Grade 3 adverse events on the FLAURA2 combination fell from 43% during induction to 14% on osimertinib alone

Grade 3 or higher events fell from 43% in induction to 14% on osimertinib alone

The analysis counts the onset of new adverse events causally related to treatment within each of the three periods. During induction, when patients were receiving osimertinib, platinum and pemetrexed together, 93% had a treatment-related event of any grade and 43% had one of grade 3 or higher. During pemetrexed maintenance the any-grade figure was 90% and the grade 3 or higher figure 25%. During the osimertinib-only period the two figures were 58% and 14%.

Those percentages rest on a re-cut of the combination arm by treatment exposure rather than by randomisation. Induction covered 276 patients, pemetrexed maintenance 201, and the osimertinib-only period 206. The abstract does not account for the difference between those denominators, so the number of patients who left the regimen before reaching a later period, and the reasons they left, are not established by the text reported here.

The measure itself is worth reading carefully. The abstract reports onset frequency, which describes how often fresh toxicity appeared within each period. It is not a measure of how many patients were living with toxicity at any given moment, and the abstract reports no such figure for any of the three periods.

Renal toxicity was the only reported category that rose during pemetrexed maintenance

Adverse events indicating renal toxicity were reported in 19% of patients during pemetrexed maintenance, against 7% during induction. The maintenance events were all grade 1 or 2, and the induction period included one grade 3 event. Pemetrexed is the agent that continues through maintenance after the platinum has stopped, and it is the only one of the toxicity categories named in the abstract that was more frequent later in treatment than at the start.

The abstract reports the frequency and the grades and nothing behind them. There are no creatinine or creatinine-clearance values, no count of dose reductions or interruptions attributed to renal events, and no statement of how many of these events resolved. Whether the higher maintenance figure reflects cumulative pemetrexed exposure or the longer observation window of that period is not addressed in the text reported here.

Maintenance ran a median of 12.4 months and the osimertinib-only period 17.8 months

The three periods are of very different length. Induction ran a median of 2.8 months, with a range of 0.1 to 4.1 months. Pemetrexed maintenance ran a median of 12.4 months, range 0.7 to 56.1. The osimertinib-only period ran a median of 17.8 months, range 0.4 to 56.8. The authors present the length of the two later periods as a finding in its own right, noting that they exceeded 12 and 17 months respectively.

That difference in length matters for reading the percentages above. An onset frequency measured over a median of 2.8 months and one measured over a median of 17.8 months are not observed for the same amount of time per patient, and the abstract reports the figures as period frequencies rather than as exposure-adjusted rates. The decline from 43% to 14% is therefore a comparison between periods of unequal duration, and the longer period is the one with the lower figure.

Discontinuation of osimertinib stayed at or below 8% across the three periods

Adverse events leading to discontinuation of osimertinib were reported at or below 8% in every period. New events indicating haematological, gastrointestinal and skin or nail toxicity were reported most frequently during induction and were mostly grade 1 or 2, with reductions in frequency in the periods that followed.

The design sets the limit on what this can show. The analysis is post hoc, it covers the combination arm alone, and its periods are defined by each patient’s own treatment exposure rather than by random assignment. It describes how the safety profile of the FLAURA2 regimen behaved over time in the patients who received it. It does not measure that profile against the osimertinib monotherapy arm, and the abstract reports no period-by-period comparator figures that would allow a reader to do so. The efficacy claim for the regimen rests on the parent trial, which the authors describe as having significantly improved survival with the addition of platinum–pemetrexed.