Baseline adipokines failed to forecast sleeve gastrectomy weight loss with any consistency
Routine preoperative adipokine profiling did not provide a dependable way to predict who would lose the most weight after laparoscopic sleeve gastrectomy. In a prospective longitudinal cohort of 83 adults with severe obesity, baseline circulating adipokines offered only scattered signals, while a standard lipid marker, HDL cholesterol, tracked postoperative weight loss more reliably across follow-up.
The question is clinically important because bariatric surgery remains the most effective treatment for severe obesity, yet the postoperative course varies from patient to patient. Some individuals lose weight rapidly and continue to improve for months, while others plateau earlier or lose less than expected. Adipokines seemed biologically plausible as predictors because they sit at the intersection of energy homeostasis, inflammation, and insulin sensitivity. In theory, a baseline hormonal profile drawn from adipose tissue might reveal a patient’s metabolic “set point” before the first stapler line is even placed.
This cohort tested that idea in a practical way. Investigators measured clinical and metabolic parameters together with eight circulating adipokines before surgery and over 12 months of follow-up. They then examined associations with postoperative weight-loss magnitude, weight-loss velocity, and the timing of weight-loss plateau using multivariable regression and survival analyses, while applying false-discovery-rate correction and bootstrap validation. The analytical framework mattered because weak biomarker studies often generate false leads when multiple candidates are tested at once.
The postoperative biology was robust. LSG produced substantial improvements in anthropometric and metabolic measures, alongside marked remodeling of circulating adipokines. In other words, the surgery clearly altered the hormonal environment linked to appetite regulation, inflammation, and insulin action. Yet those baseline hormone levels did not translate into stable prediction of who would ultimately lose more weight.
One signal did emerge early. Higher baseline adiponectin was independently associated with lower percentage total weight loss at 6 months. After that, the association lost its clinical traction. No adipokine consistently predicted long-term weight loss or the onset of weight-loss plateau once multiple testing was accounted for. That pattern is a cautionary one: a biomarker may look promising in a single time point analysis, only to lose relevance when the full longitudinal picture is examined.
The inconsistency probably reflects the biology of obesity itself. Adipokines are not isolated markers; they are downstream readouts of adipose tissue function, systemic inflammation, hepatic metabolism, insulin sensitivity, and perhaps even behavior-related factors that influence eating after surgery. A single baseline measurement may capture an unstable metabolic snapshot rather than a durable driver of postoperative response. Sleeve gastrectomy also changes the hormonal milieu so profoundly that preoperative circulating levels may be overwhelmed by postoperative physiology within weeks.
HDL outperformed adipokines as a preoperative signal
A more familiar metabolic marker proved stronger. Baseline HDL cholesterol showed a stable inverse association with postoperative weight loss across multiple follow-up time points and outperformed every adipokine tested as a preoperative predictor. In practical terms, higher HDL at baseline tracked with less postoperative weight loss, while lower HDL aligned with greater loss across the observed intervals.
This finding is not surprising on biological grounds. HDL often reflects broader metabolic health, hepatic lipid handling, insulin resistance, and the degree of cardiometabolic dysregulation that accompanies severe obesity. Unlike adipokines, which can fluctuate with adipose tissue signaling and inflammatory tone, HDL may serve as a more integrated marker of the systemic metabolic state entering surgery. It is not a direct predictor of surgical physiology, but it may be a better proxy for the terrain on which that physiology acts.
The study also speaks to the limits of precision prediction in bariatric care. Preoperative counseling often tries to convert complex biology into a simple estimate of postoperative weight loss. This cohort suggests that expensive biomarker panels may not yet improve on conventional clinical chemistry. For now, routine adipokine profiling appears to add little beyond established markers already available in standard practice.
A direct bedside implication follows from the data. If the goal is to identify patients at risk for suboptimal weight-loss response after sleeve gastrectomy, the most useful signal may still come from common metabolic measures rather than specialized endocrine assays. The appeal of personalized biomarker prediction is easy to understand. The data here show how hard it is to turn that appeal into reliable forecasting.
What the 12-month cohort can and cannot support
The study design strengthens the message. Prospective follow-up reduces some of the bias that can distort retrospective biomarker work, and the use of multivariable modeling with correction for multiple comparisons adds statistical discipline. Even so, the cohort was modest in size, with 83 patients, so smaller effects could have been missed. The report also assessed a specific surgical procedure, laparoscopic sleeve gastrectomy, which means the findings may not extend automatically to other bariatric operations with different hormonal and anatomic consequences.
The broader clinical takeaway is restrained but useful. Baseline adipokine measurement did not deliver a dependable preoperative map of postoperative weight-loss success, despite strong biological plausibility and clear postoperative hormonal remodeling. HDL cholesterol, a conventional marker, carried more consistent predictive value than any adipokine in the panel.
For clinicians, this means the hunt for a single preoperative blood test that can forecast weight-loss response remains unfinished. For researchers, it means the next useful biomarker may need to integrate multiple metabolic domains rather than rely on one adipose-derived signal alone. A future model may need to combine hormonal, lipid, inflammatory, and clinical variables if it is to outperform the simple measures already available in the clinic.