Costs and Adherence With U-500 vs High-Dose U-100 Insulin
This study provides the first large database analysis of patients taking U-500R compared with patients taking high-dose (≥150 units/day) U-100 insulin treatment using MDI or CSII (a use for U-500R that is not approved by the U.S. Food and Drug Administration) to evaluate comparative healthcare costs and adherence rates. There have been relatively few published case series focused on the use of U-500R by MDI (315 patients) and CSII (65 patients), as reported in recent reviews. No randomized clinical trials of U-500R have been completed, although 1 trial has been initiated. Only one large naturalistic study has been published to date, but it was limited to patients receiving >200 units/day of U-100 insulin by MDI. By including patients receiving ≥150 units/day of U-100 insulin and both MDI and CSII delivery modalities in this study, we achieved a much greater capture rate of highdose insulin-using patients in the database with a higher number of propensity-matched pairs available for analysis (684 in the earlier analysis [8] and 1,039 in this study). Therefore, this study has the potential to provide significant insights into the real-world utilization of these highdose insulin regimens.
The key findings are that treatment with U-500R as compared with high-dose U-100 therapy (≥150 units/day) is associated with decreased pharmacy costs, a higher proportion of patients experiencing hypoglycemia, and greater treatment adherence. In addition, U-500R–treated patients (prematch, preperiod) versus the U-100 cohort had higher overall healthcare costs, more comorbidities/complications, higher CCI scores, higher inpatient admission rates at baseline, more care by endocrinologists, more treatment by noninsulin injectables, and a higher prevalence of treatment by CSII.
Analysis of medical and pharmacy-related costs for the matched cohorts revealed that compared with treatment with ≥150 units/day of U-100 insulin, treatment with U-500R is associated with a significant decrease in mean pharmacy costs, with no significant difference in overall costs. The increase in medical costs in the U-500R cohort, largely due to non–diabetes-related costs, was offset by the reduction of pharmacy costs, driven by lower insulin costs. There were no significant differences in medical costs for the U-100 cohort; however, an increase in pharmacy costs due to higher insulin costs led to an overall cost increase. The magnitude of these cost differences reached significance only in the difference-in-differences for pharmacy costs.
Hypoglycemia-specific costs and the mean number of hypoglycemic events per patient were similar between the cohorts, but the percentage of patients with at least one hypoglycemic event during the postperiod was significantly higher in the U-500R cohort than in the U-100 cohort. The higher adherence to the insulin regimen in the U-500R cohort may have contributed to the higher percentage of patients with hypoglycemic events. However, glycemic control could not be assessed (hemoglobin A1c [HbA1c] values were not available) in this study, and tighter glycemic control in patients requiring insulin generally correlates with an increased frequency of hypoglycemia. Given the identified risks of "dose confusion" with use of U-100 insulin syringes or volumetric (tuberculin) syringes, great care must be taken in appropriate counseling of patients being switched to U-500R from U-100 insulin regimens. It was impossible in this study to assess details about the real-world instruction of the U-500R cohort that could relate to the occurrence of captured hypoglycemic events. There have been few reports of U-500R use with occurrence of severe hypoglycemia, although Ziesmer and colleagues) recently reported a case series in which 8 severe hypoglycemic episodes were observed in a 53-patient study over a mean treatment duration of 20 months. Although overinsulinization can occur and can lead to the incidence of hypoglycemic events, the majority of high-dose U-100 insulin-treated patients have high HbA1c values and are quite underinsulinized prior to initiation of U-500R. Recent American Association of Clinical Endocrinologists consensus statements/algorithms emphasize the importance of avoiding hypoglycemia in patients with T2D. For U-500R, specific treatment algorithms using U-500R have not yet been tested in randomized clinical trials.
Consistent with the previous database analysis of U-500R versus >200 units/day of U-100, adherence to U-500R was significantly higher than adherence to high-dose U-100 in this study. However, the differences in adherence rates between the U-500R and U-100 arms were slightly lower than what was reported in the >200 units/day arm of a previous study due to higher PDCs seen in the ≥150 units/day cohort of this study. The PDCs were nearly identical for the U-500R arms of the 2 studies, suggesting that delivery via CSII (not included in the previous study) did not affect the adherence rate. Similar results were found in the subgroup analysis of adherence and CSII versus MDI, in which the PDC was significantly higher for U-100 CSII users as compared with MDI users, with no significant difference in the PDCs between delivery devices in the U-500R arm. Together, these results suggest worsening adherence rates are associated with higher-dose U-100 therapy, perhaps because of volumerelated factors such as greater number of required injections, greater complexity of regimens, leakage at injection sites, and injection discomfort. Given the longer duration of action of U-500R compared with U-100, patients can be dosed 2 to 3 times daily with U-500R.
The novel finding in this study that almost one-third of the patients used U-500R via CSII (off label) in the postperiod is perhaps not surprising, given technical advances in insulin pumping over the last 3 decades and increasing experience (and improving reimbursement) with CSII in patients with T2D. Earlier reports have suggested cost savings with the use of U-500R versus U-100 insulins via CSII, including reduced costs for insulins, infusion sets, and pump batteries. Studies have also supported improved satisfaction scores with U-500R versus U-100 insulins via CSII.
There are limitations to this study. This was a retrospective analysis using claims data, not a randomized controlled trial. Retrospective analyses must rely upon diagnostic codes to identify patients, whereas formal diagnostic assessments may be more reliable. The costs of insulin may be underestimated compared with present-day values because an inflation adjustment was not factored into the analyses. These cost results are based on 2008–2010 insulin prices that do not reflect the current cost of U-500R, which is now identical (wholesale unit to unit) to the cost of Lilly's U-100R insulin. There may also be potentially confounding variables that are not captured from the MarketScan database, including glycemic control and the duration and severity of diabetes, so surrogate indicators of disease severity, including the presence of comorbidities, high CCI scores, and previous hypoglycemic events were used in this study. Hypoglycemia rates were calculated using ICD-9-CM codes, which require some medical intervention, so milder events may not have been captured (a recent study showed that patients experience hypoglycemia significantly more often than is captured in claims data [26]). The PDC as a measure of adherence may not reflect the true adherence of the insulin therapy, which considers the correct insulin dose and timing of doses in accordance with exercise, diet, and other factors. Finally, the results may not be representative of uninsured patients or elderly patients with only basic Medicare coverage (the MarketScan database does not include all Medicare claims for those older than 65 years).
Discussion
This study provides the first large database analysis of patients taking U-500R compared with patients taking high-dose (≥150 units/day) U-100 insulin treatment using MDI or CSII (a use for U-500R that is not approved by the U.S. Food and Drug Administration) to evaluate comparative healthcare costs and adherence rates. There have been relatively few published case series focused on the use of U-500R by MDI (315 patients) and CSII (65 patients), as reported in recent reviews. No randomized clinical trials of U-500R have been completed, although 1 trial has been initiated. Only one large naturalistic study has been published to date, but it was limited to patients receiving >200 units/day of U-100 insulin by MDI. By including patients receiving ≥150 units/day of U-100 insulin and both MDI and CSII delivery modalities in this study, we achieved a much greater capture rate of highdose insulin-using patients in the database with a higher number of propensity-matched pairs available for analysis (684 in the earlier analysis [8] and 1,039 in this study). Therefore, this study has the potential to provide significant insights into the real-world utilization of these highdose insulin regimens.
The key findings are that treatment with U-500R as compared with high-dose U-100 therapy (≥150 units/day) is associated with decreased pharmacy costs, a higher proportion of patients experiencing hypoglycemia, and greater treatment adherence. In addition, U-500R–treated patients (prematch, preperiod) versus the U-100 cohort had higher overall healthcare costs, more comorbidities/complications, higher CCI scores, higher inpatient admission rates at baseline, more care by endocrinologists, more treatment by noninsulin injectables, and a higher prevalence of treatment by CSII.
Analysis of medical and pharmacy-related costs for the matched cohorts revealed that compared with treatment with ≥150 units/day of U-100 insulin, treatment with U-500R is associated with a significant decrease in mean pharmacy costs, with no significant difference in overall costs. The increase in medical costs in the U-500R cohort, largely due to non–diabetes-related costs, was offset by the reduction of pharmacy costs, driven by lower insulin costs. There were no significant differences in medical costs for the U-100 cohort; however, an increase in pharmacy costs due to higher insulin costs led to an overall cost increase. The magnitude of these cost differences reached significance only in the difference-in-differences for pharmacy costs.
Hypoglycemia-specific costs and the mean number of hypoglycemic events per patient were similar between the cohorts, but the percentage of patients with at least one hypoglycemic event during the postperiod was significantly higher in the U-500R cohort than in the U-100 cohort. The higher adherence to the insulin regimen in the U-500R cohort may have contributed to the higher percentage of patients with hypoglycemic events. However, glycemic control could not be assessed (hemoglobin A1c [HbA1c] values were not available) in this study, and tighter glycemic control in patients requiring insulin generally correlates with an increased frequency of hypoglycemia. Given the identified risks of "dose confusion" with use of U-100 insulin syringes or volumetric (tuberculin) syringes, great care must be taken in appropriate counseling of patients being switched to U-500R from U-100 insulin regimens. It was impossible in this study to assess details about the real-world instruction of the U-500R cohort that could relate to the occurrence of captured hypoglycemic events. There have been few reports of U-500R use with occurrence of severe hypoglycemia, although Ziesmer and colleagues) recently reported a case series in which 8 severe hypoglycemic episodes were observed in a 53-patient study over a mean treatment duration of 20 months. Although overinsulinization can occur and can lead to the incidence of hypoglycemic events, the majority of high-dose U-100 insulin-treated patients have high HbA1c values and are quite underinsulinized prior to initiation of U-500R. Recent American Association of Clinical Endocrinologists consensus statements/algorithms emphasize the importance of avoiding hypoglycemia in patients with T2D. For U-500R, specific treatment algorithms using U-500R have not yet been tested in randomized clinical trials.
Consistent with the previous database analysis of U-500R versus >200 units/day of U-100, adherence to U-500R was significantly higher than adherence to high-dose U-100 in this study. However, the differences in adherence rates between the U-500R and U-100 arms were slightly lower than what was reported in the >200 units/day arm of a previous study due to higher PDCs seen in the ≥150 units/day cohort of this study. The PDCs were nearly identical for the U-500R arms of the 2 studies, suggesting that delivery via CSII (not included in the previous study) did not affect the adherence rate. Similar results were found in the subgroup analysis of adherence and CSII versus MDI, in which the PDC was significantly higher for U-100 CSII users as compared with MDI users, with no significant difference in the PDCs between delivery devices in the U-500R arm. Together, these results suggest worsening adherence rates are associated with higher-dose U-100 therapy, perhaps because of volumerelated factors such as greater number of required injections, greater complexity of regimens, leakage at injection sites, and injection discomfort. Given the longer duration of action of U-500R compared with U-100, patients can be dosed 2 to 3 times daily with U-500R.
The novel finding in this study that almost one-third of the patients used U-500R via CSII (off label) in the postperiod is perhaps not surprising, given technical advances in insulin pumping over the last 3 decades and increasing experience (and improving reimbursement) with CSII in patients with T2D. Earlier reports have suggested cost savings with the use of U-500R versus U-100 insulins via CSII, including reduced costs for insulins, infusion sets, and pump batteries. Studies have also supported improved satisfaction scores with U-500R versus U-100 insulins via CSII.
There are limitations to this study. This was a retrospective analysis using claims data, not a randomized controlled trial. Retrospective analyses must rely upon diagnostic codes to identify patients, whereas formal diagnostic assessments may be more reliable. The costs of insulin may be underestimated compared with present-day values because an inflation adjustment was not factored into the analyses. These cost results are based on 2008–2010 insulin prices that do not reflect the current cost of U-500R, which is now identical (wholesale unit to unit) to the cost of Lilly's U-100R insulin. There may also be potentially confounding variables that are not captured from the MarketScan database, including glycemic control and the duration and severity of diabetes, so surrogate indicators of disease severity, including the presence of comorbidities, high CCI scores, and previous hypoglycemic events were used in this study. Hypoglycemia rates were calculated using ICD-9-CM codes, which require some medical intervention, so milder events may not have been captured (a recent study showed that patients experience hypoglycemia significantly more often than is captured in claims data [26]). The PDC as a measure of adherence may not reflect the true adherence of the insulin therapy, which considers the correct insulin dose and timing of doses in accordance with exercise, diet, and other factors. Finally, the results may not be representative of uninsured patients or elderly patients with only basic Medicare coverage (the MarketScan database does not include all Medicare claims for those older than 65 years).
Source...