BariPredict: Class II obesity, younger age and concurrent depression are risk factors for eating disorders in potential bariatric patients

Outcomes from the BariPredict cohort have found that 65.8% prevalence of persons at risk of an eating disorder (ED) - opting for, but not yet undergone, bariatric surgery - was higher than expected and at-risk patient profile is defined by class II obesity, younger age and concurrent depression, according to researchers from Kuwait. The researchers said that discovering that a specific obesity class is a risk factor, could indicate a critical window in terms of BMI. Therefore, they said this patient profile should be prioritised for psychological assessment and support to potentially improve outcomes of bariatric surgery.

The study’s authors established a cohort for screening and prospective follow up from bariatric surgery - BariPredict - a single-centre longitudinal cohort study of a population of persons in Kuwait that were referred for bariatric surgery, with data are collected in a hospital setting from a preintervention multidisciplinary team screening. The study was designed to identify preoperative baseline data from the BariPredict cohort to report on the prevalence and correlates of disordered eating risk in bariatric surgery candidates.

Of the 376 patients, 275 had complete baseline data and served as the subgroup on which cross-sectional analysis was performed. The analysis investigated prevalence of ED risk and association of multiple factors (psychological, physical).

The SCOFF questionnaire was self-administered by participants attending their first preoperative appointment. Psychological screening and assessment for anxiety and depression was carried out by a clinical psychologist using the Beck Depression inventory (BDI).

Outcomes

From the 275 participants 8% (n = 22) of participants were living with class I obesity, 34.9% (n = 96) with class II obesity and 57.1% (n = 157) with class III obesity. In terms of diabetes, 64.4% (n = 177) did not have diabetes, 20.4% (n = 56) had prediabetes and 15.3% (n = 42) had type 2 diabetes (T2D), 19.6% (n = 54) were hypertensive.

Psychological screening included eating disorder risk (SCOFF)12, anxiety (KUAS)16 and depression (BDI)15. Mean SCOFF score was 2.03 (SD = 1.12), approximately two thirds (65.8%; n = 181) had a SCOFF score of 2 or more indicating a likely ED. The mean KUAS score was 33.8 (SD = 11.8). Mild anxiety was the most common category, reported by 84.4% (n = 232) of participants. Moderate and severe anxiety affected 13.1% (n = 36) and 1.5% (n = 4), respectively. Mean depression score was 12.0 (SD = 7.7), with minimal or mild depression present in 45.5% (n = 125) and 40.7% (n = 112) of participants, respectively; and moderate or severe depression present in 9.8% (n = 27) and 4% (n = 11) of participants, respectively.

Compared to participants with no indication of an ED (SCOFF < 2), those with a likely ED (SCOFF ≥ 2) were of younger age (36.9 SD = 11.6 vs. 41.6 SD = 11.7 years; p < 0.01). BMI, gender, smoking status and class of obesity were not different between SCOFF categories.

Patients with T2D were more represented in the SCOFF < 2 group than the SCOFF ≥ 2 group (22.3% vs. 11.6%; p < 0.05). When those at risk and with prediabetes, are also considered, diabetes status remained near-significantly different between SCOFF categories (p = 0.057). Participants without diabetes nor prediabetes were more represented in the SCOFF < 2 group than the SCOFF ≥ 2 group (57.4% vs. 68%). Prediabetes status was not different between groups. HbA1c was higher in the SCOFF < 2 group than the SCOFF ≥ 2 group (6.1 SD = 1.5 vs. 5.7 SD = 1.1; p < 0.05).

The SCOFF < 2 group had a lower mean anxiety score than the SCOFF ≥ 2 group (31.5 SD = 7.5 vs. 35.0 SD = 10; p < 0.01). In anxiety category I, participants with mild anxiety predominated in both comparable groups, although those in the SCOFF < 2 group were more likely to have mild anxiety (p < 0.01). In anxiety category II, participants in the SCOFF ≥ 2 group were more likely to experience moderate and/or severe anxiety (p < 0.05).

Mean depression score was lower in the SCOFF < 2 group than SCOFF ≥ 2 group (9.7 SD = 5.8 vs. 13.2 SD = 8.4; p < 0.001). Grouping scores into categories: depression category I (minimal, mild, moderate or severe; p < 0.01) and depression category II (minimal/mild or moderate/severe; p < 0.001), participants with minimal depression and minimal/mild depression were more common in the SCOFF < 2 group than SCOFF ≥ 2 group (minimal: 56.4% vs. 39.8%; minimal/mild: 95.7% vs. 81.2%), and moderate and/or severe depression were less common in the SCOFF < 2 group than SCOFF ≥ 2 group (grouped categories: 4.3% vs. 18.8%).

Age had a near-significant negative association with SCOFF score (β = − 0.01, 95% CI − 0.02 to 0.00, p = 0.066). Gender, smoking status and BMI were not associated with SCOFF score. Class II obesity was positively associated with SCOFF score (β = 0.33, 95% CI 0.05–0.61, p < 0.05).

T2D was negatively associated with SCOFF score (β = − 0.42, 95% CI − 0.79 to − 0.05, p < 0.05). Hypertension was negatively associated with SCOFF score (β = − 0.40, 95% CI − 0.74 to − 0.07, p < 0.05). HbA1c was negatively associated with SCOFF score (β = − 0.11, 95% CI − 0.21 to − 0.00, p < 0.05). Other variables (normal or prediabetes status, total cholesterol, triglycerides) were not associated with SCOFF score.

Anxiety and depression scores were positively associated with SCOFF scores (respectively β = 0.02, 95% CI 0.01–0.03, p < 0.01 and β = 0.03, 95% CI 0.02–0.05, p < 0.001). Moderate/severe anxiety was also associated with SCOFF score (β = 0.50, 95% CI 0.121–0.873, p < 0.05).

Overall, class II obesity remained positively associated with SCOFF (β = 0.34, 95% CI 0.73– 0.61, p < 0.05). Hypertension remained negatively associated with SCOFF score (β = − 0.36, 95% CI − 0.69 to − 0.61, p < 0.05). Other variables in this category (normal or prediabetes status, total cholesterol, triglycerides and HbA1c) were not associated with SCOFF score following adjustment. Depression scores were positively associated with SCOFF scores (β = 0.03, 95% CI 0.01–0.05, p < 0.01). Anxiety scores were not associated with SCOFF score following adjustment.

“We will elaborate on these finds in the continuation and follow-up of the BariPredict cohort,” the authors concluded. “While our results indicate which patient profiles should be prioritized for psychological screening and support, future work will investigate temporal dynamics of ED risk and how these impact perioperative and long-term outcomes of bariatric surgery.”

The findings were reported in the paper, ‘Cross-sectional analysis of eating disorder risk and risk correlates in candidates for bariatric surgery from the BariPredict cohort’, published in Scientific Reports. To access this paper, please click here