Jorrit W van As, Henk van den Berg, Elvira C van Dalen

Platinum-based therapy, including cisplatin, carboplatin, oxaliplatin or a combination of these, is used to treat a variety of paediatric malignancies. Unfortunately, one of the most important adverse effects is the occurrence of hearing loss or ototoxicity. There is a wide variation in the reported prevalence of platinum-induced ototoxicity and the associated risk factors. More insight into the prevalence of and risk factors for platinum-induced hearing loss is essential in order to develop less ototoxic treatment protocols for the future treatment of children with cancer and to develop adequate follow-up protocols for childhood cancer survivors treated with platinum-based therapy. To evaluate the existing evidence on the association between childhood cancer treatment including platinum analogues and the occurrence of hearing loss. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (2015, Issue 8), MEDLINE (PubMed) (1945 to 23 September 2015) and EMBASE (Ovid) (1980 to 23 September 2015). In addition, we searched reference lists of relevant articles and the conference proceedings of the International Society for Paediatric Oncology (2008 to 2014), the American Society of Pediatric Hematology/Oncology (2008 to 2015) and the International Conference on Long-Term Complications of Treatment of Children and Adolescents for Cancer (2010 to 2015). Experts in the field provided information on additional studies. All study designs, except case reports, case series (i.e. a description of non-consecutive participants) and studies including fewer than 100 participants treated with platinum-based therapy who had an ototoxicity assessment, examining the association between childhood cancer treatment including platinum analogues and the occurrence of hearing loss. Two review authors independently performed the study selection. One review author performed data extraction and risk of bias assessment, which was checked by another review author. We identified 13 eligible cohort studies including 2837 participants with a hearing test after treatment with a platinum analogue for different types of childhood cancers. All studies had methodological limitations, with regard to both internal (risk of bias) and external validity. Participants were treated with cisplatin, carboplatin or both, in varying doses. The reported prevalence of hearing loss varied considerably between 0% and 90.1%; none of the studies provided data on tinnitus. Three studies reported a prevalence of 0%, but none of these studies provided a definition for hearing loss and there might be substantial or even complete overlap in included participants between these three studies. When only studies that did provide a definition for hearing loss were included, the prevalence of hearing loss still varied widely between 1.7% and 90.1%. All studies were very heterogeneous with regard to, for example, definitions of hearing loss, used diagnostic tests, participant characteristics, (prior) anti-tumour treatment, other ototoxic drugs and length of follow-up. Therefore, pooling of results was not possible.Only two studies included a control group of people who had not received platinum treatment. In one study, the prevalence of hearing loss was 67.1% (95% confidence interval (CI) 59.3% to 74.1%) in platinum-treated participants, while in the control participants it was 7.4% (95% CI 6.2% to 8.8%). However, hearing loss was detected by screening in survivors treated with platinum analogues and by clinical presentation in control participants. It is uncertain what the effect of this difference in follow-up/diagnostic testing was. In the other study, the prevalence of hearing loss was 20.1% (95% CI 17.4% to 23.2%) in platinum-treated participants and 0.4% (95% CI 0.12% to 1.6%) in control participants. As neither study was a randomized controlled trial or controlled clinical trial, the calculation of a risk ratio was not feasible as it is very likely that both groups differed more than only the platinum treatment.Only two studies evaluated possible risk factors using multivariable analysis. One study identified a significantly higher risk of hearing loss in people treated with cisplatin 400 mg/m(2) plus carboplatin 1700 mg/m(2) as compared to treatment with cisplatin 400 mg/m(2) or less, irrespective of the definition of hearing loss. They also identified a significantly higher risk of hearing loss in people treated with non-anthracycline aminoglycosides antibiotics (using a surrogate marker) as compared to people not treated with them, for three out of four definitions of hearing loss. The other study reported that age at treatment (odds ratio less than 1 for each single-unit increase) and single maximum cisplatin dose (odds ratio greater than 1 for each single-unit increase) were significant predictors for hearing loss, while gender was not. This systematic review shows that children treated with platinum analogues are at risk for developing hearing loss, but the exact prevalence and risk factors remain unclear. There were no data available for tinnitus. Based on the currently available evidence we can only advise that children treated with platinum analogues are screened for ototoxicity in order to make it possible to diagnose hearing loss early and to take appropriate measures. However, we are unable to give recommendations for specific follow-up protocols including frequency of testing. Counselling regarding the prevention of noise pollution can be considered, such as the use of noise-limiting equipment, avoiding careers with excess noise and ototoxic medication. Before definitive conclusions on the prevalence and associated risk factors of platinum-induced ototoxicity can be made, more high-quality research is needed. Accurate and transparent reporting of findings will make it possible for readers to appraise the results of these studies critically.