br Patients and Methods Patients who had been initiated on
Patients and Methods Patients who had been initiated on a DPP-4 inhibitor (see in the online version at doi:10.1016/j.clinthera.2018.06.002) between April 2012 and June 2017, when the common threshold for dose adjustment for all non-linagliptin DPP-4 inhibitors was a creatinine clearance (CrCl) level of 50 mL/min according to the respective SPCs,3, 4, 5, 6 were identified in the Clinical Practice Research Datalink (CPRD). CPRD is a large, anonymized general practice database representative of the UK population. Quality criteria described previously were applied to define the patient cohort eligible for inclusion in the analysis. In particular, patients were to be of an acceptable CPRD research standard and registered to a general practice for at least 12 months, which had to be of “Up to Standard” at DPP-4 inhibitor initiation. Patients should have had a record of a T2DM-related medical code (see in the online version at doi:10.1016/j.clinthera.2018.06.002) before initiation of the DPP-4 inhibitor, with no history of type 1 the truth and be at least 18 years of age at T2DM diagnosis. Patients were to have at least 1 record of serum creatinine measurement before DPP-4 inhibitor initiation and at least 1 weight record around the time of DPP-4 inhibitor initiation. Patients initiated on vildagliptin should have had information on their daily dose recorded in their first prescription. To account for outliers and incorrect entries in patient records, serum creatinine and weight values <1st percentile and >99th percentile were excluded from the analysis. In addition, patients with a CrCl level >50 mL/min before or at DPP-4 inhibitor initiation who had a record of renal impairment in their clinical files were not included in the analysis. Patients included in the analysis were to have a CrCl level >50 mL/min with no recorded history of renal impairment and were assessed toward SPCs valid at the time of treatment. The analysis generated counts and percentages of patients with a CrCl level >50 mL/min who were initiated on alogliptin 12.5mg or 6.25mg, sitagliptin 50mg or 25mg, saxagliptin 2.5 mg, and vildagliptin 50 mg once daily that were lower than the SPC-specified doses, overall and according to DPP-4 inhibitor separately. CrCl levels were estimated by using the Cockcroft-Gault formula from the last serum creatinine measurement recorded before DPP-4 inhibitor initiation and also the weight measurement closest to the DPP-4 inhibitor initiation date. Patients who were prescribed vildagliptin 50 mg once daily but had a record of treatment with a sulfonylurea within 30 days of vildagliptin initiation were classified as treated in accordance with SPCs.
Results Of the 36,874 patients initiated on a DPP-4 inhibitor during the study period, 30,346 (82.3%) met the eligibility criteria. Of the 30,346 eligible patients, 3142 (10.4%) had a CrCl value >50 mL/min before DPP-4 inhibitor initiation coupled with a record of renal impairment in their clinical files, and they were therefore excluded from the analysis. Among the remaining 27,204 patients, 4011 (14.7%) had a CrCl value ≤50 mL/min before DPP-4 inhibitor initiation and were also excluded from the analysis. The remaining 23,193 (85.3%) patients with a CrCl value >50 mL/min according to their last serum creatinine measurement before or at DPP-4 inhibitor initiation with no history of renal impairment were included in the analytical cohort (). The majority of patients with CrCl values >50 mL/min were initiated on sitagliptin (14,779 of 23,193 [64%]), of whom 14% (2061 of 14,779) were on a lower dose than that specified in the SPC. Overall, 12% (2764 of 23,193) of patients with CrCl values >50 mL/min initiated on a DPP-4 inhibitor were on a lower dose than the SPC specified, which increased to 14% when patients initiated on linagliptin who were not at risk of being prescribed a lower dose were excluded (). Sensitivity analysis restricting the time period between last creatinine and weight measurement at maximum 90 days (to control for secular volatility of CrCl) generated similar results.