Effect of Ferric Carboxymaltose Versus Low-Dose Intravenous Iron Therapy and Iron Sucrose on the Total Cost of Care in Patients with Iron Deficiency Anemia: A US Claims Database Analysis

This retrospective claims analysis utilized medical and pharmacy claims from the IQVIA PharMetrics Plus database (IQVIA, Durham, NC, USA), from 01/01/2017 to 31/12/2019. IQVIA PharMetrics Plus is a longitudinal health plan database comprising deidentified medical and pharmacy claims data among commercially insured patients and a small portion of Medicare and Medicaid patients in the USA [14]. Data in the database are deidentified and compliant with the Health Insurance Portability and Accountability Act. Patients with an index claim of FCM or LDI (including iron sucrose, iron dextran, sodium ferric gluconate complex in sucrose) between 1 January 2017 and 31 December 2019 were included (Online Resource Supplementary Fig. 1). The index date was the date of their first claim for IV iron. Patients were required to have continuous enrollment for ≥ 6 months before the index date and ≥ 12 months after the index date and at least one medical claim associated with IDA (ICD10: D50.x) during the study observation period. Patients who had a medical claim for CKD stage 5 (ICD10, N18.5) and end-stage renal disease (N18.6) or dialysis [Healthcare Common Procedure Coding System Current Procedural Terminology (HCPCS CPT) 90935–90999] during the 6-month period before the index date were excluded.

Patients treated with FCM were compared with a cohort of patients treated with LDI, which included iron sucrose, iron dextran, and sodium ferric gluconate complex in sucrose. Additionally, a subanalysis was completed in which patients treated with iron sucrose were extracted from the LDI cohort for a comparison with FCM. Patients in the FCM cohort were required to receive a second dose of FCM within 21 days of the index date. Patients were excluded in the FCM cohort if they had received other IV iron infusions during the 6-month period before their index date. Patients were excluded in the LDI and iron sucrose cohorts if they had received FCM, ferumoxytol, or other IV iron infusions during the 6-month period before their index date. Ferric derisomaltose was not available in the USA during the study period.

Analyses were performed in the overall population and in subgroups of patients with cancer, HF, and CKD. In the cancer subgroup, patients were required to have a medical claim associated with a cancer diagnosis (ICD10: C00.X–C96.X) in the 6-month period before their index date, respectively. Similarly, in the HF subgroup, patients were required to have a medical claim associated with a HF diagnosis (ICD10: I50.1–I50.9) in the 6-month period before their index date, respectively. The CKD subgroup included patients with a medical claim associated with a CKD diagnosis (ICD10: N18.1–N18.5) in the 6 months prior to index IV infusion. These patient subgroups were not mutually exclusive.

Healthcare resource use data were summarized on a per-patient-per-month (PPPM) basis for the FCM group and compared with the LDI and iron sucrose groups using t-tests and logistic regression analysis. Descriptive statistics were used to summarize the differences in monthly total healthcare cost before and after index IV iron therapy. Because cost data are skewed, monthly healthcare costs after receiving index IV iron were compared using a generalized linear model with gamma distribution and log-link function [15]. The patient costs were divided into the following five mutually exclusive groups: inpatient, emergency room (ER), outpatient (IV iron infusion and related costs), outpatient other costs (excluding IV iron infusion and related costs), and pharmacy oral. The adjusted cost ratio post-index analysis was conducted for total cost, inpatient cost, and outpatient other cost. Covariates included in the monthly total cost regression models were the pre-index monthly total healthcare cost, patient age and gender, year of treatment index, and Charlson Comorbidity Index (CCI). In the cancer subgroup, the National Cancer Institute (NCI) Comorbidity Index was used instead of the CCI. Covariates included in the monthly inpatient cost and monthly outpatient cost regression models were similar except pre-index monthly total healthcare cost was replaced by pre-index monthly inpatient cost and pre-index monthly outpatient cost, respectively. To assess the robustness of the study results, three sensitivity analyses were performed to compare the costs for FCM versus LDI and iron sucrose. In the propensity score–weighting sensitivity analysis, monthly total healthcare cost before index, patient age and gender, year of treatment index date, and CCI (or NCI Comorbidity Index for the cancer subgroup) were used to estimate propensity scores, and generalized linear model with inverse probability of treatment weights of propensity scores were fitted for cost ratios. In the matched baseline total cost sensitivity analysis, only FCM and LDI/iron sucrose patients with similar baseline total cost were included in the analysis. The pre-index period was defined as the 6 months prior to the patient’s index date. The third sensitivity analysis used a generalized linear model with gamma distribution and log-link function comparing monthly total healthcare cost, inpatient cost, and outpatient cost of patients who had received at least one dose of FCM compared with the singular dose that patients received among the LDI group. Each sensitivity analysis aimed to serve different purposes, with the goal of enhancing the validity of the study findings. Propensity score weighting was used to help control for selection bias [16]. The generalized linear model with gamma distribution and log-link function was used to help address non-normality of cost data and provide more accurate estimates among skewed distributions. Analyses were conducted using SAS 9.4 software.

The subgroup of patients with cancer included data from 7829 patients treated with FCM and 6899 patients treated with LDI, with 4086 of those patients having been treated with iron sucrose (Online Resource Supplementary Table 1). Gender distribution (female: 64% for FCM and LDI and 63% for iron sucrose) and mean age were similar among the three cohorts. Mean (SD) NCI Comorbidity Index was similar among the FCM cohort at 0.52 (0.6), LDI cohort at 0.56 (0.6), and iron sucrose cohort at 0.58 (0.6) (Online Resource Supplementary Table 1). At 12 months post-index, FCM had the lowest PPPM number of IV iron infusions (0.21) versus LDI (0.35) and iron sucrose (0.39), respectively (Online Resource Supplementary Table 2). FCM had a greater decrease in the PPPM number of inpatient admission days versus LDI and iron sucrose from the pre- to the post-index period (mean difference, pre- versus post-index, − 0.11 versus − 0.04 and − 0.04, respectively). There was a smaller increase in the mean number of outpatient visits from the pre- to the post-index period in the FCM group versus LDI and iron sucrose (Online Resource Supplementary Table 2). Unadjusted mean monthly total cost post-index was lowest for FCM at $5596 PPPM versus LDI and iron sucrose at $6367 PPPM and $6820 PPPM, respectively (Fig. 2). FCM had the smallest increase in costs from pre- to post-index ($684 PPPM) and iron sucrose had the largest increase in costs ($999 PPPM). The adjusted cost ratio for post-index total cost was significantly lower for FCM versus both LDI (0.89) and iron sucrose (0.84) (P < 0.0001). Adjusted post-index inpatient and outpatient costs were 21 and 12% lower for FCM versus LDI, respectively (P < 0.0001). Similar levels of reduction were observed for FCM compared with iron sucrose (Fig. 2).

Data were analyzed from 3002 patients with HF treated with FCM and 3641 treated with LDI, with 2403 of those patients having been treated with iron sucrose (Online Resource Supplementary Table 1). Most patients were female (57, 56, and 55% for FCM, LDI, and iron sucrose, respectively). Mean age was slightly older for the FCM than the LDI and iron sucrose cohorts (69.4, 67.0, and 67.2 years, respectively). Mean (SD) CCI was highest for FCM at 1.9 (2.5), followed by LDI at 1.7 (2.5) and iron sucrose at 1.6 (2.5) (Online Resource Supplementary Table 1). Mean PPPM number of IV iron infusions was lowest for those treated with FCM (0.25) versus those treated with LDI (0.40) and iron sucrose (0.44) (Online Resource Supplementary Table 2). In patients treated with FCM, there was a decrease in number of outpatient visits versus those treated with LDI and iron sucrose from the pre-index to the post-index period (mean difference PPPM, pre- versus post-index, 0.03 versus 0.54 and 0.66, respectively). Similarly, the FCM cohort had a decrease in number of inpatient admission days versus LDI and iron sucrose during the post-index period (mean difference PPPM, pre- versus post-index, − 0.12 versus 0.05 and 0.05, respectively) (Online Resource Supplementary Table 2). Unadjusted mean monthly total cost post-index was lowest for FCM at $4081 PPPM versus LDI and iron sucrose ($5804 PPPM and $5691, respectively) (Fig. 3). Unadjusted mean monthly inpatient and outpatient costs were lowest for FCM as well. The adjusted cost ratio post-index for total cost was 0.78 for FCM versus LDI (P < 0.0001) and 0.75 for FCM versus iron sucrose (P < 0.0001). FCM had the smallest increase in costs from pre- to post-index ($40 PPPM), while LDI and iron sucrose had similar increases ($404 and $415 PPPM, respectively). The adjusted cost ratios for inpatient and outpatient costs revealed a 34% reduction in inpatient cost and a 23% reduction in outpatient cost for FCM versus LDI, and a 31% reduction in inpatient cost and a 27% reduction in outpatient cost for FCM versus iron sucrose (all P < 0.0001) (Fig. 3).

Data were analyzed from 3727 patients with CKD treated with FCM and 5068 treated with LDI, with 3489 of those patients having been treated with iron sucrose (Online Resource Supplementary Table 1). Most patients were female (61% in the FCM and 60% in the LDI and iron sucrose cohorts). Patients in the FCM cohort were slightly older [67.9 versus 64.8 (LDI) and 64.6 (iron sucrose) years]. Mean (SD) CCI was highest for those patients treated with FCM at 1.9 (2.5), followed by both LDI and iron sucrose [1.6 (2.4) and 1.6 (2.4), respectively] (Online Resource Supplementary Table 1). Mean PPPM number of IV iron infusions was lowest for the FCM cohort versus the LDI and iron sucrose cohorts post-index (0.24 versus 0.43 and 0.46, respectively) (Online Resource Supplementary Table 2). Unadjusted mean monthly total cost post-index was lowest for those in the FCM cohort at $3660 PPPM, followed by $4804 PPPM for LDI and $4916 for iron sucrose (Fig. 4). FCM had the lowest increase in costs from pre- to post-index ($478 PPPM) compared with LDI ($763 PPPM) and iron sucrose ($725 PPPM). Similarly, unadjusted costs were lowest for the FCM cohort for inpatient and outpatient costs. Adjusted post-index cost ratios showed significantly lower total costs for the FCM cohort versus LDI (0.81; P < 0.0001) and iron sucrose (0.79, P < 0.0001). The adjusted cost ratio was significantly lower for the FCM cohort for inpatient and outpatient costs as well, with a 27% reduction for both inpatient and outpatient costs versus LDI, and a 26% reduction in inpatient cost and a 30% reduction in outpatient cost versus iron sucrose (all P < 0.0001) (Fig. 4).

The results of the three sensitivity analyses were generally consistent with the main analyses (Table 1), confirming a generally significant reduction in monthly post-index healthcare costs in the FCM group versus the LDI and iron sucrose groups. The results of the three sensitivity analyses were similar to the main analyses for the cancer, HF, and CKD subgroups, confirming statistically significant lower monthly total, inpatient, and outpatient costs in the FCM group versus the LDI and iron sucrose groups during the 12 months after initiating IV iron therapy (Table 1).

When patients receiving at least one dose of FCM were included in the overall study cohort analysis, post-index monthly total cost in the FCM group was similar to LDI, with the upper bound of the 95% confidence interval of the adjusted cost ratio reaching 1.0 (Table 1). However, post-index monthly inpatient and outpatient costs remained significantly lower in the FCM group versus LDI and iron sucrose. Adjusted cost ratios derived from the propensity score weighting method provided more conservative estimates of the level of cost reduction, while results from the matched baseline cost analysis provided similar results as the main analysis. Using the propensity score weighting method, post-index monthly inpatient cost of FCM was similar to LDI and iron sucrose in the overall study cohort analysis. Monthly outpatient cost during the 12 months after index IV iron treatment for FCM was also similar to LDI in the overall study cohort analysis under the propensity score weighting method (Table 1). Distribution of propensity score weights for each comparison are presented in Online Resource Supplementary Fig. 2.