Real-world data from the VA suggests abiraterone is associated with increased hospitalizations for cardiovascular events, infections, and acute kidney injury compared to other ARPIs. This finding prompts clinicians to favor AR antagonists like enzalutamide in older, comorbid patients at higher risk for these complications.

The study utilized "interruption-free survival" as a primary endpoint, a pragmatic measure derived from real-world data. This serves as a valuable surrogate for treatment toxicity, as clinicians typically pause treatment in response to adverse events, providing a quantifiable measure of a drug's real-world tolerability.

Current Quality of Life (QoL) assessments in cancer trials fail to capture severe, long-term toxicities. They are designed for short-term effects and data collection often ceases after a patient experiences a life-changing adverse event, thus painting an inaccurately rosy picture of a drug's tolerability.

The enzalutamide arms saw discontinuation rates of 20-25% due to adverse events. This high rate reflects a different risk calculation for patients who feel healthy and are asymptomatic. Unlike in advanced disease where patients tolerate more toxicity, this population has a very low threshold for side effects, making early intervention a significant trade-off.

When a clinical trial includes a major procedure like a cystectomy for all patients, its overwhelming effect on quality of life can completely obscure the subtler side effects of the drug being tested. The measurement tool ends up capturing the impact of the surgery, not the specific therapeutic intervention.

Unlike controlled clinical trial data, real-world evidence is derived from vast, messy, and incomplete data from daily healthcare. This variability is its strength, offering deeper insights into long-term outcomes, drug interactions, and diverse patient populations that clean trial data misses.

A real-world analysis of pemigatinib reported low rates of dose reduction or discontinuation. This may be misleading, as the toxicities of FGFR inhibitors (e.g., nail, skin, eye issues) are cumulative and worsen over extended periods. The study's shorter follow-up likely didn't capture the full long-term safety profile of the drug.

Registrational clinical trials for prostate cancer drugs enroll patients who are substantially younger (median age 67-69) and healthier than the typical real-world patient (median age 73-74). This gap means trial data on efficacy and safety doesn't perfectly apply to the majority of patients seen in clinic.

Tarlatamab is being administered to patients who would have been excluded from clinical trials (e.g., lower performance status, brain mets). This real-world population experiences potentially lower efficacy and different toxicity patterns, such as more frequent ICANS, than the pristine data from the drug's approval studies would suggest.