgroup [6,12]. This leads to a critical realization: the variable of interest may not be the surgical approach ("MIS vs. Open") but rather the unmeasured confounder of surgeon expertise. The learning curve for MIS spine surgery is steep, and it is substantially steeper in the obese population. An experienced, high volume MIS surgeon may be able to complete the procedure efficiently, thus realizing the benefits of minimal invasion while minimizing the harm of prolonged operative time. Conversely, a less experienced surgeon may find the procedure so challenging that the extended duration and potential for technical errors outweigh any advantages of the approach. The current body of literature, which is dominated by retrospective database studies, almost universally fails to account for surgeon volume or specific experience with bariatric patients. This is a critical limitation and suggests that a blanket recommendation for or against MIS is inappropriate [9]. The evidence does not support the conclusion that ``MIS is better for obese patients.'' Rather, it supports the more nuanced conclusion that ``expertly performed MIS may be better,'' and inexpertly performed MIS may be worse than a well-executed open procedure [6,12,23]. The role of metabolic syndrome The data from studies by Lovecchio, Thomas, and He underscore a critical point: the risk associated with obesity is not merely mechanical [18,36,39]. Metabolic syndrome (MetS), a cluster of conditions including hypertension, hyperglycemia, and dyslipidemia that often accompanies obesity, emerges as an independent and powerful predictor of adverse events. Even after controlling for BMI, the presence of MetS significantly increases the risk for wound complications, readmissions, and a host of systemic issues [36]. This suggests that the systemic inflammatory and hormonal dysregulation characteristic of MetS is a key driver of perioperative risk, separate from the technical challenges posed by a large body habitus [36,39]. This finding has profound implications for preoperative risk stratification, suggesting that screening for and addressing the components of MetS may be as important, if not more so, than focusing on BMI alone [36,39,51]. Emerging pharmacological optimization strategies The evidence strongly supports preoperative optimization, with bariatric surgery being the most effective and cost-effective intervention currently available [13]. However, bariatric surgery is an invasive procedure with its own risks and barriers to access [13,43]. A recent review by Barkyoumb et al. proposes a novel, less invasive approach: the use of glucagon like peptide 1 receptor agonists (GLP 1 RAs) [52]. These medications, widely known for their efficacy in managing type 2 diabetes and promoting significant weight loss, offer a multidimensional solution [52]. They not only induce weight loss rivaling that of bariatric surgery but also improve glycemic control and have potentially cardioprotective and osteogenic properties [52]. The convenience and dependability of a pharmaceutical intervention make GLP 1 RAs an exceptionally promising tool for mitigating obesity related surgical risk [52]. While direct evidence on their use specifically for preoperative optimization in spine surgery is still needed, they represent a critical area for future research and could revolutionize how we prepare high risk patients for surgery [52]. Beyond the narrative case for pharmacological optimization, an emerging body of comparative evidence is beginning to test whether GLP 1 RA exposure translates into measurable perioperative benefit in obese surgical populations. Early reports from orthopedic and general surgical cohorts suggest reduced rates of wound complications and improved perioperative glycemic control in GLP 1 RA-exposed patients compared with nonmedicated obese controls, though direct evidence from elective spine surgery cohorts remains lacking. These emerging signals must be weighed against the documented risk of retained gastric contents and the resulting perioperative anesthetic guidance recommending case-by-case decisions on continuation versus holding of GLP 1 RAs prior to elective surgery. Prospective trials in spine surgery populations, controlling explicitly for duration of therapy, magnitude of preoperative weight loss, and timing of the last dose, are urgently needed to clarify the net perioperative effect of this rapidly evolving therapeutic class. Beyond the OR: the economic and equity imperative The findings of this review extend far beyond the operating room, carrying significant implications for health policy and system design [11]. The economic analysis is clear: the current model of proceeding directly to surgery in morbidly obese patients is economically inefficient [28]. The combination of higher complication rates, increased resource utilization, and frequent readmissions leads to substantially higher costs [3,14]. The cost utility analyses, showing values as high as $138,094 per QALY for ACDF in the morbidly obese, place these procedures well outside the bounds of what is typically considered cost effective [28]. In this context, the finding that a strategy of preoperative bariatric surgery is dominant both saving costs and improving outcomes over a 10-year period is a landmark piece of evidence [53]. It reframes preoperative optimization not as an elective adjunct or a clinical nicety, but as a financial and ethical imperative for any health system committed to providing value-based care. However, implementing such an optimization strategy exposes a deep seated and dangerous fault line in the healthcare system: the ``preoperative optimization disparity trap.'' The evidence on health disparities is unequivocal [48,49,54]. Racial and ethnic minority patients, particularly Black individuals, and those of lower socioeconomic status, present for spine surgery with a greater burden of comorbidity and worse baseline health [48,49]. They subsequently experience higher rates of postoperative complications, more non home discharges, and more frequent readmissions [48,54]. The reasons for this are multifactorial and rooted in structural determinants of health, including disparate access to primary care, healthy food, and safe living environments. The optimization strategy itself, bariatric surgery, is subject to these same disparities, with minority patients facing significant barriers to access and experiencing worse outcomes when they do receive it [50]. This creates a perilous potential for well-intentioned policy to cause unintended harm. A rigid, ``nonnegotiable'' institutional policy that mandates preoperative optimization (eg, achieving a target BMI or HbA1c level before surgery is approved) without concurrently providing the resources to achieve it could paradoxically worsen health inequities. Such a policy risks creating a 2-tiered system. Well resourced, predominantly white patients will successfully navigate the optimization pathway and receive safer, more effective surgery. In contrast, underserved, predominantly minority patients may be unable to access or afford these preparatory interventions and could be indefinitely denied necessary surgical care or forced to accept a higher risk procedure. Therefore, any institutional or payer level mandate for preoperative optimization must be inextricably linked with dedicated, funded programs for patient navigation, nutritional counseling, and equitable access to services like bariatric surgery [50]. Without this crucial linkage, optimization policies risk becoming another mechanism of systemic disadvantage. Strengths and limitations of the evidence synthesis The principal strength of this review lies in its comprehensive and systematic methodology. The broad search strategy, inclusion of grey literature, and adherence to PRISMA guidelines were designed to capture the totality of the available evidence. However, the conclusions must be interpreted in light of several important limitations inherent in the source material. The most significant limitation is the predominance of retrospective, nonrandomized study designs. This introduces a moderate to serious risk of bias across the evidence base, particularly from confounding by indication and selection bias. Secondly, there is considerable heterogeneity in the reporting of BMI [3,15,16]. Many studies group all patients with a BMI > 30 together, diluting the specific effect on the morbidly obese, while others use varying cutoffs (eg, BMI ≥ 35 vs. ≥40), which complicates direct comparison and pooling [15,16,32]. Finally, as previously discussed, large administrative databases, the source for many included studies, suffer from a critical lack of granularity [9]. They often fail to capture crucial clinical details and potential confounders such as the specifics of surgical technique, surgeon experience, patient socioeconomic status, and the very PROs that define surgical success. This limits the ability to draw definitive causal inferences. To allow readers to calibrate trust in our conclusions, we explicitly distinguish here between findings that are robust to risk-of-bias filtering and those that are not. The direction and approximate magnitude of association between morbid obesity and surgical site infection, venous thromboembolism, mortality, and revision surgery were preserved when restricted to studies at moderate ROBINS I risk, and were further corroborated by the contemporary anchor-validation pool for SSI reported in the Results (pooled OR 3.03, 95% CI 2.20–4.17). These signals can be regarded as the most defensible quantitative claims of the review. By contrast, the precise magnitude of cost differentials, length-of-stay increments, and the very large effect estimates reported for metabolic-syndrome attributable cervical infections and pneumonia depend heavily on administrative-database studies judged at serious ROBINS I risk. These signals should therefore be interpreted as hypothesis-generating and used cautiously in clinical or policy decisions. Formal GRADE assessment of the 4 primary outcomes (Table 3) yielded an overall certainty of evidence of Low for surgical site infection and Very Low for venous thromboembolism, mortality, and revision surgery. These ratings reflect the observational nature of the evidence base, the predominance of administrative-database designs, and additional imprecision for the latter 3 outcomes. The Low rating for SSI is supported by the consistency between the Jiang 2014 anchor, our contemporary anchor-validation pool, and the qualitative signal in the BMI ≥ 40 sensitivity subset. The implication for the value-based care framework we propose is that recommendations regarding SSI risk and its mitigation rest on a more defensible evidence base than recommendations regarding cost-effectiveness or metabolic-syndrome attributable complications, which require corroboration by higher-quality prospective studies before being incorporated into payer-level mandates.Table 3Certainty of evidence (GRADE) for review findings synthesized narratively without meta-analysis.Table 3 dummy alt textFindingNo. studies (design)Risk of biasInconsistencyIndirectnessImprecisionPublication biasOther considerationsDirection of effect across studiesCertaintySurgical site infection14SeriousNot seriousSeriousNot seriousUndetectedDose-response (+1)Consistently higher in morbidly obeseLowVenous thromboembolism6SeriousNot seriousSeriousSeriousUndetected—Consistently higher; magnitude substantialVery lowMortality6SeriousNot seriousSeriousSeriousUndetected—Consistently higher in highest BMI strataVery lowRevision/reoperation9SeriousNot seriousSeriousSeriousUndetected—Consistently higher; magnitude modestVery lowOperative time11 NRSISeriousNot seriousNot seriousNot seriousUndetected—Uniformly longer in morbidly obeseLowLength of stay10SeriousNot seriousNot seriousNot seriousUndetected—Uniformly longerLowEstimated blood loss8SeriousMinor concernsNot seriousNot seriousUndetected—Generally higher; attenuated with MISLowDirect costs/cost-utility3SeriousSeriousSeriousSeriousUndetected—Higher costs; cost-effectiveness unfavorable at BMI ≥ 40Very lowPRO improvement (any)8ModerateNot seriousNot seriousNot seriousUndetected—Morbidly obese patients improve after surgeryLowPRO—likelihood of achieving MCID5ModerateSeriousNot seriousNot seriousUndetected—Conflicting: some studies show reduced MCID achievementVery lowDose-response with BMI severity3ModerateNot seriousNot seriousSeriousUndetectedCoherent gradient (+1)Monotonic increase in complications with BMI categoryLowMIS approach mitigates obesity risk5 (Wang, Goyal, Adogwa, Krüger, Claus)SeriousSeriousSeriousSeriousUndetected—Conflicting; benefit not consistently realizedVery lowMetabolic syndrome as independent risk factor4 (Lovecchio, Thomas, He, Shahbandi)ModerateNot seriousNot seriousNot seriousUndetected—Consistently higher risk beyond BMI aloneLowPreoperative bariatric surgery improves outcomes3 (Gupta, Paranjape, Couch)SeriousSeriousNot seriousSeriousSuspected—Conflicting; medical complications reduced but surgical outcomes uncertainVery lowRacial/socioeconomic disparities in outcomes4 (Khan, Sanford, Mohanty, Masanam)SeriousNot seriousSeriousNot seriousUndetected—Consistently worse baseline, access, and postoperative outcomes in minoritised groupsLowAll findings derive from nonrandomised studies of interventions (NRSI) and therefore start at Low certainty per GRADE guidance, with downgrading applied for serious concerns within each domain and upgrading applied where dose-response or large-effect signals were identified. Domains were judged as follows. Risk of bias: assessed across contributing studies using ROBINS-I, with ``serious'' indicating that 2 or more studies were judged at serious risk or that structural confounding (eg, surgeon volume, unmeasured comorbidity) was unaddressed. Inconsistency: judged on agreement in the direction of effect across contributing studies, given the absence of pooled estimates; ``serious'' indicates direct conflict between studies. Indirectness: judged on alignment of the contributing studies with the review PICO, with attention to BMI cut-off, procedure mix, and population. Imprecision: judged on the cumulative volume of contributing data and consistency of direction rather than on confidence-interval width; ``serious'' indicates few contributing studies, sparse events, or estimates close to the null. Publication bias: judged narratively; ``suspected'' indicates a plausible asymmetry in the available literature. Other considerations: upgrades for dose-response or large-effect signals were applied where supported. Certainty levels are interpreted as: High—we are very confident the true effect lies close to the estimated effect; Moderate—we are moderately confident; Low—our confidence is limited; Very low—we have very little confidence.NRSI, nonrandomised studies of interventions; MCID, minimal clinically important difference; MIS, minimally invasive surgery; PRO, patient-reported outcome; BMI, body mass index. Table 3Certainty of evidence (GRADE) for review findings synthesized narratively without meta-analysis.Table 3 dummy alt textFindingNo. studies (design)Risk of biasInconsistencyIndirectnessImprecisionPublication biasOther considerationsDirection of effect across studiesCertaintySurgical site infection14SeriousNot seriousSeriousNot seriousUndetectedDose-response (+1)Consistently higher in morbidly obeseLowVenous thromboembolism6SeriousNot seriousSeriousSeriousUndetected—Consistently higher; magnitude substantialVery lowMortality6SeriousNot seriousSeriousSeriousUndetected—Consistently higher in highest BMI strataVery lowRevision/reoperation9SeriousNot seriousSeriousSeriousUndetected—Consistently higher; magnitude modestVery lowOperative time11 NRSISeriousNot seriousNot seriousNot seriousUndetected—Uniformly longer in morbidly obeseLowLength of stay10SeriousNot seriousNot seriousNot seriousUndetected—Uniformly longerLowEstimated blood loss8SeriousMinor concernsNot seriousNot seriousUndetected—Generally higher; attenuated with MISLowDirect costs/cost-utility3SeriousSeriousSeriousSeriousUndetected—Higher costs; cost-effectiveness unfavorable at BMI ≥ 40Very lowPRO improvement (any)8ModerateNot seriousNot seriousNot seriousUndetected—Morbidly obese patients improve after surgeryLowPRO—likelihood of achieving MCID5ModerateSeriousNot seriousNot seriousUndetected—Conflicting: some studies show reduced MCID achievementVery lowDose-response with BMI severity3ModerateNot seriousNot seriousSeriousUndetectedCoherent gradient (+1)Monotonic increase in complications with BMI categoryLowMIS approach mitigates obesity risk5 (Wang, Goyal, Adogwa, Krüger, Claus)SeriousSeriousSeriousSeriousUndetected—Conflicting; benefit not consistently realizedVery lowMetabolic syndrome as independent risk factor4 (Lovecchio, Thomas, He, Shahbandi)ModerateNot seriousNot seriousNot seriousUndetected—Consistently higher risk beyond BMI aloneLowPreoperative bariatric surgery improves outcomes3 (Gupta, Paranjape, Couch)SeriousSeriousNot seriousSeriousSuspected—Conflicting; medical complications reduced but surgical outcomes uncertainVery lowRacial/socioeconomic disparities in outcomes4 (Khan, Sanford, Mohanty, Masanam)SeriousNot seriousSeriousNot seriousUndetected—Consistently worse baseline, access, and postoperative outcomes in minoritised groupsLowAll findings derive from nonrandomised studies of interventions (NRSI) and therefore start at Low certainty per GRADE guidance, with downgrading applied for serious concerns within each domain and upgrading applied where dose-response or large-effect signals were identified. Domains were judged as follows. Risk of bias: assessed across contributing studies using ROBINS-I, with ``serious'' indicating that 2 or more studies were judged at serious risk or that structural confounding (eg, surgeon volume, unmeasured comorbidity) was unaddressed. Inconsistency: judged on agreement in the direction of effect across contributing studies, given the absence of pooled estimates; ``serious'' indicates direct conflict between studies. Indirectness: judged on alignment of the contributing studies with the review PICO, with attention to BMI cut-off, procedure mix, and population. Imprecision: judged on the cumulative volume of contributing data and consistency of direction rather than on confidence-interval width; ``serious'' indicates few contributing studies, sparse events, or estimates close to the null. Publication bias: judged narratively; ``suspected'' indicates a plausible asymmetry in the available literature. Other considerations: upgrades for dose-response or large-effect signals were applied where supported. Certainty levels are interpreted as: High—we are very confident the true effect lies close to the estimated effect; Moderate—we are moderately confident; Low—our confidence is limited; Very low—we have very little confidence.NRSI, nonrandomised studies of interventions; MCID, minimal clinically important difference; MIS, minimally invasive surgery; PRO, patient-reported outcome; BMI, body mass index. Certainty of evidence (GRADE) for review findings synthesized narratively without meta-analysis. All findings derive from nonrandomised studies of interventions (NRSI) and therefore start at Low certainty per GRADE guidance, with downgrading applied for serious concerns within each domain and upgrading applied where dose-response or large-effect signals were identified. Domains were judged as follows. Risk of bias: assessed across contributing studies using ROBINS-I, with ``serious'' indicating that 2 or more studies were judged at serious risk or that structural confounding (eg, surgeon volume, unmeasured comorbidity) was unaddressed. Inconsistency: judged on agreement in the direction of effect across contributing studies, given the absence of pooled estimates; ``serious'' indicates direct conflict between studies. Indirectness: judged on alignment of the contributing studies with the review PICO, with attention to BMI cut-off, procedure mix, and population. Imprecision: judged on the cumulative volume of contributing data and consistency of direction rather than on confidence-interval width; ``serious'' indicates few contributing studies, sparse events, or estimates close to the null. Publication bias: judged narratively; ``suspected'' indicates a plausible asymmetry in the available literature. Other considerations: upgrades for dose-response or large-effect signals were applied where supported. Certainty levels are interpreted as: High—we are very confident the true effect lies close to the estimated effect; Moderate—we are moderately confident; Low—our confidence is limited; Very low—we have very little confidence. NRSI, nonrandomised studies of interventions; MCID, minimal clinically important difference; MIS, minimally invasive surgery; PRO, patient-reported outcome; BMI, body mass index. A practical framework for counseling and decision-making in the morbidly obese spine surgery candidate Translating this evidence base into bedside practice requires moving beyond a single BMI threshold. We propose 4 domains for surgeon counseling and shared decision-making in patients with BMI ≥ 40 kg/m² considering elective spine surgery.1.Risk re-stratification beyond BMI. Preoperative screening should include the components of metabolic syndrome (HbA1c, blood pressure, lipid profile, waist circumference). The independent contribution of metabolic syndrome to perioperative risk, demonstrated in propensity-matched cohorts after adjustment for BMI, indicates that the BMI value alone should not drive the decision to proceed, defer, or decline surgery. Patients with elevated BMI but metabolically healthy profiles may carry lower risk than patients with lower BMI and active metabolic syndrome.2.Procedure selection. For decompression-only pathology (eg, central lumbar stenosis without instability), the marginal risk attributable to morbid obesity is smaller and the benefit-risk balance generally favors proceeding after standard optimization. For instrumented fusion, particularly multilevel or deformity correction, the documented increases in operative time, blood loss, SSI, revision, and cost are large enough to warrant a structured optimization pathway before scheduling. The threshold for staged or sequential procedures should be lower in this population.3.Approach selection (MIS vs. open). The evidence does not support a blanket preference for minimally invasive over open surgery in the morbidly obese. MIS may attenuate certain complications when performed by surgeons experienced with MIS techniques in obese habitus, but in lower-volume hands the prolonged operative time may negate any benefit. Surgeons should counsel patients explicitly about the surgeon-experience component of approach choice rather than presenting MIS as inherently safer.4.Optimization pathway. Where feasible and equitable, structured preoperative optimization (nutritional assessment, glycemic control, smoking cessation, physical conditioning, and where appropriate and accessible, bariatric surgery or GLP 1 receptor agonist therapy) should be offered. Mandatory BMI cutoffs imposed without funded patient navigation and access to optimization services risk worsening health inequities, as discussed in the preoperative optimization disparity trap. Patients who cannot achieve a target BMI should not be categorically denied surgery; they should be offered the lowest-risk procedure compatible with their pathology, with explicit counselling about residual risk. 1.Risk re-stratification beyond BMI. Preoperative screening should include the components of metabolic syndrome (HbA1c, blood pressure, lipid profile, waist circumference). The independent contribution of metabolic syndrome to perioperative risk, demonstrated in propensity-matched cohorts after adjustment for BMI, indicates that the BMI value alone should not drive the decision to proceed, defer, or decline surgery. Patients with elevated BMI but metabolically healthy profiles may carry lower risk than patients with lower BMI and active metabolic syndrome. Risk re-stratification beyond BMI. Preoperative screening should include the components of metabolic syndrome (HbA1c, blood pressure, lipid profile, waist circumference). The independent contribution of metabolic syndrome to perioperative risk, demonstrated in propensity-matched cohorts after adjustment for BMI, indicates that the BMI value alone should not drive the decision to proceed, defer, or decline surgery. Patients with elevated BMI but metabolically healthy profiles may carry lower risk than patients with lower BMI and active metabolic syndrome. 2.Procedure selection. For decompression-only pathology (eg, central lumbar stenosis without instability), the marginal risk attributable to morbid obesity is smaller and the benefit-risk balance generally favors proceeding after standard optimization. For instrumented fusion, particularly multilevel or deformity correction, the documented increases in operative time, blood loss, SSI, revision, and cost are large enough to warrant a structured optimization pathway before scheduling. The threshold for staged or sequential procedures should be lower in this population. Procedure selection. For decompression-only pathology (eg, central lumbar stenosis without instability), the marginal risk attributable to morbid obesity is smaller and the benefit-risk balance generally favors proceeding after standard optimization. For instrumented fusion, particularly multilevel or deformity correction, the documented increases in operative time, blood loss, SSI, revision, and cost are large enough to warrant a structured optimization pathway before scheduling. The threshold for staged or sequential procedures should be lower in this population. 3.Approach selection (MIS vs. open). The evidence does not support a blanket preference for minimally invasive over open surgery in the morbidly obese. MIS may attenuate certain complications when performed by surgeons experienced with MIS techniques in obese habitus, but in lower-volume hands the prolonged operative time may negate any benefit. Surgeons should counsel patients explicitly about the surgeon-experience component of approach choice rather than presenting MIS as inherently safer. Approach selection (MIS vs. open). The evidence does not support a blanket preference for minimally invasive over open surgery in the morbidly obese. MIS may attenuate certain complications when performed by surgeons experienced with MIS techniques in obese habitus, but in lower-volume hands the prolonged operative time may negate any benefit. Surgeons should counsel patients explicitly about the surgeon-experience component of approach choice rather than presenting MIS as inherently safer. 4.Optimization pathway. Where feasible and equitable, structured preoperative optimization (nutritional assessment, glycemic control, smoking cessation, physical conditioning, and where appropriate and accessible, bariatric surgery or GLP 1 receptor agonist therapy) should be offered. Mandatory BMI cutoffs imposed without funded patient navigation and access to optimization services risk worsening health inequities, as discussed in the preoperative optimization disparity trap. Patients who cannot achieve a target BMI should not be categorically denied surgery; they should be offered the lowest-risk procedure compatible with their pathology, with explicit counselling about residual risk. Optimization pathway. Where feasible and equitable, structured preoperative optimization (nutritional assessment, glycemic control, smoking cessation, physical conditioning, and where appropriate and accessible, bariatric surgery or GLP 1 receptor agonist therapy) should be offered. Mandatory BMI cutoffs imposed without funded patient navigation and access to optimization services risk worsening health inequities, as discussed in the preoperative optimization disparity trap. Patients who cannot achieve a target BMI should not be categorically denied surgery; they should be offered the lowest-risk procedure compatible with their pathology, with explicit counselling about residual risk. Conclusions This systematic review synthesizes current evidence on the impact of obesity on outcomes after elective spine surgery. Across procedures, higher BMI is associated with more complications, greater healthcare resource utilization, and higher reoperation rates, with the highest risk in morbid obesity (BMI ≥ 40 kg/m²) and extreme obesity (BMI ≥ 50 kg/m²). These findings support a patient centered, value-based pathway that combines preoperative risk stratification beyond BMI alone (including metabolic health, procedural complexity, and approach specific risk) with shared decision making and scalable, equitable optimization to improve outcomes without widening disparities. The direction of association between morbid obesity and excess perioperative risk is robust to risk-of-bias filtering and was further corroborated by a contemporary anchor-validation pool for surgical site infection. However, the precise magnitude of effect for cost, length of stay, and metabolic-syndrome attributable risk relies on administrative-database evidence judged at serious risk of bias on ROBINS I and should be interpreted with corresponding caution when used to inform institutional or payer-level policy. Data availability All data analyzed in this study are derived from previously published reports and publicly available sources. Ethics statement Ethics approval was not required because this work synthesizes previously published data and includes no new human participant data. Funding No funding was received for this work. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.