The Ruler Matters: How Weight-Loss Metrics Can Flip Who Counts as a Responder

Two patients each lose 25 kilograms in the year after a sleeve gastrectomy. Same operation, same scale, same number of kilograms gone. Whether either of them is later described as a success — or quietly filed among the “suboptimal responders” — turns on a decision that has nothing to do with the surgery itself: which ruler you measure them with, and where that ruler starts.

That is the argument of a new analysis of 17,568 patients published in Obesity Surgery, the journal of the International Federation for the Surgery of Obesity and Metabolic Disorders. It is a retrospective analysis of a national quality registry, not a randomised trial, so it describes associations rather than cause and effect. But the paper’s point is methodological, and it arrives as the specialty is still trying to agree on a single way to report what weight-loss surgery actually achieves.

Two rulers, two traditions

The field has long spoken in two units.

Percent excess weight loss, or %EWL, measures how much of a patient’s excess weight above a reference figure has come off. It dominated the literature for decades, and it belongs to the era when surgery was reserved for patients with very high body mass index. Over time, a 50% EWL threshold hardened into an informal definition of success.

Percent total weight loss, or %TWL, is blunter: kilograms lost divided by starting weight. Because it needs no reference weight, it travels better across populations whose BMIs differ — which is why, as the Freiburg authors note, a 2024 IFSO consensus adopted a reporting threshold of less than 20% TWL to flag a suboptimal response, while the older <50% EWL bar survives mainly as a bridge to earlier studies.

Both are legitimate. They are not interchangeable.

What the registry showed

Researchers led by Mira Fink at the Medical Center of the University of Freiburg studied adults in the StuDoQ|MBE registry who had primary metabolic-bariatric surgery and complete two-year follow-up. They applied multivariable generalised linear mixed-effects models — statistical machinery built to separate the influence of one preoperative factor while holding the others steady.

Two preoperative variables were tested against response: starting BMI, and how much weight a patient had already shed before the operating day. Two thresholds were applied to the same outcomes: the historical <50% EWL line and the IFSO-consistent <20% TWL line.

The two verdicts did not agree. A higher starting BMI made a patient more likely to fall short of the older %EWL target — intuitive enough, since excess weight is the denominator. Switch to %TWL and the relationship reverses: a higher starting BMI was associated with a lower probability of being classified a suboptimal responder.

The same data, two defensible formulas, opposite conclusions about which patients struggle.

Where the ruler begins

The second finding is subtler and, for patients, more pointed. Response depends on the baseline as much as the formula.

The team calculated weight loss two ways: cumulatively, from the earliest documented weight before any preoperative weight-reduction effort, and postsurgically, from the weight recorded on the day of surgery. Preoperative weight loss then looked protective in one frame and harmful in the other. When the whole trajectory was counted from the start, losing weight before surgery reduced the odds of a suboptimal label. When only the surgical phase counted from the day of the operation, that same preoperative loss raised them — because the patient had already improved the number the second ruler starts from.

What the paper does not claim

The authors are unusually direct about the limits. The registry lacked sufficiently standardised data on improvement of obesity-related complications, so the clinical half of the IFSO definition could not be incorporated; the analysis is weight-based by necessity. Residual confounding and unmeasured differences between patients remain in any observational registry, and the cohort is German. A threshold, they write, should not be read as synonymous with clinical treatment failure, because clinical response after metabolic-bariatric surgery is multidimensional: glycaemic control, sleep apnoea, blood pressure, joint pain and quality of life all move on their own schedules.

That caution connects to a distinction the field has been drawing for years. In a prospective cohort such as the LABS study, funded by NIDDK — which followed thousands of patients for years rather than reconstructing them from billing and registry records — weight is only one of several outcomes tracked, precisely because the operation’s effects do not reduce to a percentage.

What to ask when you read the numbers

For anyone outside the operating theatre — including people weighing their own BMI against eligibility criteria — the usable lesson is a reading habit, not a rule.

When a headline declares that a procedure produced “poor weight loss” in some fraction of patients, two questions cut through the noise: which metric, and which baseline? A study reporting %TWL and one reporting %EWL can look as though they studied different operations. A report that measures from the day of surgery and one that measures from the first day of the preoperative diet will disagree about who did well, and neither is lying.

The 2022 ASMBS and IFSO indications update, which lowered the BMI threshold for surgery to 35 regardless of comorbidity, is a reminder that the entry criteria themselves were redrawn recently — and any percentage quoted today sits on top of those definitions. This archive is independent of NIH and NIDDK and does not recommend surgery for or against anyone; what the evidence supports here is narrower and more useful. Agreement on the ruler is the first thing a field needs before it can argue about the outcome.