Nasal outcomes
The postoperative Lund-Kennedy scores, SNOT-22 scores and TIBSIT scores were summarised in Fig. 2. Postoperative Lund-Kennedy score was significantly lower in the transseptal group than in the transnasal group (P < . 001). Post-hoc analysis revealed that the Lund-Kennedy scores were also lower at every timepoint after surgery (2.72 vs, 5.08 at 2 weeks, 0.88 vs. 3.52 at 1 month, 0.48 vs. 2.12 at 3 months, 0.04 vs. 1.76 at 6 months, all adjusted P < . 01). SNOT-22 showed also a lower postoperative scores in the transseptal group than in the transsnasal group at 6 months after surgery (4.40 vs. 7.36, P = . 005). Nevertheless, the TIBSIT scores were not significantly different between the two approaches (41.6 vs. 42.4 preoperatively, P = . 68; 41.6 vs. 40.2 postoperatively,P = . 46). The olfactory status based on normal values of different gender and age remained unchanged for all patients.
Table 3 showed the effect of transseptal approach on the postoperative Lund-Kennedy scores and SNOT-22 scores using regression models. The univariate model revealed a significant lower Lund-Kennedy scores and SNOT-22 scores in the transseptal group (β, -2.09; 95% confidence interval (CI), (-2.54, -1.64), P < .001 and β, -2.96; 95% CI, (-4.98, -0.93), P = .005, respectively). After controlling age, sex, Hardy grade, presence of regional invasion and endocrine function, transseptal approach remains an independent factor for lower Lund-Kennedy scores and SNOT-22 scores (β, -2.24; 95% CI, (-2.73, -1.75), P < .001 and β, -2.46; 95% CI, (-4.60, -0.93), P = .02, respectively).