周末医疗脉搏：减重不止于卡路里？GLP-1s与癌症风险

Woody Polglase+19173448314 woody.polglase@bernsteinsg.comSusannah Ludwig+41582723127 susannah.ludwig@bernsteinsg.comRebecca Liang, Ph.D.+85221232656rebecca.liang@bernsteinsg.comMiki Sogi, Ph.D.+81367776991 mikisogi@bernsteinsg.comJustin Smith+442077625899 justin.smith@bernsteinsg.comLee Hambright+19173448429 lee.hambright@bernsteinsg.comWilliam Pickering, MD+1917344 8340 wiliam.pickering@bernsteinsg.comLance Wilkes+19173448501 lance.wilkes@bernsteinsg.comDelphine Le Louet+33142139293 delphine.le-louet@bernsteinsg.com a dozen tumor types, survival outcomes on immunotherapy andCDK4/6inhibitors,hematologicmalignancies,cardioprotectionduring anthracycline chemotherapy, and BCG response in bladdercancer.AscO itselfappears tohavetakennote-a dedicatedsegment,with three Oral Presentations on GLP-1s and cancer, anda discussant session ("Weighingthe Evidence: GLP-1 ReceptorAgonists and Cancer Risk"),in the Prevention, Risk Reduction andGenetics track.This is perhaps asignalthat thefield has movedfrom a fringe hypothesis to a mainstream research agenda. and presentations referencing GLP-1s.Most of these wereobservational studies,using electronic health recordsdatabases,howevertheir results were prettystriking.Across numerousstudies,hazardratiosthatdescribed>30%reduction incancerincidence. Whilst the data is informative, it is unfortuantely not asclear-cut as the seductive HR's may lead one to believe. Firstly,retrospective observational studies are only hypothesis generatingand cannotbeusedtodetermine causality.Secondarily,thereare a number of design limitations with observational trials thatcan hinder interpretability,which we discuss below.This is nottosay that GLP-1s do nothing for cancer; there is a growing bodyof evidenceto suggestboth weight-dependentand independentmechanism by which GLP-1s impact cancer (although mostresearch is still preclinical).Ultimately,we need prospective,randomised, controlled, blinded, clinical trials to assess thesewonderdrugs'abilityto impact cancer diagnoses and progression.And the first - INSPIRE - a multisite international clinical trialevaluatingwhether GLP-1receptoragonists canlowerbreastcancer incidence in high-risk women, might be close to starting. WHATWASSEENATASCO? ASIZEABLEINCREASEIN(OBSERVATIONAL)STUDIES The volume of GLP-1 related work at this year's ASCO annualmeeting seems to be a signal in itself.A search of the ASCOabstracts databasereturns 44GLP-1and cancer-relatedpresentationsfromthe2026Annual Meeting,upfrom13atASCO2025,four at ASC0 2024,and one in the years prior.The 2026programme also reflects a meaningful qualitative shift: whereearlier abstracts werelargely confinedtoweight managementin breast cancer survivors orisolated single-tumorincidence control for confounders. between the two arms that are central to interpretingthe results.The GLP-1-exposed group carried a substantially higher diseaseburden: 21.3% had a Charlson comorbidity score of 3-4 (high)versus 13.6% in the unexposed group,and 19.7% scored 5+(very high) versus 10.4%. Heart disease (CHF/MI) was present in13.0% of GLP-1 users versus 7.3% of controls, renal disease in13.9% versus 7.1%,and liver disease was also more prevalent.Most strikingly, 53.4% of GLP-1-exposed women had Type 2diabetes comparedto just17.6% in the unexposedarm-a near-threefold difference that reflects the drugs'primary indicated use.The GLP-1 group also skewed toward higher BMI classes,with26.9% inObeseClass2(BMI35-40)and39.8%inObeseClass 3(BMI>=40)versus 15.7% and 12.7% respectively in controls. ordinarily predict worse cancer outcomes -GLP-1 exposure wasassociatedwitha~30%reduction inbreastcancerincidence.Breastcancerwas diagnosed in2.5%of thenon-exposedgroupversus 1.6% of GLP-1-exposed women, with true positive imagingrates of 2.4% versus 1.6% respectively.The authors were carefulto frame these findings as hypothesis-generating, explicitlynotingthattheobservational design cannotdemonstrate causalityand that the sicker, heavier baseline profile of GLP-1 users addsfurther complexity to causal inference. Indeed, the fact that aprotective signal emerged despite the GLP-1 cohort carryingsubstantiallygreater comorbidity burden-more diabetes,moreheart disease, morerenal disease, and higher BMI -makes the 30%risk reduction arguably more striking, and strengthens the case thatthe association warrants rigorous prospective investigation. THEORALSESSIONON GLP-1'SINCANCERSession#1-BreastCancer-Association ofGLP-1agonistswith breast cancer incidence in women. This session presented data from Dr.Elizabeth McDonald(PennMedicine/Abramson CancerCenter)on aretrospectiveobservationalstudy,andalsopublished inJcOOncologyPractice(D0l:10.1200/0P-26-00485)examiningwhetherGLP-1receptoragonists reduce breast cancer incidence. The study was motivatedbythree convergingfacts:breast canceris themost commoncancerinwomenworldwide,excessweightisoneofthemostimportant modifiable riskfactors for the disease, and GLP-1agonists are now in widespread use. Th