Leukemia & Lymphoma

ISSN: 1042-8194 (Print) 1029-2403 (Online) Journal homepage: http://www.tandfonline.com/loi/ilal20

Medical complications, resource utilization and costs in patients with myelofibrosis by frequency of blood transfusion and iron chelation therapy Francis Vekeman, Wendy Y. Cheng, Medha Sasane, Lynn Huynh, Mei Sheng Duh, Carole Paley & Ruben A. Mesa To cite this article: Francis Vekeman, Wendy Y. Cheng, Medha Sasane, Lynn Huynh, Mei Sheng Duh, Carole Paley & Ruben A. Mesa (2015): Medical complications, resource utilization and costs in patients with myelofibrosis by frequency of blood transfusion and iron chelation therapy, Leukemia & Lymphoma To link to this article: http://dx.doi.org/10.3109/10428194.2015.1016933

View supplementary material

Accepted author version posted online: 13 Feb 2015. Published online: 30 Mar 2015. Submit your article to this journal

Article views: 71

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ilal20 Download by: [Universite Laval]

Date: 05 November 2015, At: 18:12

Leukemia & Lymphoma, 2015; Early Online: 1–9 © 2015 Informa UK, Ltd. ISSN: 1042-8194 print / 1029-2403 online DOI: 10.3109/10428194.2015.1016933

original article: CLINICAL

Medical complications, resource utilization and costs in patients with myelofibrosis by frequency of blood transfusion and iron chelation therapy Francis Vekeman1, Wendy Y. Cheng2, Medha Sasane3, Lynn Huynh2, Mei Sheng Duh2, Carole Paley3 & Ruben A. Mesa4 1Groupe d’analyse, Ltée, Montréal, Québec, Canada, 2Analysis Group, Inc., Boston, MA, USA, 3Novartis Pharmaceuticals

Downloaded by [Universite Laval] at 18:12 05 November 2015

Corporation, East Hanover, NJ, USA and  4Mayo Clinic Cancer Center, Scottsdale, AZ, USA

or leukocytosis, thrombocytopenia or thrombocytosis, and multi-organ extramedullary hemopoiesis [3,4]. Patients diagnosed with advanced MF experience constitutional symptoms, pain, splenic infarction, portal hypertension and dyspnea as a result of massive splenomegaly [3]. Myelofibrosis is a rare condition affecting 1.5 out of 100 000 people annually [1]. Median survival is about 5 years, and can range between 2 years and more than 10 years depending on the presence of well-defined prognostic indicators [5]. Treatment of MF usually includes red blood cell (RBC) transfusions, radiation, chemotherapy or stem cell transplant. Treatment focusing on symptom management and optimizing cell counts is available for patients with splenomegaly and anemia. Patients with MF develop anemia because of the decreased marrow reserve [6]. For symptomatic patients with MF with anemia, RBC transfusion is the standard therapy [3]. A consequence of regular transfusion in patients with MF is iron overload, which can lead to multiple comorbid conditions and severe complications, including significant impairment in vital organs, primarily the liver, heart and endocrine glands [3,7–16]. Iron overload has also been associated with shorter survival rates [6]. Transfusion-induced iron overload is a substantial clinical problem among transfusion-dependent (TD) patients with MF, which can be treated with iron chelation therapy (ICT) such as deferoxamine (Desferal®) or deferasirox (ICL670, Exjade®) [17]. In a retrospective study in patients with primary MF, median overall survival (OS) was longer among TD ICT patients compared with TD ICT patients (median OS was not reached for TD ICT patients vs. 58 months for TD ICT patients) [6]. In three case report studies, patients with primary MF were reported to no longer require transfusion after receiving ICT for a substantial period of time [18–20]. Despite the clinical benefits of ICT, the health resource utilization (RU) and cost associated with ICT in patients with TD MF, in particular, remains undefined. One study has examined the overall health RU and cost associated

Abstract Iron chelation therapies (ICTs) can help eliminate iron surplus in erythrocyte transfusion-dependent (TD) patients with myelofibrosis (MF). The study assessed adjusted incidence rate ratios (aIRRs) of MF-related complications and resource utilization (RU) and adjusted mean monthly inpatient cost differences in patients with TD MF treated with versus without ICT (ICT vs. ICT) using data from two healthcare claims databases. Patients with  2 MF International Classification of Diseases, 9th Revision (ICD-9) diagnosis codes  30 days apart were included. Among 571 patients with TD MF, 103 (18%) were ICT and 468 (82%) were ICT. ICT patients had lower rates of thrombocytopenia (aIRR: 0.55; p  0.001), pancytopenia (0.53; p  0.001), emergency room visits (0.84 [95% confidence interval: 0.74–0.96]) and inpatient stays (0.75 [0.64–0.87]), but higher rates of outpatient visits (1.21 [1.18–1.23]). Adjusted mean complication-related inpatient cost difference per month was lower in ICT patients ($1804 [$570]; p  0.004). ICT patients had significantly lower rates of acute care, but higher rates of outpatient care. Keywords: Myelofibrosis, transfusion dependent, iron chelation therapy, medical resource utilization, economic cost

Introduction Myelofibrosis (MF) is a chronic malignant hematological disease and is a part of a group of conditions referred to as myeloproliferative neoplasms, which include MF, polycythemia vera and essential thrombocythemia [1,2]. It is a disease of the bone marrow in which collagen builds up fibrous scar tissue inside the marrow cavity, resulting in low blood counts. Myelofibrosis may be present as a primary condition or secondary condition due to progression of polycythemia vera or essential thrombocythemia [1]. The clinical features include progressive anemia, leukopenia

Correspondence: Francis Vekeman, MA, Vice President, Groupe d’analyse, Ltée, 1000 De La Gauchetiere West, Suite 1200, Montreal, H3B 4W5, QC, Canada. Tel: 514-394-4442. Fax: 514-394-4461. E-mail: [email protected] Received 22 September 2014; revised 29 January 2015; accepted 2 February 2015

1

2  F. Vekeman et al. with myeloproliferative neoplasms, but lacked a focus on transfusion dependence and ICT [21]. Given the limited information on the RU and cost impact of MF on the US healthcare system, especially among patients with frequent RBC transfusions and ICT use, this study aimed to examine and compare RBC transfusion patterns, incidence of MF-related complications, incidence of all-cause and MF-related RU and cost in patients with TD MF with versus without ICT (ICT vs. ICT) from a US managed care perspective.

Methods

Downloaded by [Universite Laval] at 18:12 05 November 2015

Data sources The study was based on a retrospective analysis of health insurance claims from two databases: the Truven Health Analytics MarketScan [22] (2000–2012) and IMS Health LifeLink PharMetrics Integrated Database [23] (2001–2012). The claims databases are fully de-identified and Health Insurance Portability and Accountability Act (HIPAA)-compliant.

Study design and patient population Patients with  2 International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis codes for MF (238.76 or 289.83) that were  30 days apart and who were  18 years of age at first MF diagnosis were identified. Patients with transfusion dependence were so defined if they had  3 RBC transfusion events within any 3-month period [24]. Definition of RBC transfusion event and the associated codes can be found in the Supplementary Table I to be found online at http://informahealthcare. com/doi/abs/10.3109/10428194.2015.1016933. Index date was defined as the date of the first RBC transfusion of the first TD episode after the first MF diagnosis. Patients had  6 months of continuous enrollment for both medical and pharmacy coverage prior to the index date. Figure 1 illustrates the study design scheme. Patients with TD MF were stratified based on their ICT utilization (with  1 claim of ICT since MF diagnosis vs. none). ICT codes can be found in the Supplementary Table II to be found online at http://informahealthcare.com/doi/abs/10.

3109/10428194.2015.1016933. The proportion of patients with TD MF receiving ICT was reported.

Outcome measures The number of RBC transfusion events per patient per year and the occurrence of RBC transfusion were captured for all patients with MF. The frequency of MF-related complications per patient per year was identified through ICD-9-CM codes. The list of MF-related complications and ICD-9-CM codes can be found in the Supplementary Table III to be found online at http://informahealthcare.com/doi/abs/10.3109/10428194. 2015.1016933. All-cause, MF-related and MF complicationrelated health RU included inpatient stays, emergency room (ER) visits and outpatient visits. Healthcare costs were broken down by all-cause, MF-related and MF complication-related costs. MF-related refers to events (i.e. RU and costs) that were associated with claims for MF diagnosis, MF complication, MF medication (azacitidine, busulfan, cladribine, danazol, decitabine, erythropoiesis-stimulating agents, hydroxyurea, immunomodulatory agents, interferon a, methylprednisone and ruxolitinib [Janus kinase (JAK)1/2 inhibitors]), surgical procedure (stem cell transplant, radiation therapy and splenectomy) or ICT, whereas MF complication-related refers to events associated with  1 claim for MF complication.

Statistical analysis Baseline descriptive statistics were summarized and compared between the two groups using c2 tests for categorical variables and two-sided t-tests for continuous variables. The Kaplan–Meier method was used to assess the time from first MF diagnosis to first RBC transfusion. Time to first RBC transfusion between TD ICT and ICT patients was compared using the log-rank test. The incidence rates of MF-related complications, allcause and MF-related RU were computed and compared between the two groups using adjusted incidence rate ratios (aIRRs). The aIRRs were estimated based on Poisson models controlling for baseline conditions found statistically different in the two groups, including Charlson Comorbidity Index (CCI, a weighted index including 17 conditions predictive of

Figure 1. Study design scheme.

Myelofibrosis  3 1-year mortality, was calculated during the 6-month baseline period prior to the index date) [25], prior history of essential thrombocythemia, anemia, acute leukemia, bleeding infections, leukocytosis and portal hypertension. Healthcare costs from the payer’s perspective were inflation-adjusted to 2012 US dollars (USD) and were computed using weighted mean cost per person per month (PPPM). To quantify the adjusted cost difference between TD ICT vs. ICT patients, linear regression models were used. All data manipulation and statistical analyses were performed using SAS release 9.3 (SAS Institute, Inc., Cary, NC). Additional details on methods are available in the Supplementary Material S4 to be found online at http://informahealthcare.com/ doi/abs/10.3109/10428194.2015.1016933.

Results Downloaded by [Universite Laval] at 18:12 05 November 2015

Baseline demographics and clinical characteristics A total of 571 TD patients with MF meeting eligibility criteria were included in the study, of whom 103 (18%) had received at least one ICT since MF diagnosis and 468 (82%) had never (Table I). The mean age was similar between patients with TD MF with vs. without ICT (mean [standard deviation, SD]: 67.2 [10.4] years vs. 66.6 [11.7] years; p  0.653). There were significantly more men among patients with ICT than those without ICT (percent of male: 70.9% vs. 58.3%; p  0.018). The proportions of patients with MF insured by different types of health plans were similar between patients with TD MF with vs. without ICT. The prevalence of the majority of comorbidities and complications assessed at baseline was similar between TD ICT and ICT patients (Table II). The proportions of patients in the ICT and ICT groups with prior history of myelodysplastic syndrome (MDS) prior to their first observed MF diagnosis during the baseline period were

69.9% ICT and 74.8% ICT patients (p  0.307). The only exceptions were essential thrombocythemia (ICT: 11.7% vs. ICT: 20.9%; p  0.030), hypothyroidism (5.8% vs. 13.0%; p  0.04), acute leukemia (2.9% vs. 13.7%; p  0.002), leukocytosis (2.9% vs. 12.0%; p  0.006), infections (1.0% vs. 8.5%; p  0.007) and portal hypertension (3.9% vs. 0.6%; p  0.023), all of which were significantly lower among ICT patients during the baseline period. Anemia was found to be significantly higher in the ICT vs. ICT group (86.4% vs. 77.4%; p  0.041). Additionally, the CCI was significantly lower among ICT patients (mean [SD]: 1.8 [1.8] vs. 2.3 [2.1]; p  0.006).

Blood transfusions The mean [SD] observation time was significantly longer for ICT patients than ICT patients (22.2 [13.9] months vs. 12.6 [11.6] months; p  0.001) (Table II). The mean (SD) number of RBC transfusion events per year was similar between the two groups, with 22.4 (19.5) events/year in ICT patients compared to 22.2 (28.5) events/year in ICT patients (p  0.943), although ICT patients progressed from MF diagnosis to RBC transfusion dependence significantly quicker than ICT patients (ICT: median 2.9 months vs. ICT: median 4.3 months; p  0.003) (Table III). Among ICT patients, therapy was initiated after a median time of 5.6 months following the first evidence of TD.

MF-related complications Patients with ICT had a significantly higher incidence rate of anemia, and lower rates of thrombocytopenia and pancytopenia (Table IV). After adjusting for potential confounders, the aIRR for ICT vs. ICT patients was 1.56 (95% confidence interval [CI]: 1.24–1.97; p  0.001) for anemia, 0.55 (0.40–0.75; p  0.001) for thrombocytopenia and 0.53 (0.37–0.76; p  0.001) for pancytopenia. The incidence rates

Table I. Baseline demographic characteristics of transfusion-dependent MF ICT and ICT patients. Characteristic Demographics Age, mean (SD) Age group, n (%)   20–49 years   50–64 years   65  years Male, n (%) Region, n (%)   Midwest/North Central   East/Northeast   South   West   Unknown Health plan, n (%)   HMO   POS   PPO   Other‡   Unknown

Transfusion-dependent ICT (n  103) 67.2

(10.4)

5 41 57 73

(4.9%) (39.8%) (55.3%) (70.9%)

34 13 29 27 0 7 3 68 25 0

Transfusion-dependent ICT (n  468) 66.6

p-Value†

(11.7)

0.653

25 195 248 273

(5.3%) (41.7%) (53.0%) (58.3%)

0.841 0.728 0.665 0.018*

(33.0%) (12.6%) (28.2%) (26.2%) (0.0%)

171 57 159 80 1

(36.5%) (12.2%) (34.0%) (17.1%) (0.2%)

0.499 0.902 0.255 0.032* 1.000

(6.8%) (2.9%) (66.0%) (24.3%) (0.0%)

32 36 278 115 7

(6.8%) (7.7%) (59.4%) (24.6%) (1.5%)

0.988 0.082 0.213 0.949 0.361

­ F, myelofibrosis; ICT, iron chelation therapy; HMO, health maintenance organization; POS, point of service; PPO, preferred M provider organization. *p-Value  0.05. †p-Value for comparison of transfusion-dependent patients with ICT vs. without ICT is based on c2 tests for categorical variables and t-tests for continuous variables. ‡Other includes extended provider organization, basic/comprehensive health plan, high deductible health plan, dental coverage, indemnity plan and multiple health plans.

4  F. Vekeman et al. Table II. Medical history of transfusion-dependent ICT and ICT patients with MF.

Downloaded by [Universite Laval] at 18:12 05 November 2015

Medical history Length of observation (months), mean (SD) Prior history of essential thrombocythemia, n (%)‡ Prior history of polycythemia vera, n (%)‡ Prior history of myelodysplastic syndrome, n (%)‡ Charlson Comorbidity Index, mean (SD) Complications, n (%)   Any complication   Anemia   Pain   Myelogenous leukemia   Complications of extramedullary hematopoiesis   Splenomegaly   Thrombocytopenia   Pancytopenia   Acute leukemia   Low grade myelodysplastic syndrome lesions   Leukocytosis   Hardening and inflammation of bone tissue and gout, gouty arthritis   Hypothyroidism   Bleeding   Infections   Thrombosis§   Lymphadenopathy   Hepatomegaly   Portal hypertension

Transfusion-dependent ICT (n  103) 22.2 12 12 72 1.8 102 89 43 42 41 29 28 22 3 12 3 10 6 0 1 6 4 4 4

Transfusion-dependent ICT (n  468)

p-Value†

(13.9) (11.7%) (11.7%) (69.9%) (1.8)

12.6 98 64 350 2.3

(11.6) (20.9%) (13.7%) (74.8%) (2.1)

0.030* 0.584 0.307 0.006*

(99.0%) (86.4%) (41.7%) (40.8%) (39.8%) (28.2%) (27.2%) (21.4%) (2.9%) (11.7%) (2.9%) (9.7%) (5.8%) (0.0%) (1.0%) (5.8%) (3.9%) (3.9%) (3.9%)

452 362 211 215 154 134 159 137 64 54 56 39 61 7 40 34 30 25 3

(96.6%) (77.4%) (45.1%) (45.9%) (32.9%) (28.6%) (34.0%) (29.3%) (13.7%) (11.5%) (12.0%) (8.3%) (13.0%) (1.5%) (8.5%) (7.3%) (6.4%) (5.3%) (0.6%)

0.333 0.041* 0.537 0.340 0.181 0.923 0.184 0.105 0.002* 0.974 0.006* 0.652 0.040* 0.361 0.007* 0.604 0.327 0.542 0.023*

­ F, myelofibrosis; ICT, iron chelation therapy. M *p-Value  0.05. †p-Value for comparison of transfusion-dependent patients with ICT vs. without ICT is based on c2 tests for categorical variables and t-tests for continuous variables. ‡Patients were identified for essential thrombocythemia, polycythemia vera or myelodysplastic syndrome prior to the MF diagnosis date. §Thrombosis included portal vein, other venous embolism and thrombosis of unspecified site, and arterial embolism and thrombosis.

of other MF-related complications considered were similar between the two groups.

All-cause and MF-related RU Among MF-related medications prescribed, more ICT patients received danazol (ICT: 10.7% vs. ICT: 3.6%, p  0.003), Epogen (ICT: 14.6% vs. ICT: 5.8%, p  0.002), Procrit (ICT: 18.4% vs. ICT: 9.0%, p  0.005), thalidomide (ICT: 15.5% vs. ICT: 6.6%, p  0.003) and methylprednisone (ICT: 17.5% vs. ICT: 9.2%, p  0.014) (Table V). No difference was observed among patients receiving ruxolitinib (ICT: 1.9% vs. ICT: 2.6%, p  1.000), general hematopoietic cell transplant (ICT: 1.0% vs. ICT: 0.6%, p  0.550), allogeneic hematopoietic stem cell transplant (ICT: 7.8% vs. ICT: 7.5%, p  0.920) and autologous hematopoietic stem cell transplant (ICT: 5.8% vs. ICT: 7.3%, p  0.604). Patients with ICT had significantly lower rates of all-cause and MF-related acute care but higher rates of outpatient visits (Table VI). In particular, ICT patients had a mean of

1.26 all-cause inpatient stays per person-year compared to 1.83 all-cause inpatient stays per person-year among ICT patients. After adjusting for potential confounders, the aIRR of inpatient stays for ICT patients vs. ICT patients was 0.75 (95% CI: 0.64–0.87; p  0.001). The mean all-cause inpatient days were 11.03 days per person-year in ICT patients compared to 22.52 days per person-year in ICT patients (aIRR: 0.52 [0.50–0.55]; p  0.001). The mean all-cause outpatient visits were 70.96 visits per person-year for ICT patients and 62.17 visits per person-year for ICT patients (aIRR: 1.21 [1.18–1.23]; p  0.001). The mean all-cause ER visits were 1.63 visits per person-year for ICT patients and 2.08 visits per person-year for ICT patients (aIRR: 0.84 [0.74–0.96]; p  0.011). Similar trends were observed for MF-related and MF complication-related RU. In an exploratory analysis assessing the medical conditions associated with outpatient visits, the distribution of primary diagnoses for these visits was similar between ICT and ICT patients. The most common primary reason for

Table III. Comparison of blood transfusions and median time to events during the study period. Transfusion-dependent ICT (n  103) Mean number of blood transfusions per year (SD) Time to events, median (range) in months‡   Time from first MF diagnosis to first blood transfusion   Time from first MF diagnosis to transfusion dependence   Time from first MF to first ICT   Time from first dependence on blood transfusion to first ICT

22.4 (19.5) 0.9 (0.03, 33.80) 2.9 (0.03, 33.80) 8.2 (0.07, 49.53) 5.6 ( 21.33, 34.87)

Transfusion-dependent ICT (n  468) 22.2 (28.5) 1.7 (0.03, 60.23) 4.3 (0.03, 60.23) NA NA

p-Value† 0.943 0.001* 0.003* NA NA

­ F, myelofibrosis; ICT, iron chelation therapy. M *p-Value  0.05. †p-Value for analysis of mean blood transfusions was obtained from Student’s t-test, whereas p-values for time to event analyses were obtained from log-rank tests. ‡NA stands for “not applicable,” which applied to the group that was not at risk for the event.

Myelofibrosis  5 Table IV. Adjusted rate of MF-related complications during the study period among transfusion-dependent ICT and ICT patients with MF. Transfusion-dependent ICT (n  103)

Complications, n (%)

Number of Personpatients with years of  1 event observation†

Downloaded by [Universite Laval] at 18:12 05 November 2015

[A] Anemia Pain Myelogenous leukemia Splenomegaly Complications of extramedullary hematopoiesis Thrombocytopenia Pancytopenia Low grade myelodysplastic syndrome lesions Hardening and inflammation of bone tissue and gout, gouty arthritis Acute leukemia Infections Leukocytosis Thrombosis¶ Hepatomegaly Lymphadenopathy Portal hypertension Bleeding

[B]

Incidence rate

Transfusion-dependent ICT (n  468) Number of Personpatients with years of  1 event observation

[A]/[B]  [C]

Incidence rate

[D]

[E]

[D]/[E]  [F]

Adjusted incidence rate ratio‡

(95% CI)

p-Value§

100 72 71

6.95 75.90 87.71

14.38 0.95 0.81

423 306 293

67.17 253.40 223.67

6.30 1.21 1.31

1.56 0.93 0.83

(1.24, 1.97)  0.001* (0.71, 1.22) 0.603 (0.63, 1.09) 0.173

47 61

110.81 90.19

0.42 0.68

174 188

338.56 321.11

0.51 0.59

0.80 1.25

(0.57, 1.12) (0.91, 1.71)

53 39 30

130.93 136.11 142.88

0.40 0.29 0.21

247 187 75

293.36 337.19 420.51

0.84 0.55 0.18

0.55 0.53 1.45

(0.40, 0.75)  0.001* (0.37, 0.76)  0.001* (0.92, 2.28) 0.105

20

161.75

0.12

57

434.26

0.13

0.91

(0.53, 1.56)

0.729

23 28 17 15 15 11 8 5

169.85 155.28 174.84 173.18 174.78 175.72 182.55 181.99

0.14 0.18 0.10 0.09 0.09 0.06 0.04 0.03

116 98 76 63 49 42 19 10

412.04 396.12 433.96 438.35 448.07 449.45 476.27 481.11

0.28 0.25 0.18 0.14 0.11 0.09 0.04 0.02

0.67 0.91 0.79 0.64 0.87 0.68 0.92 1.52

(0.42, 1.07) (0.59, 1.42) (0.45, 1.40) (0.35, 1.17) (0.47, 1.60) (0.34, 1.35) (0.39, 2.19) (0.49, 4.70)

0.092 0.686 0.424 0.151 0.654 0.269 0.852 0.463

0.191 0.172

I­ CT, iron chelation therapy; MF, myelofibrosis; CI, confidence interval. *p-Value  0.05. †Person-years of observation were rounded to the nearest hundredth. ‡The following baseline characteristics were controlled: Charlson Comorbidity Index, prior history of essential thrombocythemia, anemia, acute leukemia, bleeding and infections, leukocytosis and portal hypertension. §p-Value for comparison of incidence rates of complications in transfusion-dependent patients with ICT vs. without ICT is based on a Poisson distribution. ¶Thrombosis included portal vein, other venous embolism and thrombosis of unspecified site, and arterial embolism and thrombosis.

outpatient visits was unspecified MDS or MF (ICT: 34% of all outpatient visits vs. ICT: 29%), followed by anemia, which was observed in a slightly higher proportion of outpatient visits in ICT patients (ICT: 12% vs. ICT: 7%).

All-cause and MF-related costs All-cause and MF-related medical service costs were similar between ICT patients and ICT patients (Table VII). Pharmacy costs were significantly higher in ICT patients than in ICT patients. Mean (SD) all-cause pharmacy costs were estimated at $2464 ($265) per person per month for ICT patients compared to $1120 ($106) for ICT patients. After adjusting for potential confounders, the mean (SD) cost difference was $1426 ($359) (p  0.004). In investigating whether the mean cost difference may be driven by treatments for MF-related complications, a complication-specific analysis indicated that inpatient costs were significantly lower in ICT patients (mean [SD] adjusted cost difference: $1804 [$570], p  0.004).

Discussion Evidence on the clinical outcomes and RU of ICT among patients with TD MF is lacking to date. In the present study, ICT patients with TD MF were found to have similar clinical profiles to ICT patients. Rates of MF-related complications were generally similar between the two groups, although

ICT patients had significantly higher rates of anemia and lower rates of thrombocytopenia and pancytopenia as well as lower CCI. While RU for outpatient care was higher in ICT patients, the use of acute care, including ER and inpatient visits, was significantly lower among ICT patients. In terms of costs, while lower all-cause and MF-related costs were observed between ICT vs. ICT patients with TD MF, such differences were not statistically significant. Anemia is the primary indication for frequent blood RBC transfusions. Hence, it is expected that patients with greater RBC transfusion levels have higher rates of anemia. Additionally, as frequent RBC transfusion may lead to iron overload, ICT is more often used in patients who receive a greater volume of blood. However, in this study, ICT patients received a similar number of RBC transfusion events to ICT patients, seemingly suggesting that the requirement for RBC transfusion was not different between the two groups of patients. A few possibilities may explain the role of ICT in relation to the lower rates of selected complications and acute care use. One possibility may be the presence of a confounding by indication bias, where physicians select patients with better prognosis to receive ICT as is supported by lower CCI observed in ICT patients. Patients receiving ICT may be linked to a more aggressive track of medical care, as suggested by higher rates of outpatient visits in ICT patients in this study. Current guidelines for MF management do not have a clear recommendation on the use of ICT [3]. Based on a study by

6  F. Vekeman et al. Table V. Proportion of patients with MF-related medications and surgical procedures during the study period.

Downloaded by [Universite Laval] at 18:12 05 November 2015

Transfusion-dependent ICT (n  103) Medications, n (%)   Any medication   Azacitidine   Busulfan   Cladribine   Danazol   Decitabine   Erythropoiesis-stimulating agents    Darbepoetin alfa    Epoietin alfa (Epogen)    Epoietin alfa (Procrit)    Methoxy polyethylene glycolepoietin beta   Hydroxyurea   Immunomodulatory agents    Lenalidomide    Thalidomide    Pomalidomide   Interferon a   Methylprednisolone   Prednisone   Ruxolitinib Surgical procedures, n (%)   Any procedure   Transplant    General hematopoietic cell transplant    Allogeneic hematopoietic stem cell transplant    Autologous hematopoietic stem cell transplant   Radiation therapy    General radiation    Brachytherapy    Hyperthemia    Nuclear medicine therapy    Proton beam delivery   Splenectomy

Transfusion-dependent ICT (n  468)

p-Value†

82 8 0 0 11 6

(79.6%) (7.8%) (0.0%) (0.0%) (10.7%) (5.8%)

301 26 2 0 17 17

(64.3%) (5.6%) (0.4%) (0.0%) (3.6%) (3.6%)

0.003* 0.391 1.000 — 0.003* 0.279

16 15 19 0 13

(15.5%) (14.6%) (18.4%) (0.0%) (12.6%)

47 27 42 0 80

(10.0%) (5.8%) (9.0%) (0.0%) (17.1%)

0.107 0.002* 0.005* — 0.266

14 16 0 1 18 51 2

(13.6%) (15.5%) (0.0%) (1.0%) (17.5%) (49.5%) (1.9%)

41 31 0 1 43 185 12

(8.8%) (6.6%) (0.0%) (0.2%) (9.2%) (39.5%) (2.6%)

0.132 0.003* — 0.329 0.014* 0.063 1.000

29

(28.2%)

131

(28.0%)

0.973

1 8 6

(1.0%) (7.8%) (5.8%)

3 35 34

(0.6%) (7.5%) (7.3%)

0.550 0.920 0.604

13 0 0 0 0 8

(12.6%) (0.0%) (0.0%) (0.0%) (0.0%) (7.8%)

64 1 0 0 0 34

(13.7%) (0.2%) (0.0%) (0.0%) (0.0%) (7.3%)

0.777 1.000 — — — 0.860

­ F, myelofibrosis; ICT, iron chelation therapy. M *p-Value  0.05. †p-Value for comparison of transfusion-dependent patients with ICT vs. without ICT is based on c2 tests for categorical variables and t-tests for continuous variables.

Tefferi and colleagues, who have shown that hyperferritinemia is not linked to decreased survival in patients with MF, the British Committee for Standards in Haematology (BCSH)

suggests that ICT should not be used routinely [3,26]. Nonetheless, the use of ICT has been proposed more frequently as increased research suggests that ICT may be linked to

Table VI. Adjusted rate of resource utilization during the study period among transfusion-dependent ICT and ICT patients with MF. Transfusion-dependent ICT (n  103)

Transfusion-dependent ICT (n  468)

Number of events

Personyears of observation

Incidence rate

Number of events

Personyears of observation

Incidence rate

[A]

[B]

[A]/[B]  [C]

[D]

[E]

[D]/[E]  [F]

Adjusted incidence rate ratio†

188.59 188.59 188.59 188.59

1.63 1.26 11.03 70.96

1010 886 10 914 30 126

484.60 484.60 484.60 484.60

2.08 1.83 22.52 62.17

0.84 0.75 0.52 1.21

(0.74, 0.96) 0.011* (0.64, 0.87)  0.001* (0.50, 0.55)  0.001* (1.18, 1.23)  0.001*

188.59 188.59 188.59 188.59

0.87 1.15 10.44 49.31

566 811 10 332 20 703

484.60 484.60 484.60 484.60

1.17 1.67 21.32 42.72

0.81 0.76 0.55 1.20

(0.68, 0.97) 0.025* (0.65, 0.88)  0.001* (0.52, 0.58)  0.001* (1.17, 1.23)  0.001*

188.59 188.59 188.59 188.59

0.99 1.07 9.53 32.75

611 783 10 067 15 055

484.60 484.60 484.60 484.60

1.26 1.62 20.77 31.07

0.88 0.74 0.52 1.15

(0.74, 1.05) 0.148 (0.63, 0.87)  0.001* (0.49, 0.54)  0.001* (1.11, 1.18)  0.001*

All-cause resource utilization   ER visits 307   Inpatient stays 237   Inpatient days 2080   Outpatient visits 13 382 MF-related resource utilization   ER visits 165   Inpatient stays 216   Inpatient days 1968   Outpatient visits 9299 Complication-related resource    utilization   ER visits 186   Inpatient stays 202   Inpatient days 1798   Outpatient visits 6176

(95% CI)

p-Value‡

I­ CT, iron chelation therapy; MF, myelofibrosis; ER, emergency room; CI, confidence interval. *p-Value  0.05. †The following baseline characteristics were controlled for: Charlson Comorbidity Index, prior history of essential thrombocythemia, anemia, acute leukemia, bleeding and infections, leukocytosis and portal hypertension. ‡p-Value for comparison of incidence rates of complications in transfusion-dependent patients with ICT vs. without ICT is based on a Poisson distribution.

Myelofibrosis  7 Table VII. Unadjusted and adjusted mean difference in all-cause, MF-related and complication-related cost during the study period between transfusion-dependent ICT and ICT patients with MF. TransfusionTransfusiondependent ICT dependent ICT (n  103) (n  468)

Downloaded by [Universite Laval] at 18:12 05 November 2015

[A] All-cause cost (PPPM)§, mean (SD)   Total medical services cost $10 075    ER $95    Inpatient $4704    Outpatient $5291   Pharmacy $2464 MF-related cost (PPPM)§, mean (SD)   Total medical services cost $9045    ER $58    Inpatient $4387    Outpatient $4602   Pharmacy $2464 Complication-related cost (PPPM)§, mean (SD)   Total medical services cost $1689    ER $49    Inpatient $703    Outpatient $936

[B]

Unadjusted cost difference [A]  [B]

p-Value‡

Adjusted cost difference†

p-Value‡

($1261) ($15) ($933) ($503) ($265)

$12 539 $106 $7759 $4666 $1120

($924) ($12) ($759) ($293) ($106)

$2465 $11 $3055 $625 $1344

($1570) ($19) ($1204) ($584) ($275)

0.112 0.495 0.02 0.495 0.004

$545 $8 $1901 $1568 $1426

($1671) ($15) ($1253) ($695) ($359)

0.615 0.587 0.132 0.291 0.004*

($1,219) ($14) ($916) ($480) ($265)

$10 981 $71 $6898 $4006 $1120

($848) ($10) ($708) ($261) ($106)

$1936 $13 $2511 $596 $1344

($1497) ($17) ($1161) ($557) ($275)

0.18 0.423 0.04 0.363 0.004

$325 $13 $1457 $1398 $1426

($1602) ($12) ($1204) ($662) ($359)

0.794 0.295 0.224 0.120 0.004*

($296) ($10) ($209) ($135)

$4069 $63 $2566 $1440

($536) ($8) ($444) ($154)

$2380 $14 $1862 $504

($621) ($12) ($497) ($203)

0.004 0.216 0.004 0.024

$1911 $10 $1804 $268

($625) ($11) ($570) ($255)

0.004* 0.323 0.004* 0.259

ICT, iron chelation therapy; MF, myelofibrosis; ER, emergency room. *p-Value  0.05. †The mean adjusted difference in cost was based on a two-part weighted multivariate regression model adjusting for Charlson Comorbidity Index, prior history of essential thrombocythemia, anemia, acute leukemia, bleeding and infections, leukocytosis and portal hypertension. Weight was computed based on the person-time observed for each patient. ‡p-Values are estimated using a non-parametric bootstrap re-sampling technique with 500 iterations. §PPPM stands for “per patient per month.” PPPM was calculated by dividing the cost incurred over the observation period by the person-time in months observed for each patient.

improved erythropoiesis in patients with MF [18–20,27–29]. Similarly, the role of ICT is well established in other blood disorders such as thalassemia and sickle cell disease, and gaining acceptance in MDS [13–16,30–33]. According to the American Society of Hematology guidelines for MDS, ICT is recommended for patients with life expectancy greater than a year, and without comorbidities that would limit prognosis [34]. With no established treatment guidelines for the use of ICT in patients with TD MF, its prescription relies more heavily on the treating physician’s choice. Another explanation for the role of ICT in relation to lower rates of selected complications may lie in differences in gender, geographic region and socioeconomic status (SES). According to a study by Viprakasit et al., physician treatment practices, cost and patient reimbursement influenced ICT prescription rates [35]. In fact, significant differences were observed in the distributions of gender and geographic region between ICT and ICT patients in the present study. Thus, given the variance in clinical practice, ICT was more likely to be channeled toward patients with a better prognosis. While cost may have been a driving factor for how likely ICT was used in the two cohorts, such that patients at lower SES may have less access to ICT, the necessary information to assess SES impact was not possible in claims data. However, all patients included in this study were privately insured or had a component of private insurance (for Medicare patients), and thus the mean SES level was likely higher than that in the overall US population, as Medicaid patients were excluded. Nonetheless, conclusions regarding the SES between ICT and ICT patients could not be reached. Another possibility is that ICT is a marker for earlier and more regular interventions, which in turn may contribute

to improve the complication profile and lower utilization of acute care. The higher utilization rate of outpatient care observed in the present study among ICT vs. ICT patients corroborates the hypothesis of a possible link between closer monitoring and improved outcomes, especially given that such outpatient visits were linked to the MF condition itself, rather than adverse events associated with ICT. This suggests that the increased rate of outpatient visits was not a marker for worse prognosis, but rather potentially linked to the higher ICT use. Finally, although no causal inferences can be made on the basis of the data used in this retrospective study, ICT may also help to alleviate certain symptoms of MF, thus contributing to improve outcomes in this patient population. According to a retrospective medical chart review study by Leitch and colleagues on patients with MF TD, ICT was linked to lower iron levels, slower progression to acute myelogenous leukemia and prolonged OS [6]. While the association between ICT and clinical outcomes in patients with MF has been shown in retrospective and case series studies [6,18,19], and the role of ICT in improved survival is well established in other blood disorders such as thalassemia and is gaining acceptance in MDS [13,14,36–38], the impact of ICT on clinical outcomes in patients with MF awaits further evidence from randomized clinical trials. This study has a few notable limitations. First, the lack of complete clinical data limited the ability to fully assess a patient’s MF severity, such that the prognosis of MF between groups could not be confirmed and compared. Similarly, severity of certain manifestations of MF such as thrombocytopenia, anemia, myeloproliferative neoplasm subtype and risk group could not be determined with medi-

Downloaded by [Universite Laval] at 18:12 05 November 2015

8  F. Vekeman et al. cal claims databases. Possible heterogeneity of underlying conditions preceding MF diagnosis may exist. Nonetheless, other clinical characteristics, such as CCI and baseline ­MF-related complications, allowed for the assessment and comparisons of the patient groups. Baseline characteristics indicate that ICT vs. ICT patients were largely comparable when becoming TD. Second, our definition of patients with ICT utilization (ICT) was based on a low threshold (i.e. at least one claim of ICT since their MF diagnosis). As such, some heterogeneity was expected in this cohort. Based on our exploratory analysis, the mean number of ICTs received among this cohort was 10.2 (SD  13.0), with a median of 6.0 and an interquartile range of 2.0–14.0. Further, with a mean of 30 days of supply per oral ICT use, there was considerable variance in the total duration of ICT. This may be driven partly by patients who dropped out early due to side effects or other reasons and may have received fewer ICTs as a consequence. Third, consistent with a major limitation of claims data, the validity of clinical diagnosis relied on the accuracy with which the conditions were coded. Approximately 70% of patients with MF had a prior diagnosis of MDS, suggesting that patients with MF in the study sample may have a worse prognosis compared to patients with MF without MDS diagnosis, irrespective of ICT status. However, this proportion of patients with MDS was higher than expected; it may be an artifact of the non-specificity of ICD-9 codes which encompasses MDS and a series of conditions related to MDS, including MF diagnosis. Nonetheless, while a high rate of patients in the study sample had prior history of MDS, the distribution of MDS between comparison cohorts (ICT vs. ICT) was not different. Furthermore, MF-related complications may have been misclassified or underreported. Nonetheless, misclassification is expected to be non-differential between comparator groups, such that results reported may be biased toward the null. Lastly, confounding by indication may be present and not fully accounted for by the adjustment of baseline comorbidity and complication in regression models. The reason for residual confounding may be due to the classification of patients with ICT, which was based on any evidence of the administration of ICT before or after the baseline period. Furthermore, as aforementioned, physicians may have been more likely to administer ICT to patients with a better prognosis.

Conclusions The present study provides an overview of the real-world complications and RU in ICT vs. ICT patients with TD MF, and serves as one of the first studies to highlight the role of ICT in this patient population. Yet, due to the nonrandomized nature of the study and possible heterogeneity underlying the conditions preceding MF diagnosis, the potential short-term and long-term benefits of ICT in MF cannot be validated with the present study and require further investigations with prospective clinical trials. Future studies would benefit from the assessment of disease severity of patients with MF, thereby providing evidence on any potential differential benefit of ICT by MF severity.­­­­

Acknowledgement This work was supported by Novartis Pharmaceuticals Corporation, East Hanover, New Jersey, USA. Potential conflict of interest:  Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.

References [1]  Hellmann A. Myeloproliferative syndromes: diagnosis and therapeutic options. Polskie Archiwum Medycyny Wewnetrznej 2008;118:756–760. [2]  Emanuel RM, Dueck AC, Geyer HL, et  al. Myeloproliferative neoplasm (MPN) symptom assessment form total symptom score: prospective international assessment of an abbreviated symptom burden scoring system among patients with MPNs. J Clin Oncol 2012;30:4098–4103. [3]  Reilly JT, McMullin MF, Beer PA, et al. Guideline for the diagnosis and management of myelofibrosis. Br J Haematol 2012;158:453–471. [4]  Abdel-Wahab OI, Levine RL. Primary myelofibrosis: update on definition, pathogenesis, and treatment. Annu Rev Med 2009;60: 233–245. [5]  Rollison DE, Howlader N, Smith MT, et  al. Epidemiology of myelodysplastic syndromes and chronic myeloproliferative disorders in the United States, 2001-2004, using data from the NAACCR and SEER programs. Blood 2008;112:45–52. [6]  Leitch HA, Chase JM, Goodman TA, et  al. Improved survival in red blood cell transfusion dependent patients with primary myelofibrosis (PMF) receiving iron chelation therapy. Hematol Oncol 2010;28:40–48. [7]  Brown K, Subramony C, May W, et  al. Hepatic iron overload in children with sickle cell anemia on chronic transfusion therapy. J Pediatr Hematol Oncol 2009;31:309–312. [8]  Walter PB, Harmatz P, Vichinsky E. Iron metabolism and iron chelation in sickle cell disease. Acta Haematol 2009;122:174–183. [9]  Darbari DS, Kple-Faget P, Kwagyan J, et al. Circumstances of death in adult sickle cell disease patients. Am J Hematol 2006;81:858–863. [10]  Olivieri NF. Progression of iron overload in sickle cell disease. Semin Hematol 2001;38(Suppl. 1):57–62. [11]  Ballas SK. Iron overload is a determinant of morbidity and mortality in adult patients with sickle cell disease. Semin Hematol 2001;38(Suppl. 1):30–36. [12]  Harmatz P, Butensky E, Quirolo K, et al. Severity of iron overload in patients with sickle cell disease receiving chronic red blood cell transfusion therapy. Blood 2000;96:76–79. [13]  Brittenham GM, Griffith PM, Nienhuis AW, et  al. Efficacy of deferoxamine in preventing complications of iron overload in patients with thalassemia major. N Engl J Med 1994;331:567–573. [14]  Davis BA, Porter JB. Long-term outcome of continuous 24-hour deferoxamine infusion via indwelling intravenous catheters in highrisk beta-thalassemia. Blood 2000;95:1229–1236. [15]  Daar S, Pathare A, Nick H, et al. Reduction in labile plasma iron during treatment with deferasirox, a once-daily oral iron chelator, in heavily iron-overloaded patients with beta-thalassaemia. Eur J Haematol 2009;82:454–457. [16]  Taher A, El-Beshlawy A, Elalfy MS, et  al. Efficacy and safety of deferasirox, an oral iron chelator, in heavily iron-overloaded patients with beta-thalassaemia: the ESCALATOR study. Eur J Haematol 2009;82:458–465. [17]  Harvey RD. Myelodysplastic syndromes and the role of iron overload. Am J Health Syst Pharm 2010;67(Suppl. 2):S3–S9. [18]  Tesch H, Ihling C. Loss of transfusion dependency following deferasirox treatment of iron overload in a woman with myelofibrosis and spherocytosis - a case report. Onkologie 2013;36:205–208. [19]  Di Tucci AA, Murru R, Alberti D, et  al. Correction of anemia in a transfusion-dependent patient with primary myelofibrosis receiving iron chelation therapy with deferasirox (Exjade, ICL670). Eur J Haematol 2007;78:540–542. [20]  Lisette del C, Enrico B, Eleonora A, et  al. Mayor erythropoietic response after deferasirox treatment in a transfusion-dependent anemic patient with primary myelofibrosis. Case Rep Hematol 2013;2013:520712.

Downloaded by [Universite Laval] at 18:12 05 November 2015

Myelofibrosis  9 [21]  Mehta J, Wang H, Fryzek JP, et  al. Health resource utilization and cost associated with myeloproliferative neoplasms in a large United States health plan. Leuk Lymphoma 2014;55:2368–2374. [22]  Truven Health Analytics MarketScan Research Database. Available from: http://truvenhealth.com/your-healthcare-focus/ analytic-research/marketscan-research-databases [23]  IMS LifeLink PharMetrics Plus™ – U.S. Available from: hwww. imshealth.com/deployedfiles/imshealth/Global/Content/Home%20 Page%20Content/Real-World%20insights/IMS_Lifelink_pharmetrics_ Plus.pdf [24]  Gale RP, Barosi G, Barbui T, et  al. What are RBC-transfusiondependence and -independence? Leuk Res 2011;35:8–11. [25]  Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987;40:373–383. [26]  Tefferi A, Pardanani A, Lim KH, et al. TET2 mutations and their clinical correlates in polycythemia vera, essential thrombocythemia and myelofibrosis. Leukemia 2009;23:905–911. [27]  Marsh JH, Hundert M, Schulman P. Deferoxamine-induced restoration of haematopoiesis in myelofibrosis secondary to myelodysplasia. Br J Haematol 1990;76:148–149. [28]  Messa E, Cilloni D, Messa F, et al. Deferasirox treatment improved the hemoglobin level and decreased transfusion requirements in four patients with the myelodysplastic syndrome and primary myelofibrosis. Acta Haematol 2008;120:70–74. [29]  Smeets ME, Vreugdenhil G, Holdrinet RS. Improvement of erythropoiesis during treatment with deferiprone in a patient with myelofibrosis and transfusional hemosiderosis. Am J Hematol 1996;51:243–244.

Supplementary material available online Appendix Tables showing ICD-9 codes, and further study design and methods. Supplementary Appendix Tables I–III.

[30]  Gattermann N, Finelli C, Della Porta M, et  al. Hematologic responses to deferasirox therapy in transfusion-dependent patients with myelodysplastic syndromes. Haematologica 2012;97: 1364–1371. [31]  Rose C, Brechignac S, Vassilief D, et  al. Does iron chelation therapy improve survival in regularly transfused lower risk MDS patients? A multicenter study by the GFM (Groupe Francophone des Myelodysplasies). Leuk Res 2010;34:864–870. [32]  Cohen AR, Martin MB. Iron chelation therapy in sickle cell disease. Semin Hematol 2001;38(Suppl. 1):69–72. [33]  Inati A, Khoriaty E, Musallam KM, et  al. Iron chelation therapy for patients with sickle cell disease and iron overload. Am J Hematol 2010;85:782–786. [34]  Bennett JM; MDS Foundation’s Working Group on Transfusional Iron Overload. Consensus statement on iron overload in myelodysplastic syndromes. Am J Hematol 2008;83:858–861. [35]  Viprakasit V, Gattermann N, Lee JW, et al. Geographical variations in current clinical practice on transfusions and iron chelation therapy across various transfusion-dependent anaemias. Blood Transfus 2013;11:108–122. [36]  Leitch HA. Improving clinical outcome in patients with myelodysplastic syndrome and iron overload using iron chelation therapy. Leuk Res 2007;31(Suppl. 3):S7–S9. [37]  Rose C, Brechignac S, Vassilief D, et  al. Positive impact of iron chelation therapy (CT) on survival in regularly transfused MDS patients. A prospective analysis by the GFM. Blood 2007;110:80A–81A. [38]  Olivieri NF, Nathan DG, MacMillan JH, et al. Survival in medically treated patients with homozygous beta-thalassemia. N Engl J Med 1994;331:574–578.

Medical complications, resource utilization and costs in patients with myelofibrosis by frequency of blood transfusion and iron chelation therapy.

Iron chelation therapies (ICTs) can help eliminate iron surplus in erythrocyte transfusion-dependent (TD) patients with myelofibrosis (MF). The study ...
604KB Sizes 0 Downloads 8 Views