The Pharmacy Line: What a Regain Predictor Actually Measures

A 242-patient sleeve cohort found first-year psychotropic use raised the hazard of weight regain by 79%. Two other cohorts point the other way, and the outcome's own threshold was only settled in 2026.

weight lossmental healthsleeve gastrectomyresearch

A three-item checklist that sorts sleeve patients into four groups is not, on its own, remarkable. What makes the one published this week in Obesity Surgery worth reading is the middle item. Among 242 patients operated at a single Taiwanese hospital, the most actionable of the three predictors of regaining more than a fifth of the weight lost in the first year was whether a psychotropic medication appeared in the chart during that same year — a hazard ratio of 1.79, against 0.96 for each point of twelve-month total weight loss.

The cohort deserves stating precisely. Median follow-up ran 37 months, and 89 patients — 36.8% — met the study’s definition of recurrent weight gain. Those who regained had lost less by six months (26.9% versus 29.2% of body weight) and by twelve (31.5% versus 35.1%). Preoperative triglycerides pointed the wrong way: lower values, higher risk, reported without much explanation. Then the count. Zero, one, two or three factors present gave regain rates of 0%, 20.7%, 44.4% and 63.2%. Discrimination was modest — AUC 0.700, and 0.60 after bootstrap correction — and the authors call the medication item a marker of “psychological or behavioral drivers” rather than a cause.

That is where the piece gets interesting, because two other cohorts have measured something close to the same variable and come back with the opposite sign.

The depression that helps and the treatment that protects

Nine hundred and thirty patients at one US institution, operated between 2018 and 2024, had PHQ-9 scores recorded at their standard preoperative psychological evaluation. The 128 with a depression diagnosis lost more excess weight than those without: 56.9% versus 52.0% at one year, 59.0% versus 50.4% at two, no difference after that. Within the depressed group, milder scores did better than severe ones.

The sharper result sits in the untreated subgroup. Depressed patients not taking antidepressants had overall complication rates of 11.5% against 0%, and 30-day readmissions of 15.4% against 3.0%, compared with depressed patients on therapy. Those cells are small enough that a single event moves a percentage, but the direction reverses the intuition that psychiatric illness is the risk. Here the medication reads as protective.

A Dublin cohort of 315 patients splits the difference by drug class rather than diagnosis. Antidepressants were used by 22.2%, and overall use did not predict twelve-month weight loss (β −2.59, p=0.084). Only agents classed as medium risk for weight gain did. “Psychotropic” as a database field pools drugs that behave differently — and the pooled estimate is what most models, including this week’s, are built on. The same denominator problem dogs the field’s headline numbers, as we wrote when the sleeve-versus-bypass metric question last came up.

A threshold agreed only this year

There is a second reason to handle the 36.8% carefully. The Taiwanese study defines recurrence as more than 20% of the weight lost by twelve months. An international Delphi exercise published in Surgery for Obesity and Related Diseases this year, with 66 experts across five specialties, settled on a different line: recurrent weight gain as more than 25% of lost weight measured from nadir, endorsed by 70.3% of participants. Nonresponse became under 10% total weight loss at twelve months, at 73.0%.

Those numbers matter because the consensus was weakest exactly where it was needed. Quantitative threshold items drew agreement on 28.1% of propositions — lowest of the seven domains — and the authors conclude that this area “requires additional research.” Voting moved from 26.2% of items agreed in the first round to 40.9% in the second, on response rates of 54.5% and 57.6%.

The vacuum being filled is measurable. A systematic review of 119 papers from 2020 to 2023 found that 45% did not define weight recurrence at all, or did not separate partial response from nonresponse; among those that did, definitions varied widely. Only 2.5% proposed a structured management algorithm.

The label moves as well

One further complication for that hazard ratio. A Dutch record review of 163 patients taking lithium, tricyclics or clozapine found 27 of them — 16.6% — needed a psychotropic dose adjustment within the first postoperative year. Paired pre- and postoperative plasma concentrations existed for three patients. Three had psychiatric symptoms severe enough to require hospitalisation or intensive monitoring.

So “first-year psychotropic use” bundles at least three different things: a stable prescription for a drug that does not touch weight, a dose that was chasing altered absorption after the operation, and an untreated illness that changes eating. Nothing in the model separates them, which is how the same variable can raise a hazard in Kaohsiung and lower a complication rate in a US health system.

That is the honest shape of prediction here. The count is real and the arithmetic transparent; a corrected C-statistic of 0.60 is a fair warning that this is a prompt for a conversation at a follow-up visit, not a triage rule. Whether the patient returns for that visit is its own literature, and its own problem. The LABS cohort, funded by NIDDK and followed on a fixed schedule for years, remains the closest thing the field has to a design in which these numbers could be checked — the study archive sets out what was measured and what was not.

niddklabs.org is an independent archive. It is not affiliated with NIH or NIDDK, and nothing here is medical advice.

Sources

  1. Su YS et al. One-Year Weight Loss, Metabolic Factors, and Psychotropic Medications Identify Patients at Risk for Recurrent Weight Gain After Sleeve Gastrectomy. Obesity Surgery, 16 Nov 2026.
  2. Wills MV et al. International expert consensus on definitions and management of weight recurrence and suboptimal response after metabolic and bariatric surgery: a Delphi study. Surgery for Obesity and Related Diseases 2026;22(7):753-761.
  3. Majid SF et al. Proposed Definitions and Clinical Recommendations for the Management of Weight Recurrence, Partial Response, and Nonresponse Following Metabolic and Bariatric Surgery. Obesity Surgery 2025;35(12):5556-5567.
  4. Maric E et al. Impact of preoperative depression severity and antidepressant medication use on perioperative outcomes and long-term weight loss following primary bariatric surgery. Surgical Endoscopy 2026;40(8):7133-7142.
  5. Haugh S et al. Impact of antidepressant use on post-operative weight loss following bariatric surgery. Irish Journal of Medical Science 2024;193(2):903-908.
  6. Wever SAL et al. Prevalence of Psychotropic Drug Treatment Interventions in Patients Undergoing Metabolic and Bariatric Surgery: A Retrospective Study. Obesity Surgery 2026;36(8):4104-4111.
From the blog

New research notes