Dosing fluorouracil by blood level rather than body surface area halved severe toxicity

Design

Systematic review and meta-analysis of comparative randomised and non-randomised studies, random-effects models, RoB 2 and ROBINS-I, GRADE

Population

Five studies, 809 unique patients with colorectal cancer, predominantly metastatic disease on infusional fluorouracil within FOLFOX or FOLFIRI

Primary outcome

Severe or grade 3 and above toxicity with pharmacokinetically guided versus body surface area dosing

Effect

Severe toxicity RR 0.50 (95% CI 0.33-0.76, p=0.001, I-squared 79%); severe diarrhoea 0.33 (0.18-0.62); objective response 1.50 (1.24-1.80); disease control 1.18 (1.07-1.30)

Fluorouracil has been dosed by body surface area for fifty years, and body surface area predicts fluorouracil exposure poorly - interpatient variability in clearance is several-fold, which means some patients are systematically underdosed and others systematically poisoned. Pharmacokinetically guided dosing, adjusting to a measured area under the curve, has existed for decades and is barely used. This meta-analysis assembled the five comparative studies, randomised and non-randomised, covering 809 patients with colorectal cancer.

Pharmacokinetically guided dosing halved severe toxicity: risk ratio 0.50 (95% CI 0.33-0.76, p=0.001), though with substantial heterogeneity (I-squared 79%). Severe diarrhoea fell most clearly (RR 0.33, 0.18-0.62). Mucositis, neutropenia and hand-foot syndrome showed no significant difference. Objective response was higher (RR 1.50, 1.24-1.80) as was disease control (1.18, 1.07-1.30).

The combination is what makes this compelling: less toxicity and more response, which is what you would predict if body surface area dosing is simultaneously overdosing some patients and underdosing others. That is a coherent mechanism rather than a fluke of two separate analyses.

The evidence is nonetheless thin - five clinically heterogeneous studies, and no demonstrated progression-free or overall survival benefit, which the authors state plainly. The findings apply to infusional fluorouracil within FOLFOX or FOLFIRI in metastatic disease, not to bolus regimens or capecitabine.

Where an assay exists, this is a strong argument for using it in patients on infusional fluorouracil, and a stronger one for the units that already own the equipment and do not use it. Where it does not, the argument is for making the case to acquire it - the intervention costs a blood test per cycle and the toxicity it prevents costs admissions.

• Use pharmacokinetic dose adjustment for infusional fluorouracil where the assay is available
• This applies to infusional 5-FU in FOLFOX or FOLFIRI, not to bolus regimens or capecitabine
• Expect the largest reduction in severe diarrhoea; other toxicities did not differ significantly
• Do not claim a survival benefit - it was not established
• Combine with DPYD screening rather than treating one as a substitute for the other

The statistics, in plain English

A risk ratio of 0.50 for severe toxicity means half as many patients affected, but I-squared of 79% says the five studies disagreed considerably about the size of that effect - so the direction is more trustworthy than the number. One study reported toxicity per treatment cycle rather than per patient, which inflates the denominator and mixes two different units of analysis. Objective response, at RR 1.50, is a surrogate: more tumours shrank, and the analysis could not show that patients lived longer or progressed later, which is the outcome that would settle it.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.