Does Amoxicillin Improve Outcomes in Patients with Purulent Rhinorrhea?

In this trial, as in daily practice, we did not know the precise diagnosis of included patients. Moreover, despite our frequent requests, participating FPs included only a minority of eligible patients. Concern might arise that only patients with mild disease were studied. We made 3 efforts to verify that the population was truly representative. First, we determined that the personal characteristics and severity of symptoms of patients of low-recruiting FPs (who tend to include patients with worse symptoms³⁸) were no different from those of patients included by high recruiters. Second, an analysis of questionnaires from all eligible but excluded patients over a 6-week period showed that included and excluded patients were very much alike. The analysis also showed that in only 3% of patients did the FP consider the subject too ill to be included. Third, the results obtained on plain radiography of the maxillary sinuses were in line with the imaging results of other family practice populations with clinical suspicion of rhinosinusitis.^11,19-21

With regard to the methodology, we wish to clarify certain choices. Amoxicillin was selected because it is recommended as the first-line drug for rhinosinusitis in several practice guidelines^39-41 and the sensitivity of respiratory pathogens to it was sufficient in our geographic area at the start of the trial.⁴²* To evaluate symptoms, we chose the 20 items of the SNOT-20 questionnaire (Table 1), an abbreviated version of the RSOM-31,²⁹ a disease-specific quality-of-life test for sinusitis. These 20 items include not only all classic rhinosinusitis symptoms but also a number of more subjective symptoms, such as sleep disturbances and reduced productivity, which may also severely inconvenience patients. Any beneficial effect of amoxicillin on these symptoms would be just as important as an effect on the classic sinusitis symptoms.

Outcome measures were mainly self-assessed by patients, since in this kind of pathology, for which subjective inconvenience is often greater than objective signs might indicate, the patient is in our view the best and only judge of symptom improvement. The main outcome measure, disappearance of perceived worst symptoms, was designed to take into account the heterogeneity of clinical presentations.

Conclusions

Patients with an acute upper respiratory tract infection with purulent rhinorrhea (and without signs of complications of sinusitis) represent a large, clearly defined, clinically recognizable group. Our results show that amoxicillin provides no clinically important benefits for this population. The implication for practice is that whatever diagnosis is suspected, all these patients can safely be treated with symptomatic therapy only. Patients should, however, be informed that whichever treatment is chosen, symptoms can last for a long time. In the rare event that symptoms worsen, they should consult their FP for antibiotic therapy. If patients are clearly distressed by the purulent rhinorrhea itself, this trial suggests reasons for considering the use of amoxicillin, but potential patient benefits still probably do not outweigh the disadvantages.

* For an expanded version of this table, see Table W1.

ACKNOWLEDGMENTS

The authors wish to thank all participating family physicians and patients and Erna Eeckhout, Adrienne Dubron, Anselme Derese, MD, PhD, and John Marshall for their invaluable help.