OBJECTIVES: Neuropathy, a common side effect of taxanes, is often dose-limiting and may result in changes in treatment.  This study examined the occurrence of neuropathy in claims data from commercially insured US patients with BC treated with adjuvant DC or PC. METHODS: This retrospective database analysis used eligibility, medical and pharmacy claims data from a large US health care organization, including subjects with a claim for BC and a claim for DC or PC-containing chemotherapy from 1/1/98-12/31/05. Subjects were stratified by dosing interval (weekly (qw) or Q 21 days (q3w)). Neuropathy was defined using ICD-9-CM codes 356.4, 356.8, 356.9, 357.2, 357.4, 357.5, 357.6, 357.7, 357.8x, 357.9, 377.34, 354.4, 354.5, 354.8, 354.9, 355.7x, 355.8 and 355.9. Neuropathy grade could not be assessed by claims data. Subjects were followed until the earliest of date of death or last enrollment or 12/31/06. Chi-square was used to compare descriptive variables. Logistic regression (LR) was used to examine the independent association of index medication and neuropathy. Covariates included age, geographic region, baseline co-morbidity score and use of medications for neuropathy. RESULTS: A total of 3619 subjects were identified for PC (n=329, qw; 1685, q3w) or DC (n=204, qw; 1045, q3w). A significantly lower frequency of neuropathy was seen in the follow-up period for DC-based treatments compared to PC (7.0% vs 10.6%, p<0.001). Differences were also noted when stratifying by dosing interval (6.7% vs 10.0%; p=0.003 in q3w, 9.3% vs 13.7%; p=0.061 in qw). After adjusting for covariates, the odds of neuropathy remained significantly lower with DC-based treatment (OR=0.70, CI =0.559, 0.920; p=0.010).  CONCLUSIONS: Less neuropathy was noted with DC-based treatment compared to PC. This difference persisted with stratification by dosing interval. The lower occurrence of neuropathy with DC may favor maintenance of dose intensity.