Results
The search rendered 4590 studies, 742 duplicates were removed, and 3442 were deemed ineligible after screening titles and abstracts. The reference review resulted in the addition of one article. The resulting 35 full texts were screened, of which twenty-three studies were selected for data extraction and analysis (Gandara et al., 1995, Somlo et al., 1995, Kemp et al., 1996, Madasu et al., 1997, Planting et al., 1999, Ekborn et al., 2004, Zuur et al., 2007, Yıldırım et al., 2010, Riga et al., 2013, Yoo et al., 2014, Marshak et al., 2014, Ishikawa et al., 2015, Crabb et al., 2017, Nasr et al., 2018, Delarestaghi et al., 2018, Rolland et al., 2019, Duinkerken et al., 2021, Fernandez et al., 2021, Moreno et al., 2022, Weijl et al., 2004, Villani et al., 2016, Scasso et al., 2017, Campbell et al., 2022). Figure 1 depicts the PRISMA complete screening process. Publication dates were 1995–2022, with studies conducted in 14 different countries, with 5 studies from the United States, 4 from the Netherlands, 2 from Canada and Italy, and one study from Sweden, Turkey, Greece, Spain, Israel, Japan, United Kingdom, Egypt, Iran, and India. Studies consisted of 18 controlled trials (Gandara et al., 1995, Somlo et al., 1995, Kemp et al., 1996, Planting et al. 1999, Zuur et al. 2007, Yıldırım et al. 2010, Riga et al. 2013, Yoo et al. 2014, Marshak et al. 2014, Crabb et al. 2017, Nasr et al. 2018, Delarestaghi et al. 2018, Rolland et al. 2019, Duinkerken et al. 2021, Moreno et al. 2022, Weijl et al. 2004, Villani et al. 2016, Campbell et al. 2022) and 5 quasi-experimental studies (Madasu et al. 1997, Ekborn et al. 2004, Ishikawa et al. 2015, Fernandez et al. 2021, Scasso et al. 2017). The median number of patients per study was 73 and ranged from 11 to 277. Of note, only four RCT had a low risk of bias, seven had some concern of bias, and seven had a high risk of bias. Across the 18 RCT, the most common sources of bias were related to the outcome measurement and the selection of results. In the quasi-experimental studies quality assessment, two studies were of high quality and three were rated as having poor methodological quality. The source of bias came from the comparability and outcome evaluations. Table 1 and Table 2 presents the quality assessment for all of the studies. In total 11 interventions were used for cisplatin-ototoxicity, 9 pharmacological interventions were assessed in 19 studies(Gandara et al., 1995, Somlo et al., 1995, Kemp et al., 1996, Madasu et al. 1997, Planting et al. 1999, Ekborn et al. 2004, Zuur et al. 2007, Yıldırım et al. 2010, Riga et al. 2013, Yoo et al. 2014, Marshak et al. 2014, Ishikawa et al. 2015, Crabb et al. 2017, Nasr et al. 2018, Delarestaghi et al. 2018, Rolland et al. 2019, Duinkerken et al. 2021, Fernandez et al. 2021, Moreno et al. 2022) and 2 non-pharmacological interventions assessed in 4 studies( Weijl et al. 2004, Villani et al. 2016, Scasso et al. 2017, Campbell et al. 2022). All of the studies assessed platinum-ototoxicity prevention, except for one that evaluated ototoxicity treatment (Nasr et al. 2018). Although we searched for platinum-induced ototoxicity, all studies assessed cisplatin and none of the studies included other platinum agents or other types of chemotherapy agents. All of the studies interpreted cisplatin-induced ototoxicity (CiO) outcome as hearing loss, five studies also considered tinnitus (Planting et al., 1999, Madasu et al., 1997, Ishikawa et al., 2015, Yoo et al., 2014, Scasso et al., 2017), and only two included vestibular disturbances(Madasu et al., 1997, Ishikawa et al., 2015). All of the studies used an audiometry test to examine ototoxicity. The study’s characteristics for the pharmacologic and nonpharmacologic interventions appear in Table 3 and Table 4, respectively.
Table 1. Assessment of the risk of bias in clinical trials.