Drug allergy labels come with a fifth more postoperative complications

Design

Prospective multicentre observational cohort study, 21 UK hospitals

Population

13,646 adults undergoing common elective and emergency surgical procedures; 3,924 (29%) with at least one drug allergy label

Primary outcome

Composite of postoperative complications within 30 days

Effect

25% vs 20%, OR 1.21 (95% CI 1.10-1.34); any postoperative infection OR 1.24 (1.11-1.38); allergic drug reactions OR 3.00 (1.77-5.09); no difference in mortality

Twenty-one UK hospitals prospectively followed 13,646 adults through hip or knee replacement, long bone fracture fixation, colorectal resection, transurethral resection of prostate or bladder tumour, caesarean delivery or hysterectomy. Patients who had taken antibiotics in the preceding fortnight were excluded. The primary outcome was a 30-day composite of infections, anastomotic leak, acute respiratory distress syndrome, myocardial infarction, bleeding, pulmonary embolism, stroke, antimicrobial side-effects and death.

Almost a third - 3,924 patients, 29% - carried at least one drug allergy label. They had more complications: 989 of 3,924 (25%) against 1,926 of 9,722 (20%), odds ratio 1.21 (95% CI 1.10-1.34). Surgical site infection was commoner (9% against 8%, OR 1.19, 95% CI 1.03-1.38), as was any postoperative infection (19% against 15%, OR 1.24, 95% CI 1.11-1.38). Allergic drug reactions occurred in 31 labelled patients against 29 unlabelled, an odds ratio of 3.00 (95% CI 1.77-5.09). Mortality did not differ.

The mechanism the authors point to is the obvious one: an estimated 90% of these labels are wrong, and the label makes the team reach for a second-choice antimicrobial that is less effective, more toxic or both. The action this supports is not clever. It is asking about the allergy properly at the pre-operative visit, correcting the record when the history does not support the label, and referring for formal assessment when it matters and the history is equivocal. A de-labelling pathway is one of the few perioperative interventions that is cheap, permanent and benefits every subsequent admission.

• Interrogate every allergy label at the pre-operative assessment rather than transcribing it forward
• Correct the record when the history does not support the label - a corrected label outlives this admission
• Where the label blocks first-choice prophylaxis, establish a route to formal allergy assessment
• Do not dismiss labels wholesale: labelled patients had three times the rate of allergic drug reactions
• Audit which antimicrobial prophylaxis your labelled patients actually receive - that is where the infection difference sits

Why it matters

A word copied forward through a patient's record is measurably changing what happens to them in theatre.

Don't overread it

Observational: the labels may partly be a marker of the kind of patient who accumulates them rather than a cause of the complications.

The statistics, in plain English

An odds ratio of 1.21 is modest, and in an observational study of this kind it could partly reflect the patients rather than the labels - people with more comorbidity accumulate more allergy labels over more healthcare contacts. The infection-specific findings are what make the mechanism plausible, because they are exactly what a second-choice antimicrobial would produce. The tripled rate of allergic drug reactions is a genuine and separate finding: labelled patients are also more likely to react, which is a reason to investigate labels rather than simply ignore them. Note that the percentages printed for that outcome in the abstract are inconsistent with the counts; the odds ratio and the raw numbers are the reliable figures.

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.