Clinical Effectiveness and Safety of Acupuncture in the Treatment of Irradiation-Induced Xerostomia in Patients with Head and Neck Cancer: A Systematic Review

Abstract

Irradiation-induced xerostomia seriously reduces quality of life for patients with head and neck cancer (HNC). Anecdotal evidence suggests that acupuncture may be beneficial. To systematically review evidence on clinical effectiveness and safety of acupuncture in irradiation-induced xerostomia in patients with HNC. A detailed search was performed to identify randomised controlled trials (RCTs) and systematic reviews of RCTs on acupuncture in irradiation-induced xerostomia, using AMED, BNIA, CINAHL, Cochrane, Embase, HPSI, PsycInfo and Medline. Grey literature was explored and 11 journals hand searched. Search terms included: acupuncture, xerostomia, salivary hypofunction, hyposalivation, dry mouth, radiotherapy, irradiation, brachytherapy, external beam. Two authors independently extracted data for analysis using predefined selection criteria and quality indicators. 43 of the 61 articles identified were excluded on title/abstract. 18 articles underwent full-text review; three were deemed eligible for inclusion. Two trials had moderate risk of bias; one had high risk. Two trials compared acupuncture with sham acupuncture; one control arm received 'usual care'. Outcome measurements included salivary flow rates (SFRs) in two trials and subjective questionnaires in three. All three trials reported significant reduction in xerostomia versus baseline SFR (p<0.05); one reported greater effect in the intervention group for stimulated SFR (p<0.01). Subjective assessment reported significant differences between real acupuncture and control in two trials (p<0.02-0.05). Insufficient evidence was presented to undertake risk/benefit assessment. Limited evidence suggests that acupuncture is beneficial for irradiation-induced xerostomia. Although current evidence is insufficient to recommend this intervention, it is sufficient to justify further studies. Highlighted methodological limitations must be dealt with.