Detecting Medicine Safety Signals Using Prescription Sequence Symmetry Analysis of a National Prescribing Data Set.

INTRODUCTION: Medicine safety signal detection methods employed by the medicine regulator in Australia (Therapeutic Goods Administration [TGA], Department of Health) rely predominantly on analysis of spontaneous adverse event (AE) reports, sponsor notifications or information shared by international agencies. The limitations of these methods and the availability of large administrative health data sets has given rise to greater interest in the use of administrative health data to support pharmacovigilance (PV). OBJECTIVE: We explored whether prescription sequence symmetry analysis (PSSA) of Pharmaceutical Benefits Scheme (PBS) data can enhance signal detection by the TGA, using the AE, heart failure (HF) as a case study. METHODS: We applied the PSSA method to all single-ingredient medicines dispensed under the PBS between 2012 and 2016, using furosemide initiation as a proxy for new-onset HF. A signal was considered present if the lower limit of the 95% confidence interval for the adjusted sequence ratio was > 1. We excluded medicines known to cause HF, indicated for HF treatment or indicated for diseases that may contribute to HF. RESULTS: Of the 654 tested medicines, 26 potential new HF signals were detected by PSSA. Five signals had additional support for the possible association provided by biological plausibility, consistency and disproportionate reporting of cases of HF to the TGA and the World Health Organization; and clinical impact. CONCLUSION: PSSA was able to identify potential signals for further evaluation. With the increasing availability of different administrative health data sources, the strengths and weaknesses of methods used to analyse these data for the purpose of regulatory PV should be evaluated.

Surgical margins and risk of locoregional recurrence in invasive breast cancer: an analysis of 10-year data from the Breast Cancer Treatment Quality Assurance Project.

AIM: There is debate as to what constitutes an adequate excision margin to reduce the risk of locoregional recurrence (LRR) after breast cancer surgery. We have investigated the relationship between surgical margin distance and LRR in women with invasive breast cancer (IBC). METHODS: Tumour free margin distances were extracted from histopathology reports for women with IBC, treated by either breast conserving surgery or mastectomy, enrolled in the Breast Cancer Treatment Group Quality Assurance Project from July 1997 to June 2007. Cox proportional hazards regression analyses were conducted to compare the risk of LRR for involved margins compared with negative margins, measured in increments rounded to the nearest mm. RESULTS: 88 of 2300 patients (3.8%) experienced an LRR after a mean follow-up of 7.9 years. An involved margin, or a margin of 1 mm was associated with an increased risk of LRR (HR 2.72, 95% CI 1.30-5.69), whilst margin distances of 2 mm or greater were not. Risk of LRR with margin distances <2 mm was particularly high amongst those not receiving radiotherapy (RT). CONCLUSION: Based on our findings, we recommend that a tumour free margin distance of 2 mm be adopted as an adequate margin of excision for IBC, in the setting of patients receiving standard adjuvant RT and adjuvant drug therapies as dictated by the current clinical treatment paradigms.

Sentinel node biopsy in larger or multifocal breast cancers: to do or not to do.

BACKGROUND: The use of sentinel node biopsy (SNB) in breast cancer patients with large and/or multifocal tumours is controversial. METHODS: A review of clinical records was undertaken for 213 consecutive patients undergoing SNB for invasive breast cancer from September 2000 to February 2006. The results of SNB and axillary dissection were compared for patients with unifocal or multifocal tumours less than 3 cm and 3 cm or larger. Patient outcomes were also assessed. RESULTS: The mean number of sentinel nodes removed per patient increased from 2.33 in 2000 to 4.17 in 2006. For patients with unifocal tumours less than 3 cm, 47 of 147 (32.0%) were sentinel node positive compared with 15 of 30 (50%) for multifocal tumours less than 3 cm (P = 0.04), 19 of 28 (67.9%) for unifocal tumours 3 cm or larger (P < 0.001) and 7 of 8 (87.5%) for multifocal tumours 3 cm or larger (P = 0.003). Following axillary dissection, 20 of 48 (41.7%) patients with sentinel node macrometastases were found to have positive non-sentinel nodes, compared with 4 of 20 (20.0%) and 1 of 8 (12.5%) for patients with sentinel node micrometastases and isolated tumour cells. The mean total number of positive nodes was 1.74 compared with 4.21 for unifocal tumours less than or greater than 3 cm, respectively (P = 0.004). No axillary recurrences were detected during the follow-up period. CONCLUSION: Although patients with large and/or multifocal tumours were more likely to have a positive sentinel node, the findings provide some indication that SNB may be reliable for staging the axilla in these patients.