Introduction: Clinical trials have identified peripheral oedema (PO) as an adverse event of vildagliptin (an oral anti-diabetic drug [OAD]). A post-marketing study (PMS) was conducted to advance the understanding of vildagliptin use and particular safety concerns identified within the risk-management plan. PMS objectives included comparing the hazards between vildagliptin monotherapy and combination therapy for selected a priori identified risks, including PO.
Aim: This study was a per-protocol supplementary analysis to investigate the pattern of onset and effect of vildagliptin combination therapy on PO risk.
Methods: The PMS used an observational cohort design. OAD exposure, selected risk factors and outcome data were collected from general practitioners in England regarding vildagliptin users for the 6-month period after starting treatment. Data analysis comprised univariate case/non-case analysis, time-to-onset analysis and Cox proportional hazard models to estimate hazard ratios (HR) of PO adjusting for selected patients' baseline characteristics.
Results: The study cohort included 4828 patients (median age 63 years; interquartile range [IQR] 54-71), 2692 of whom were male (55.8 %). The crude cumulative hazard of PO was 19.09 cases (95 % confidence interval [CI] 13.54-26.10) per 1000 person-years; 50 % of cases occurred during the first 34 days of treatment. A significantly faster time to PO onset was observed in patients prescribed concomitant sulfonylureas versus other treatment combinations (log rank test [LRT] p = 0.0365); in patients with a prior history of PO (LRT p < 0.001), arrhythmia (LRT p = 0.0003) or hypertension (LRT p = 0.0125); and in patients aged ≥60 years (LRT p = 0.0047). Similarly, the case/non-case univariate analysis indicated that patients with PO were older; had a higher prevalence of a history of either arrhythmia, hypertension or PO; and frequently used a sulfonylurea in combination. In the hazard function analysis, only sex and prior PO history had a profound effect on risk of PO after starting vildagliptin. Furthermore, effect modification was observed between sex and prior PO history; in male patients of average age (62 years), the HR was 12.84 (95 % CI 4.96-33.23); in females, it was 1.44 (95 % CI 0.32-6.40).
Conclusions: In this planned supplementary analysis, the findings suggest that PO occurred most frequently within 1 month after starting treatment with vildagliptin, and previous PO history and male sex in elderly patients were important predictors of this risk. The observation that concomitant use of a sulfonylurea may also increase PO risk early after starting treatment should be taken into consideration if prescribing OADs in combination with vildagliptin.