The Sniffin’ Stick identification subtest, a validated tool to evaluate the sense of smell, is based on the recognition of 16 different odours. The patient is required to choose an answer from among four proposed odours, which introduces the possibility of obtaining random correct answers, especially in patients with an altered sense of smell.

This study was designed to evaluate the influence of these random correct answers on interpretation of the simplified version of the Sniffin’ Stick test comprising threshold and identification tests in patients with nasal polyposis.

Forty-two consecutive patients with nasal polyposis operated according to the nasalization procedure were enrolled in this prospective study. Odour threshold and identification tests of the Sniffin’ Stick kit were performed before and 1 month after surgery. Random correct answers on the identification (I) test (I_H) were subtracted from the global number of correct answers (I_G) to calculate a real identification score (I_R), corresponding to the number of correct answers unrelated to chance.

Two groups of patients were identified: one group with no random correct answers (I_H0) (n=17) and another group giving 1 to 7 random correct answers (I_H1–7) (n=25). In the I_H1–7 group, significantly more patients had an immeasurable threshold (T=0) than a measurable threshold (21 versus 4, P=0.0001). In this subgroup of 21 patients [I_H1–7, T=0], the mean I_R score was significantly lower than the mean I_G score (P<0.0001) and 13 patients were classified as [I_R=0; T=0]. Among these 13 patients classified as severe anosmia [I_R=0; T=0] preoperatively, only 3 remained severe anosmic [I_R=0; T=0] postoperatively.

Random answers to the I identification test were more numerous among patients unable to detect n-Butanol on the T threshold test than among patients able to detect n-butanol. Calculation of the I_R identification score allows more precise interpretation of the results of the identification test in patients with severe anosmia.