Low-Dose Macrolide Therapy in CRS: A Negative Large Trial Against a Positive Subgroup Signal
Two real trials that disagree, and a patient whose own lab values sit inside the one that found a benefit.
Odalys R., a 46-year-old mother of two teenagers who spends most weekends coaching her younger daughter's travel soccer team, has had chronic sinus congestion, facial pressure, and thick postnasal drainage for over a year, without nasal polyps on endoscopy and with a SNOT-22 score of 54 despite three months of maximized intranasal corticosteroid therapy and twice-daily saline irrigation, both confirmed adherent. CT imaging shows diffuse mucosal thickening across her maxillary and ethmoid sinuses without polypoid change. Her total serum IgE, checked as part of the allergist's workup, came back at 12 IU/mL — low enough to be notable — and she has no personal or family history of atopic disease.
Her phenotype sits at the center of a disagreement between the only two randomized trials in this space, and the disagreement is narrower than it first looks. The later of the two, the MACS trial (Videler et al., 2011), gave weekly azithromycin to 60 patients and found no significant benefit over placebo on SNOT-22 — the result most often cited to move specialists away from macrolides. But MACS enrolled recalcitrant disease with and without polyps, 92% of it already surgically treated, its own authors raised both under-dosing and under-powering as explanations, and it never stratified for IgE at all. The earlier trial (Wallwork et al., 2006) gave daily roxithromycin to 64 polyp-free CRS patients and did stratify, finding the benefit concentrated in the low-IgE subgroup — and her IgE of 12 IU/mL sits far inside that split, as does her polyp-free, non-eosinophilic disease. The trial that found the signal is the one whose entry criteria she actually meets; the trial that missed it is the one that never looked for her.
She has no other chronic medical conditions, doesn't smoke, and works in a role — dental hygienist — with regular occupational exposure to aerosolized particulates that she and her allergist have both wondered, without any way to confirm it, might have some bearing on her chronic mucosal symptoms. Macrolides like azithromycin are proposed to work in CRS not through their antimicrobial activity at the low, immunomodulatory doses used for this indication, but through a separate anti-inflammatory effect on neutrophil function and mucus secretion — a mechanism distinct from, and not dependent on, any active bacterial infection being present, which is part of why the drug is being considered for her at all despite no evidence of ongoing bacterial superinfection on today's exam.
Reading two trials that disagree, in the patient they disagree about
The most recent randomized evidence we have is the MACS trial — Videler's group, azithromycin against placebo, and it found no significant benefit. I don't think we should offer a months-long antibiotic course, with real resistance and adverse-effect exposure, on the strength of an older trial's post-hoc subgroup split.
MACS didn't contradict Wallwork's finding so much as fail to look for it — it pooled polyp and non-polyp disease, never stratified by IgE, and at 60 patients was actually the smaller of the two trials. A benefit confined to low-IgE, polyp-free patients washes straight out of an aggregate like that.
Her IgE is 12. Her eosinophils are normal. She has no polyps. That is three-for-three on Wallwork's entry and stratification criteria, and nowhere near the recalcitrant, mostly post-surgical population MACS actually studied — she isn't a generic CRS patient being asked to bet on an aggregate null result.
I don't think either trial should be thrown out to favor the other — they're both real, and they plausibly disagree for a real, phenotype-based reason rather than one being simply wrong.
Rather than deciding in the abstract which trial to trust more, test the hypothesis directly in her — but name the gap honestly first: Wallwork's signal was roxithromycin given daily, and what we would actually prescribe here is azithromycin, which is the drug MACS used and found nothing with. We are extrapolating across agent and schedule, not only across populations. That argues for a defined 8-week course with SNOT-22 reassessment and a pre-specified stopping rule, not an open-ended commitment on faith that her phenotype reproduces Wallwork's result through a different macrolide.
Agreed: an 8-week, phenotype-motivated trial of low-dose azithromycin added to her existing regimen, with SNOT-22 reassessment and an explicit stopping rule rather than an open-ended course either way.
Genuinely unresolved, and named as such rather than papered over: whether her low-IgE phenotype will actually reproduce Wallwork's finding through a different macrolide on a different schedule, or whether MACS's null result reflects something true about macrolide therapy in CRS more broadly that an earlier stratified trial mistook for a subgroup effect. The 8-week reassessment, not today's discussion, is what will settle it for her specifically — the literature itself remains unsettled either way.