2018 Section 5 - Rhinology and Allergic Disorders

Sukato et al

Figure 1. PRISMA flowchart demonstrating the process of article selection. AHI, Apnea-Hypopnea Index; ESS, Epworth Sleepiness Scale; PSQI, Pittsburgh Sleep Quality Index; sdRSDI, sleep domain of Rhinosinusitis Disability Index; sdSNOT-22, sleep domain of Sino-Nasal Outcome Test–22.

The methodological quality of each study was evaluated with the MINORS criteria ( Tables 1 and 2 ). The MINORS scores ranged from 11 to 15, with 16 being a perfect score. 34 The mean and median MINORS scores were 12.6 and 12, respectively. Studies with scores . 11 were considered to have low risk of bias. Most studies were deficient in cate- gories of power calculation and blinded evaluations. The majority of these studies also lost . 5% of their sample size at follow-up. All of the studies had clearly stated aims and consecutive sampling. Given that the studies aimed to evalu- ate QOL and sleep quality, all had validated end points appropriate to their objectives. Two studies were case series with chart reviews or were retrospective in nature, 16,17 but all studies involved prospective collection of data. Epworth Sleepiness Scale Three studies 13-15 evaluated sleep outcomes with the ESS ( Table 3 ). A forest plot is displayed in Figure 2 , which demonstrates a random effects meta-analysis for the change in ESS scores after surgery. An SMD of 2 0.94 (diamond) was obtained for the pooled result, which indicated a large and statis- tically significant ( P = .007) improvement. However, the clinical

significance may be lower because the broad 95% CI was not able to rule out a small effect size (SMD, 2 0.26). The I 2 of 99% was consistent with high heterogeneity among studies. Two studies 14,15 included patients with nasal polyposis, while 1 study 13 specifically evaluated patients without nasal polyps. A study had to be excluded from the meta-analysis due to the use of median and interquartile range, as opposed to mean and standard deviation. 18 Pittsburgh Sleep Quality Index The PSQI was utilized as the outcome instrument in 2 stud- ies 12,13 ( Table 3 ). The forest plot shown in Figure 3 demonstrates a random effects meta-analysis for the change in the PSQI. The SMD of the pooled result was 2 0.80 (dia- mond), which was consistent with a large and statistically significant ( P = .017) improvement. The clinical signifi- cance was uncertain because the broad 95% CI could rule out a trivial effect size (SMD, 2 0.14). The I 2 of 99% indi- cated high heterogeneity among studies. Rotenberg and Pang 13 excluded patients with nasal poly- posis, while Alt et al 12 included patients suffering from nasal polyposis.

135