Preoperative depression and outcomes after metabolic and bariatric surgery: A systematic narrative review

Preoperative depression is prevalent among patients undergoing metabolic and bariatric surgery (MBS) and is a potentially modifiable risk factor. However, the impact of preoperative depression on MBS outcomes has not been systematically reviewed. A search of MEDLINE, Embase, Cochrane, and PsychINFO (inception to June 2023) was conducted for studies reporting associations between preoperative depression and any clinical or patient‐reported outcomes after MBS. Eighteen studies (5 prospective and 13 retrospective) reporting on 5933 participants were included. Most participants underwent gastric bypass or sleeve gastrectomy. Meta‐analyses were not conducted due to heterogeneity in reported outcomes; findings were instead synthesized using a narrative and tabular approach. Across 13 studies (n = 3390) the associations between preoperative depression and weight loss outcomes at 6–72 months were mixed overall. This may be related to differences in cohort characteristics, outcome definitions, and instruments used to measure depression. A small number of studies reported that preoperative depression was associated with lower quality of life, worse acute pain, and more perioperative complications after surgery. Most of the included studies were deemed to be at high risk of bias, resulting in low or very low certainty of evidence according to the Risk of Bias In Non‐randomized Studies ‐ of Exposure (ROBINS‐E) tool. While the impact of preoperative depression on weight loss after MBS remains unclear, there is early evidence that depression has negative consequences on other patient‐important outcomes. Adequately powered studies using more sophisticated statistical methods are needed to accurately estimate these associations.

Obesity and depression are among the top public health concerns, given the rapid increase in global prevalence of both conditions in recent decades. 1,2The two exhibit a complex bidirectional relationship, with each being a strong independent risk factor for exacerbating the other. 3,4Furthermore, obesity and depression can have compounding negative effects on medical comorbidity, disability, and health-related quality of life. 5,6r patients with obesity, metabolic and bariatric surgery (MBS) is considered when lifestyle changes and medical management options have been exhausted or if the patients meet surgical indications. 7S is considered for patients with body mass index (BMI) ≥ 35 kg/ m 2 or BMI 30-34.9kg/m 2 with metabolic disease. 8Advances in comprehensive metabolic and bariatric care programs and operative technique have led to MBS being increasingly accessible to patients in developed countries. 9Laparoscopic sleeve gastrectomy and gastric bypass procedures are the most commonly performed MBS and have been established as safe and effective treatments for obesity, resulting in substantial and sustained weight loss, reductions in incidences of cardiovascular disease, diabetes, fatty liver disease, malignancy, and improved overall health-related quality of life. 10,11However, the high burden of pre-existing mental health disorders such as anxiety and depression in this patient population pose a potential risk to the short-and long-term success of MBS. 12 For instance, anhedonia, mood dysregulation, and altered appetite may impair motivation and adherence to the lifestyle and behavioral adaptations that are recommended after undergoing these procedures. 13,14ere is mounting evidence from the broader surgical literature that preoperative mental health is an important predictor of adverse clinical and patient-reported outcome measures such as severe pain, hospital length of stay, and perioperative complications. 15,16Preoperative depression is increasingly being recognized as an important modifiable risk factor, with psychosocial interventions such as cognitive behavioral therapy initiated before surgery demonstrating feasibility and effectiveness for mitigating the risks of adverse postoperative outcomes. 17,18Specifically in the MBS setting, several cohort studies have reported on weight loss outcomes for patients with preoperative depressive symptoms. 19,20However, there appears to be little known about the impact of preoperative depression on other patient-reported and perioperative outcomes after MBS.A better understanding of these relationships could guide the development and timely implementation of tailored interventions in the perioperative period, especially given the high burden of untreated depression and other psychiatric concerns among patients with obesity seeking MBS. 21Therefore, our primary aim was to conduct a comprehensive review of the literature describing the impact of preoperative depression on various outcomes after MBS.

| MATERIALS AND METHODS
The present review is part of a consortium of studies on preoperative depression and perioperative outcomes for various surgical procedures (PROSPERO ID: CRD42022296532).This systematic review of etiology and risk was conducted and reported according to the best practice methods outlined in the JBI Manual for Evidence Synthesis 22 and Synthesis without meta-analysis (SWiM) reporting guidelines. 23

| Eligibility criteria
Studies were eligible for inclusion if they i) measured depression before MBS as either a binary or ordinal variable; and ii) included an association-type analysis (e.g., comparison or regression) between depression and any relevant outcome after surgery (listed below).Eligible participants were adult patients (≥18 years of age) undergoing any type of MBS (e.g., sleeve gastrectomy, gastric bypass).Both prospective and retrospective cohort studies were included.Crosssectional studies, case-control, case series, conference abstracts, editorials, reviews, and basic science articles were excluded.

| Outcomes
Outcomes of interest included but were not limited to (i) MBS-specific outcomes (e.g., weight loss and eating behaviors); (ii) other clinical outcomes (e.g., death, organ dysfunction, admission to an intensive care unit [ICU], re-operation, infection, and anastomotic leak); (iii) patientreported outcomes (e.g., pain, disability, and quality of life); or (iv) healthcare utilization (e.g., hospital length of stay and readmission).

| Study selection and data extraction
All articles were assessed for eligibility in two stages, initially based on titles and abstracts followed by the full texts.Each stage was performed by two independent reviewers using standardized piloted forms on the online Covidence software (Veritas Health Innovation, Melbourne, Australia).Discrepancies were resolved by consensus between the same two reviewers.If disagreement persisted, a third reviewer was consulted.Reviewers were not blinded to authors, institutions, or the journal where the manuscript was published.
Data extraction from each included article was independently performed by two reviewers using a pilot-tested form.Discrepancies were resolved by consensus by the two reviewers involved.Data extracted included study characteristics (author, year of publication, and country), patient characteristics (preoperative depressive status or symptom severity, surgical procedure), outcomes of interest and their corresponding point estimates with standard errors (or confidence intervals) and p-values, and any covariates adjusted for in statistical analyses.

| Rias of bias
Using the Risk Of Bias In Non-randomized Studies -Exposure (ROBINS-E) tool, 24 pairs of reviewers independently assessed the articles for risk of bias, assigning a rating of "low," "some concerns," "high," or "very high" risk of bias across seven domains: confounding, measurement of exposure, selection bias, post-exposure interventions, missing data, measurement of outcome, and reporting bias.An overall risk of bias rating for each study was determined by the highest rating assigned to the study across the seven domains.The pooled ratings for each domain were displayed in a stacked bar plot.

| Data synthesis and analysis
Meta-analysis was considered if at least five studies reported the same outcome and measured preoperative depression using similar definitions.However, in this review, these criteria were not met and a narrative and tabular synthesis of outcome data was conducted.

| Certainty of evidence
The overall certainty of evidence for outcomes from included studies was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework as "very low," "low," "moderate," or "high" certainty. 25This has been described as an appropriate approach regardless of whether effect estimates are pooled and meta-analyses are performed. 26Evidence for each outcome started at "moderate" quality and were downgraded by one level for serious study limitations (risk of biases), indirectness of evidence, serious inconsistency, imprecision of effect estimates, or potential publication bias.Evidence for each outcome were upgraded by one level for large effect sizes, dose-response gradient, or plausible residual confounding that would increase the certainty of effect.

| Study selection and characteristics
The broader search for preoperative depression and outcomes after any surgical procedure retrieved 9769 articles for screening, of which 20 reported association-type analyses between preoperative depression and various outcomes after MBS (Figure 1).Two of these articles were identified as sequential follow-ups of the same prospective cohort, and the more recent one was removed for having greater loss to follow-up. 27,28Therefore, a total of 19 studies were included in this review.No additional articles were identified from searching reference lists.
Across the 19 included studies, a total 6005 participants underwent MBS.The studies were published from 1985 to 2022, with 14 (74%) prospective and 5 (26%) retrospective in design.Nine (47%) of the studies were conducted in the US, and the rest in Germany, France, Italy, Portugal, Turkey, Israel, and Brazil.Patients undergoing gastric bypass procedures were included in 14 (74%) studies, sleeve gastrectomy in 9 (47%) studies, and gastric banding in 3 (16%) studies.In 6 (32%) studies, the cohort was composed of more than one type of MBS procedure.Preoperative depression was measured as an ordinal variable in 9 (47%) and as a binary variable in 13 (68%) of studies (both methods were reported in three studies).Median follow-up across all studies was 12 months (range: 24 h to 36 months).Full baseline characteristics of the included studies are presented in Table 1.The definitions and instruments used to measure depression varied widely.The Beck Depression Inventory (BDI) was the most frequently used instrument (eight studies), operationalized either as an ordinal measure of symptom severity or by dichotomizing at a threshold score; however, the threshold varied between studies.Other methods of measuring depression included pre-existing history or diagnoses (i.e., self-reported and physician diagnosed), and instruments such as the BDI short form (BDI-SF), Symptom Checklist-90-Revised (SCL-90-R), Structured Clinical Interview for DSM Disorders (SCID), Center for Epidemiologic Studies Depression Scale (CES-D), and Hamilton Depression Rating Scale (HAM-D).
The overall findings were mixed (Table 2).In six studies, preoperative depression was associated with various weight loss measures, T A B L E 2 Results from studies reporting associational analyses between preoperative depression and weight loss outcomes.Estimates are only presented from 12-month follow-up, though no differences were observed in the outcomes between groups throughout 3-72 months.
though the direction of the associations was mixed. 19,29,30,33,35,37pressive symptom severity as an ordinal measure was associated with a greater degree of total weight loss, % total weight loss, or % excess weight loss in three studies, but associated with a lesser degree of % excess weight loss or BMI loss in four studies.There was no association observed in the seven remaining studies.Only four studies adjusted for potential confounders in analyses: two adjusted for baseline weight metrics 29,33 and the other two adjusted for age and sex. 37,38However, even among these four studies, the associations were mixed.There did not appear to be any influence of year of publication on the associations reported in studies.GRADE certainty of evidence: low.

| Quality of life
Two studies (n = 511) reported health-related quality of life (HRQoL) as an outcome (Table 3), one using the full 36-item version of the Short Form (SF-36) and the other using the abbreviated 12-item version (SF-12). 34,39In both studies, the mental component score (MCS) and physical component score (PCS) of the instruments were considered as separate outcomes.Preoperative depression was associated with lower postoperative MCS and PCS (i.e., worse quality of life).
However, in one study, the association between depression and PCS was inconsistent throughout 6-, 12-, and 24-month follow-up and did not adjust for any covariates. 34The other study reported negative associations between BDI score with SF-12-PCS at 6 and 12 months after surgery. 39Of note, different sets of covariates were included in the separate analytic models for MCS and PCS in this study, raising suspicion for multiple testing and selective reporting of results.
GRADE certainty of evidence: low.

| Eating behaviors
The Eating Disorder Examination Questionnaire (EDE-Q) was measured as an outcome in one study of 357 patients (Table 3). 34Patients with baseline depression (BDI ≥ 15) had worse EDE-Q scores at 6, 12, and 24 months after surgery compared with those who were not classified as depressed.Another study measured the change in Night Eating Questionnaire (NEQ) score as an outcome (Table 3). 40Patients with depression (BDI-SF ≥ 4) had similar NEQ scores compared with the non-depressed group after surgery.However, the depressed group had a higher score at baseline (16.0 ± 7.7 vs. 10.8 ± 6.4), and the mean NEQ score for subjects without depression did not change after surgery.Furthermore, the time at which subjects completed the NEQ after surgery varied widely, ranging from 6 to 33 months.
GRADE certainty of evidence: very low.

| Postoperative pain
One prospective study of 166 patients undergoing Roux-en-Y gastric bypass surgery reported postoperative pain severity using the visual analog scale at rest (VASr) and induced by coughing (VASi) (Table 3). 41In separate multivariable regression models, HAM-D score was positively associated with both VASr and VASi after adjusting for baseline anxiety and alexithymia scores.No other confounders were adjusted for, and repeated outcome measures were not captured or reported beyond 36 h.GRADE certainty of evidence: very low.

| Postoperative opioid analgesia use
Two prospective studies (n = 188) reported postoperative in-hospital opioid analgesia use (Table 3). 41,42One study reported that preoperative depression (HAM-D ≥ 18) was positively associated with the number of tramadol patient-controlled analgesia (PCA) requests at 36 h after surgery, after adjusting for baseline anxiety and alexithymia. 41The other study found no association between BDI score and milligrams of morphine consumed within 24 h after adjusting for distress tolerance, anxiety, and pain catastrophizing. 42The standard error around this estimate was not reported.GRADE certainty of evidence: very low.

| Cognitive functions
One prospective study of 67 patients measured attention, executive function, memory, and language using the IntegNeuro battery of cognitive tests before and 12 months after surgery (Table 3). 43The degree of improvement in attention, executive function, memory, and language domain scores was not significantly different between those with versus those without a lifetime history of depression.GRADE certainty of evidence: very low.

| Perioperative complications
A single prospective study of 2146 patients evaluated the outcome of perioperative complications within 30 days of surgery (Table 3), defined as the composite of death, deep vein thrombosis or venous thromboembolism, reintervention using percutaneous, endoscopic, or operative techniques, or failure to be discharged from the hospital. 44eoperative depression, measured as an ordinal BDI score, was not associated with the composite outcome.When the cohort was dichotomized using a BDI threshold of ≥10, those with depression had greater odds of complications (adjusted OR 1.77, 95% CI: 1.03 to 3.05, p = 0.04).While these effect estimates were described as adjusted ORs, the specific covariates were not reported.GRADE certainty of evidence: very low.

| Risk of bias
The risk of bias assessment for included studies is depicted in Figure 2. A traffic light plot of domain ratings for each study can be T A B L E 3 Results from studies reporting associational analyses between preoperative depression and outcomes not related to weight loss.found in Appendix S5.Fourteen studies were rated as having a "high" overall risk of bias and one was rated as "very high."Domains in which most studies had a "high" or "very high" risk of bias were related to (i) a lack of accounting for potentially important confounders (or at least basic demographic variables and baseline patientreported outcome measure scores), (ii) measurement of the exposure, and (iii) measurement of the outcome.

| Quality of evidence assessment
A summary of the findings for the various outcomes captured in this review and quality of evidence for each as assessed using the GRADE framework are reported in Table 4.

| DISCUSSION
This study presents a comprehensive review of studies reporting on the relationship between preoperative depression and various clinical and patient-reported outcomes after MBS.Most studies reported weight loss outcomes after surgery.While some studies demonstrated that subjects with depressive symptoms at baseline experienced less weight loss after surgery, some reported the opposite, and others found no association at all.Heterogeneity in definitions of weight loss outcomes, preoperative depression, and covariates included in models are suspected to contribute to these mixed findings.Our review also Risk of bias assessments using the risk of bias in nonrandomized studies -of exposure (ROBINS-E) tool.
T A B L E 4 GRADE summary of findings.included preliminary evidence that depression before MBS was associated with worse quality of life, pain, and more perioperative complications.
Previous longitudinal studies have reported on the inverse association between postoperative measurements of depression and weight loss after MBS. 12 In this present review, no clear relationship between preoperative depression and weight loss outcomes was found, suggesting that patients' depressive state after surgery may be more influential.The reasons for this are unclear, but postoperative mood and motivation are speculated to be key drivers of engaging in new lifestyle behaviors related to feeding, physical activity, and sleep needed to sustain weight loss after surgery. 12In keeping with this hypothesis, clinical trials of perioperative cognitive behavioral therapy (CBT), delivered before and after surgery, have been effective in reducing disordered eating behaviors and promoting weight loss. 45ture studies should characterize the mediating role of adaptability to behavioral changes for patients with depression undergoing MBS.
The relationship between MBS and depression in the perioperative period is complex.MBS can lead to sustained improvements in depressive symptoms for nearly half of the patients with depressive symptoms prior to surgery, 46 likely through a renewed sense of motivation and autonomy over their health. 47,48However, not all patients respond in this way.About 15% of patients experience persistent depression after MBS, with some studies reporting even higher rates of suicidality than the general population. 49,50Whether this phenomenon is related to the relative drop-off in psychologic follow-up, 51 the high incidence of chronic pain after surgery related to gastric ulcers, 52 or the guilt of failing to adhere to strict lifestyle changes and experiencing relapse in metabolic dysfunctions, 53 there is a clear need for more longitudinal multidisciplinary support after surgery.
Our review also highlighted the paucity of studies reporting the relationship between preoperative depression and other important outcomes after MBS such as pain, opioid use, quality of life, and perioperative complications.The attenuating effect of preoperative depression on expected improvements in quality of life after MBS is of particular interest, as these improvements are an important holistic measure of surgical success to both patients and clinicians. 54 There are several limitations to consider in this systematic review.
First, the patients in the included studies were heterogeneous in demographics and types of MBS, and the studies themselves were conducted across different decades and countries.These differences could influence baseline patient morbidity, surgical techniques, and perioperative care programs.For instance, contemporary MBS programs now involve intense multidisciplinary optimization and followup before and after surgery, while the same may not have been offered to patients in earlier decades. 55Meta-regression techniques could be used to determine the impact of these factors, but a larger number of studies reporting similar outcomes are first needed.Second, the definitions and instruments used to measure depression before surgery varied widely.Even among studies that operationalized depression as a binary variable, some considered a pre-existing diagnosis of depression while others only required patients to meet a threshold score at a single point in time from one of many depressionscreening instruments.While majority of the instruments used have demonstrated good psychometric properties in various patient populations and settings, 56 there is a relative lack of evidence to support their use specifically in surgical or perioperative settings.The factors associated with depression in surgical patients may be different compared with non-surgical patients, which could affect the validity of these instruments in this setting. 15,57Third, many of the included studies had small sample sizes and limited power to detect meaningful differences in outcomes.Larger sample sizes may have allowed for analytic techniques such as matching methods to improve cohort bal-

| CONCLUSION
The impact of preoperative depression status or symptom severity on weight loss outcomes after MBS remains unclear.Well-designed and adequately powered studies using robust analytic methods are needed to precisely estimate the impact of preoperative depression on various outcomes after MBS.

AUTHOR CONTRIBUTIONS
An academic librarian developed the search strategies for the MED-LINE, EMBASE, Cochrane, and PsycINFO databases from their inception to May 25, 2021 (Appendices S1-S4).Given the long interval since the initial search, an updated search was conducted on June 30, 2023.An additional manual search was conducted from the reference list of included articles.

1
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram of systematic literature search, article screening, and review.T A B L E 1 Characteristics of included studies.
ance and generalized models to account for repeated outcome measures.Furthermore, other factors such as sociodemographic factors, comorbidities, the perioperative use of medications causing weight loss or weight gain (e.g., serotonergic antidepressants and antipsychotics), psychological factors (e.g., motivation), and lifestyle factors (e.g., exercise and diet) were not accounted for in most studies.The consideration of such variables as confounders or mediators in statistical analyses, if done so appropriately, would provide more robust estimates of the independent association of depression with postoperative outcomes.