Does Infusion Time of Day Change Lung Cancer Outcomes? A Retracted Trial Raises Doubts

Circadian rhythms govern immune cell trafficking and drug metabolism, so the time of day a cancer drug is given could be a free, modifiable factor in outcomes. Retrospective studies in melanoma, lung, kidney, and esophageal cancers have linked earlier immunotherapy infusions to better survival, but results conflict: a CheckMate 238 post hoc analysis and an eight-trial lung cancer meta-analysis (i-TIMES, n=1,550) found no meaningful difference. Retrospective work is also vulnerable to data-derived cutoffs and socioeconomic confounding.

This commentary evaluates LungTIME-C01, a single-center Chinese phase III trial that randomized 210 treatment-naive patients with stage IIIC–IV NSCLC to receive the first four cycles of anti–PD-1 chemoimmunotherapy before or after 3:00 pm. It reported progression-free survival of 11.3 vs 5.7 months (HR 0.40), overall survival of 28 vs 16.8 months (HR 0.42), and response rates of 69.5% vs 56.2%, favoring early dosing. The authors argue these effects exceed the benefit of adding immunotherapy to chemotherapy, which makes verification essential.

The commentary identifies several concerns. The registered protocol changed repeatedly: endpoints shifted, and the timing threshold moved from 6 pm vs 6 am, to after 9 am vs after 3 pm, to before vs after 3 pm roughly a year before data cutoff. These changes were not explained in the paper. No required number of PFS events was specified. The trial reported no treatment discontinuations for adverse events, whereas KEYNOTE-189 and KEYNOTE-407 saw about 13–14%. No crossover between arms was reported, and the protocol indicated patients missing scheduled times should be withdrawn, which would undermine an intention-to-treat analysis.

Mechanistically, the authors say the effect may not be due to immunotherapy timing alone. PD-1 antibodies have half-lives of weeks and sustained receptor occupancy, so a few hours' shift is unlikely to change steady-state exposure unless only the first dose matters. Carboplatin, nab-paclitaxel, and pemetrexed have short half-lives and diurnal variation in metabolism, DNA damage response, and hematopoiesis. Leukopenia was more frequent in the early arm, consistent with a chemotherapy chronotoxicity signal. Late infusions running to 8:39 pm may also have affected staffing, monitoring, and dose delivery.

The commentary recommends factorial trials that randomize chemotherapy and immunotherapy timing separately, tests of cycle 1 day 1 timing alone, a PD-L1 ≥50% anti–PD-1 monotherapy cohort, paired tumor and blood sampling, and multicenter validation. A note added in proof states that the Nature Medicine paper was retracted because editors no longer had confidence in the results, so it should not guide practice. Whether infusion timing matters remains unresolved.