Residual late risk of severe adverse events in metabolic bariatric surgery patients after discharged emergency department visit.

BACKGROUND: When a patient returns with an acute problem after metabolic bariatric surgery, the emergency physician usually decides whether to admit or discharge the patient and which dangerous diagnoses to pursue. That decision is made with little operative detail and no clear sense of what is risky at a given point after surgery. National data have not helped. The major bariatric registry captured outcomes only through 30 postoperative days, and claims and single-center series miss out-of-network visits and rarely attach a reason to a visit. OBJECTIVES: To address four practical emergency questions: how often a late serious adverse event (late SAE) was preceded by a discharged emergency department (ED) visit; how much risk remains after such a visit ends in discharge; what the likely diagnoses are by procedure and time since surgery; and whether an exploratory risk model built from routinely available information separates higher-risk from lower-risk groups at the end of the first postoperative month. METHODS: This retrospective national cohort included 180,540 adults who underwent initial sleeve gastrectomy (SG) or Roux-en-Y gastric bypass (RYGB) in 2023, using the first registry release with structured follow-up beyond 30 days. The exposure was the highest-order acute-care contact within 30 days, ranked reoperation, readmission, discharged ED visit only, or none. The outcome was any serious event from day 31 to day 273. Models were adjusted for thirteen prespecified covariates. SG and RYGB were analyzed separately. All estimates identify groups at differing risk rather than individual patients. RESULTS: Among patients with a late SAE, 21.5% of SG and 23.3% of RYGB patients had a prior discharged ED visit within 30 days; these are lower bound estimates. A discharged ED visit was followed by a late serious adverse event in 2.8% of SG and 9.2% of RYGB patients, against 1.0% and 4.0% with no 30-day contact (adjusted odds ratio 2.8 and 2.4). Defining the exposure as any discharged ED visit, irrespective of later escalation, gave a larger association (3.53 and 2.71), and all four outcome components pointed the same way. The number needed to follow for one excess late SAE was 55 for SG and 19 for RYGB. Likely diagnoses split by procedure and changed over time, with nonspecific diagnoses followed by reflux-related diagnoses after SG, and anastomotic stricture followed by marginal ulcer and then internal hernia after RYGB. An exploratory risk model anchored at day 30 was well calibrated but discriminated modestly (area under the curve 0.70 and 0.63), and its sensitivity was low, so it does not identify which patients will have an event and would require external validation before any clinical use. CONCLUSIONS: A discharged ED visit after metabolic bariatric surgery identifies a group that stays at higher risk for months, and the later diagnoses follow a procedure- and time-specific order. Routinely available information gives a calibrated estimate of that risk at the end of the first postoperative month but does not predict the course of an individual patient. These associations support evaluating structured follow-up after an early ED visit.