Current Medical Research & Opinion
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
0300-7995 doi:10.1185/03007995.2014.915802
Vol. 30, No. 8, 2014, 1461–1471
Article FT-0123.R1/915802 All rights reserved: reproduction in whole or part not permitted
Original article Healthcare resource use and relapses with fingolimod versus natalizumab for treating multiple sclerosis: a retrospective US claims database analysis
Niklas Bergvall Raquel Lahoz Tracy Reynolds Novartis Pharma AG, Basel, Switzerland
Jonathan R. Korn
Abstract Objective: Healthcare resource utilization in patients with multiple sclerosis (MS) is linked to relapses and disease progression. This retrospective cohort database analysis compared healthcare resource use and proxy measures of relapse outcomes in patients with active disease who switched to fingolimod or natalizumab.
IMS Health, Waltham, MA, USA Address for correspondence: Dr. Niklas Bergvall, Novartis Pharma AG, Postfach, CH-4002 Basel, Switzerland. Tel.: +41 61 324 03 84; Fax: +41 61 696 80 19; [email protected] Keywords: Administrative claims – Effectiveness – Fingolimod – Healthcare resource – Multiple sclerosis – Natalizumab – Relapses Accepted: 14 April 2014; published online: 2 May 2014 Citation: Curr Med Res Opin 2014; 30:1461–71
Methods: Using administrative claims data from the US PharMetrics Plus database, we identified patients with an MS diagnosis and a claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012 (index period) who had experienced a relapse (identified using a claims-based algorithm) and used other diseasemodifying therapies (DMTs) in the previous year. Patients in the fingolimod and natalizumab cohorts were propensity score matched (1:1). MS-related inpatient stays, corticosteroid use and the proportion of patients experiencing claims-based relapses were assessed in the pre-index and post-index persistence periods. Time to first claims-based relapse in the post-index persistence period was assessed using a Kaplan–Meier curve. Results: The study included 623 unmatched patients (299 and 324 patients in the fingolimod and natalizumab cohorts, respectively) and 370 matched patients (185 in each cohort). In the matched analysis, MS-related inpatient stays and corticosteroid use were similar in the fingolimod and natalizumab cohorts during the post-index persistence period, and were significantly reduced versus the pre-index period (p50.01). A similar proportion of patients in the fingolimod and natalizumab cohorts were free from claims-based relapses in the persistence period (68.1% and 68.6%, respectively). There was no significant difference in the likelihood of experiencing a claims-based relapse (p ¼ 0.8696). Limitation: Identification of relapses is based on database claims rather than on clinical assessment. Conclusions: In analyses of patients with MS with a history of relapse and DMT use, fingolimod and natalizumab reduce healthcare resource utilization and have similar effectiveness in a real-world setting.
Introduction Several new therapies for the treatment of multiple sclerosis (MS) have been approved for use in the past 10 years1. All disease-modifying therapies (DMTs) ! 2014 Informa UK Ltd www.cmrojournal.com
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1461
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
August 2014
for this chronic, neurodegenerative disease aim to reduce the number of relapses (acute attacks of neurological dysfunction) and delay the progression to irreversible disability. The new therapies have different mechanisms of action and modes of administration compared with conventional therapies1, such as interferons and glatiramer acetate2–4, which have been in routine use for over two decades. Interferons and glatiramer acetate, which are administered by either subcutaneous or intramuscular injection, are recommended as first-line treatments for MS, but these therapies are increasingly perceived to have partial effectiveness and a suboptimal tolerability profile1. Fingolimod (Gilenya*), a sphingosine 1-phosphate receptor modulator, was the first oral therapy approved for the treatment of MS. It was approved by the US Food and Drug Administration (FDA)5 in September 2010, and was approved in the EU in March 2011 for the treatment of patients with high disease activity despite previous treatment with interferon beta or with rapidly evolving severe relapsing–remitting MS (RRMS)6. Natalizumab (Tysabriy), a recombinant alpha 4 integrin antagonist administered by monthly infusions in an outpatient setting, received initial approval in the USA in 2004 and was approved in the EU in 2006, and has the same licensed indications as fingolimod7,8. There have been no head-to-head clinical studies comparing the efficacy of fingolimod and natalizumab. Data from phase 3, randomized, controlled trials (RCTs) have demonstrated the efficacy of these therapies over placebo and interferon across clinical and magnetic resonance imaging (MRI; used to detect central nervous system lesions) endpoints9–13. Fingolimod was assessed in three phase 3 trials: the 24-month FREEDOMS (FTY720 Research Evaluating Effects of Daily Oral Therapy in Multiple Sclerosis)11 and FREEDOMS II9 trials, and the 12-month, active-controlled TRANSFORMS (Trial Assessing Injectable Interferon versus FTY720 Oral in Relapsing–Remitting MS)10. These trials showed that fingolimod 0.5 mg delays confirmed disability progression, as well as significantly reduces relapses and improves MRI and brain atrophy outcomes compared with placebo or intramuscular interferon beta-1a (Avonexz)9–11. The efficacy of natalizumab for reducing disability progression, relapse rates and MRI endpoints relative to placebo was demonstrated in the 24-month AFFIRM (Natalizumab Safety and Efficacy in Relapsing–Remitting Multiple Sclerosis) trial12. The 24-month SENTINEL (Safety and Efficacy of Natalizumab in Combination with Interferon beta-1a in Patients with Relapsing–Remitting MS) trial demonstrated the efficacy of natalizumab in combination *Gilenya is a registered trade name of Novartis Pharma AG, Basel, Switzerland. yTysabri is a registered trade name of Biogen Idec, Cambridge, MA, USA. zAvonex is a registered trade name of Biogen Idec, Cambridge, MA, USA.
1462
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
with intramuscular interferon beta-1a (Avonex) compared with interferon beta-1a monotherapy13. The disabling nature and lifelong impact of MS imposes a considerable burden on healthcare systems. Burdenof-illness studies conducted in Europe and North America have demonstrated that disease progression to advanced states of disability14–18 and relapses14,15 are associated with increased healthcare resource utilization, and direct and indirect costs. The symptoms of mild relapses can typically be treated with corticosteroids (administered either orally or intravenously), while treatment of severe relapses might involve admission to hospital19. Relapses requiring hospitalization and/or corticosteroid treatment put a greater burden on healthcare resources than relapses not requiring these interventions. Several studies have shown that relapse severity is a factor influencing the resource utilization associated with MS14,15. Patients with high relapse activity (two or more relapses per year) also had increased resource utilization20. DMTs that reduce the frequency of relapses regardless of disease severity and slow the progression to disability could reduce healthcare resource use and the socioeconomic burden of the disease. Administrative claims databases provide a robust source of evidence on healthcare resource use and real-world insights into the comparative effectiveness of therapies for MS. Evidence from these databases, which reflect real-world clinical experiences, can be used to complement data from clinical trials. MS relapses are not recorded directly in administrative claims databases, but can be identified using diagnosis codes and claims that have been shown to correlate with the occurrence of clinical relapses21–24. Retrospective database studies assessing these proxy measures of relapses have been valuable in confirming that switching to more efficacious therapies in patients experiencing breakthrough disease is clinically superior to cycling between first-line injectable DMTs in a real-world clinical practice setting24,25. One 12-month outpatient observational study has compared the realworld comparative effectiveness of fingolimod and natalizumab as second-line therapies, and results suggest that the clinical effectiveness of both treatments is similar with regard to stabilizing disability progression and improving relapse rates26. We have conducted a retrospective realworld analysis using a US health insurance claims database to compare healthcare resource use and proxy measures of relapses in patients with active disease who switched to fingolimod or natalizumab.
Patients and methods Data source This study was a retrospective cohort analysis of administrative health plan claims data from the PharMetrics Plus database. This database, which contains adjudicated www.cmrojournal.com ! 2014 Informa UK Ltd
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
medical and pharmacy claims for over 87 million health plan members, has been described in detail elsewhere23,24. By having integrated claims data (i.e. data from several places of services [inpatient, outpatient and pharmacy]) for over 100,000 patients with MS in 2011, the database is considered to be a valuable source of real-world data on a commercially insured population of patients with MS in the USA. The database recorded inpatient and outpatient diagnoses and procedures (using International Classification of Diseases, Ninth Revision, Clinical Modification [ICD9-CM] codes), patient demographics, dates of the first and last healthcare plan enrollment, product type and payer type, dates of service, retail/mail-order prescription records, detailed information on pharmacy and medical benefit, and inpatient stay and provider details. The dates of claims were also available for services provided, along with the amount charged by the provider and the amounts allowed and paid for by healthcare plans. For hospital stays, detailed diagnosis and procedural ICD-9-CM codes were recorded but not the names and descriptions of medications, including DMTs. For outpatient visits, medications and DMTs prescribed were recorded. For full details, see Bergvall et al. 2013 and 201423,24. The PharMetrics Plus database meets all Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliance standards regarding the protection of humans. As patient information was de-identified to maintain anonymity, approval from the Institutional Review Board and patient-informed consent were not necessary. The study was designed, implemented and reported in line with the Guidelines for Good Pharmacoepidemiology Practices (GPP) of the International Society for Pharmacoepidemiology27, the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines28, and with the ethical principles described in the Declaration of Helsinki. Access to the PharMetrics Plus database requires a licensing agreement under which open access to the data used in this study is not permitted.
Patient selection Patients with a diagnosis of MS (ICD-9-CM code 340), who had at least one prescription or administration claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012 (index period) were identified in the database. Claims for fingolimod or natalizumab were identified using National Drug Codes (NDCs) (Supplementary table 1) for prescriptions or through procedural codes (Supplementary table 2) for DMTs administered to patients in a clinical setting. The index date was the date of the first prescription fill or DMT administration. For inclusion in the study, patients were required to meet all of the following criteria: at least 360 days of continuous ! 2014 Informa UK Ltd www.cmrojournal.com
August 2014
Patients with at least one prescription/ administration claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012, and 24 months of continuous enrollment (12 months pre- and post-index) (n=3523)
Excluded (n=2900) • No evidence of at least one pre- or post-index MS diagnosis (n=72) • Less than 18 years of age at index (n=7) • Receipt of more than one DMT at index (n=3) • Receipt of index DMT in the pre-index period (n=1375) • Data quality issues (invalid enrollment dates, age, gender, coverage information) (n=7) • Missing data for index therapy (days’ supply on NDC claim) (n=8) • No pre-index claims-based relapse (n=1141) • No pre-index DMT use (n=287)
Unmatched fingolimod cohort (n=299)
Unmatched natalizumab cohort (n=324)
Patients dropped during propensity score matching (n=253)
Matched fingolimod cohort (n=185)
Matched natalizumab cohort (n=185)
Figure 1. Attrition of the study sample, by reason. Patients were propensity-score matched within strata (number of claims-based relapses and the presence of non-index DMT in the pre-index period) based on criteria such as age, gender, prescribing physician specialty, prior use of dalfampridine, presence of a pre-index claims-based relapse within 90 days of the index date, presence and timing of MRI use, and the number of pre-index corticosteroid claims. DMT, disease-modifying therapy; MS, multiple sclerosis; NDC, National Drug Code.
pre-index date health plan enrollment; at least 360 days of continuous post-index date enrollment; at least one MS diagnosis code (ICD-9-CM code 340) in either the preindex or post-index period; and at least 18 years of age at index date. Additionally, to ensure inclusion of those with active disease, eligible patients had to have experienced at least one MS-related claims-based relapse (see definition below) and used DMTs (Supplementary table 1) in the pre-index period (Figure 1). Patients were excluded from the study if they met any of the following criteria: receipt of more than one DMT on the index date; receipt of fingolimod or natalizumab in the 360-day pre-index period; data quality issues (invalid enrollment date, incomplete claims Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1463
Current Medical Research & Opinion Volume 30, Number 8
August 2014
data, missing/invalid age and/or gender data); or had missing days’ supply in claims data in the post-index period for the index therapy. Patients meeting the inclusion criteria were assigned to one of two unmatched cohorts depending on their index DMT: fingolimod or natalizumab cohort.
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Propensity score matching Analyses were also performed on matched cohorts of patients. Patients in the fingolimod cohort were randomly matched (1:1) to patients in the natalizumab cohort using a propensity score matching approach similar to that used by Bergvall et al.23,29. To ensure that patients receiving fingolimod were directly matched on specific characteristics to those receiving natalizumab, a logistic regression model was created to match patients within strata (number of claims-based relapses in the pre-index period [1 versus 2þ claimsbased relapses]). Use of fingolimod therapy was the dependent variable. The independent variables, established as of the index date and during the pre-index period, were entered stepwise into the model and included characteristics such as age, gender, prescribing physician specialty, prior use of dalfampridine, presence of a preindex claims-based relapse within 90 days of the index date, presence and timing of MRI use, and the number of corticosteroid claims. Pairs of patients with similar propensity scores were matched using the ‘nearest neighbor’ approach, discarding cases defined by a minimal difference (e.g. 0.01) from the fitted probability of DMT use.
Study measures Patient demographics and baseline clinical characteristics, which were evaluated during the pre-index period, were recorded for matched and unmatched cohorts of patients. The baseline characteristics that were assessed included prescribing physician specialty for index DMT, Charlson comorbidity index (CCI) score, number of unique medications, MS-related symptoms, previous use of dalfampridine and comorbidities of interest (e.g. dyslipidemia, depression, diabetes mellitus, history of cardiovascular disease [CVD]). Healthcare resource use was evaluated during the preindex and post-index period. Parameters assessed included the proportion of patients requiring MS-related inpatient stays or MS-related emergency department visits, and the number of MRI claims or MS-related outpatient claims (which included natalizumab infusion). MS-related resource use required a primary MS diagnosis or a claim for a DMT of interest. The mean and median number of days’ supply for corticosteroid use (all, oral and intravenous) was calculated for the total number of patients in each cohort, not just those patients receiving corticosteroids. Resource use in the post-index persistence period was 1464
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
annualized to enable direct comparison with resource use in the 360 day pre-index period. As there are no specific codes to directly record relapses in administrative databases, an algorithm using diagnosis codes and claims that has been shown to correlate with the occurrence of clinical relapses23,26,30,31 was used to identify relapses in this study. Relapses identified using this algorithm can be considered a proxy measure of relapses and are termed claims-based relapses to distinguish them from relapses reported in clinical trials. The algorithm defined an inpatient claims-based relapse as an inpatient visit with an ICD-9-CM code 340 in the primary position. An outpatient claims-based relapse was defined as an outpatient visit with an ICD-9-CM code 340 and oral or intravenous corticosteroid use within 7 days of the outpatient visit. Only outpatient visits without a primary exclusionary diagnosis code (e.g. asthma, uveitis, gout and rheumatoid arthritis) were included in the analyses. Claims-based relapse events that occurred within the same 30-day period were treated as a single claims-based relapse. To determine the effectiveness of fingolimod and natalizumab, the proportion of patients experiencing claimsbased relapses and the time to claims-based relapse in the post-index period while persistent with medication (termed the persistence period) was assessed. Persistence was measured as the time in consecutive days from initiation of the index DMT until discontinuation, receipt of another DMT or the end of the study, whichever occurred first. Discontinuation was defined as a gap in exposure to the index DMT of at least 60 days following the date when the index DMT should next have been dispensed or administered.
Statistical analyses For categorical variables, data are presented as number of cases and proportion of total patients observed in each category. For continuous variables, data were summarized using the mean (95% confidence interval [CI]), standard deviation (SD) and median. For healthcare resource utilization, the mean and median number of claims/visits/days’ supply was measured. Statistical significance (p50.05) of the differences in the distribution of these variables was calculated using the McNemar test and the Wilcoxon signed-rank test for categorical and continuous variables, respectively. The statistical significance of differences in healthcare resource use between the pre-index and postindex persistence period (annualized) was also calculated for each healthcare resource parameter for both matched and unmatched cohorts. To calculate the statistical significance of differences in healthcare resources utilization between treatment cohorts in the post-index persistence period generalized estimating equations (GEEs) were used with a binary distribution and logit link for inpatient and www.cmrojournal.com ! 2014 Informa UK Ltd
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
emergency department visits, and a negative binomial distribution and log link for corticosteroid days’ supply. To calculate the likelihood of experiencing a claimsbased relapse while persistent with index DMT a GEE with a binary distribution and logit link function was used for the matched cohorts. The dependent variable was the presence of relapse while persistent and the offset variable was the log of the number of years on therapy. Time to first claims-based relapse (in days) while persistent with the index DMT in the matched cohorts was estimated during the post-index persistence period using a Kaplan–Meier analysis, with separate survival curves for the fingolimod and natalizumab cohorts. The time to claims-based relapse among those with a relapse was also assessed. The probability of experiencing a claims-based relapse over time was calculated based on the number of patients still being followed through the post-index period. Patients were followed until claims-based relapse, discontinuation of index therapy or the end of the available data period (360 days post-index), whichever occurred first. The statistical significance of the differences between curves in the matched analyses was assessed using a logrank test.
August 2014
included in the study were female and the median age of patients at index date was 44 years in both cohorts. Owing to the matched nature of the cohorts, frequencies of MS symptoms and comorbidities in the pre-index period were similar in both treatment groups. This was with the exception of swallowing problems, which were more common in the natalizumab cohort than in the fingolimod cohort (p ¼ 0.0064) although this affected less than 5% of the total patients. Pre-index use of dalfampridine was similar in both cohorts, but CCI scores, which are a general measure of clinical comorbidity, were significantly higher in the fingolimod cohort than in the natalizumab cohort; 37% of patients in the fingolimod cohort had a CCI score of at least 1 compared with 27% in the natalizumab cohort (p ¼ 0.0426; mean scores: 0.65 and 0.49, respectively). The characteristics of patients in the unmatched cohorts are also provided in Table 1; there were significant differences between the treatment cohorts for median age, healthcare costs, CCI scores and for several symptoms and comorbidities, including numbness, walking, balance and coordination, and bladder dysfunction (all p50.05).
Persistence with fingolimod and natalizumab
Results Study attrition In total, 3523 patients were initially identified as having at least one claim for the prescription or administration of fingolimod or natalizumab between 1 October 2010 and 30 June 2012, with 24 months continuous enrollment (12 months pre-index and 12 months post-index). To include patients with active disease who switched to fingolimod or natalizumab, our study excluded patients with no evidence of previous DMT use (n ¼ 287; 3.6%) or no claims-based relapses (n ¼ 1141; 14.4%) in the pre-index period. Overall, 623 patients met all the inclusion/exclusion criteria, with 299 and 324 patients in the unmatched fingolimod and natalizumab cohorts, respectively. During propensity score matching, a further 253 patients were dropped due to no available match with a similar propensity score, with some of these being excluded to allow for an equal number of patients in each cohort (Figure 1). The final matched analyses included 370 patients, with 185 patients in each of the fingolimod and natalizumab cohorts.
Pre-index patient demographics and clinical characteristics Pre-index demographics and clinical characteristics of patients in the matched fingolimod and natalizumab cohorts are shown in Table 1. Three-quarters of patients ! 2014 Informa UK Ltd www.cmrojournal.com
During the entire 360-day post-index period, a similar proportion of patients in both the fingolimod and natalizumab cohorts were persistent with their medication (matched: 71.9% and 76.2%, respectively, p ¼ 0.3427; unmatched: 72.9% and 76.9%, respectively, p ¼ 0.2565). The length of the persistence period was also similar between the treatment cohorts (matched: mean SD, 297 112 days and 301 113 days, respectively; unmatched: mean SD, 297 113 days and 304 109 days, respectively).
Healthcare resource utilization Inpatient stays and emergency department visits During the pre-index period, 9.2% and 8.6% of patients in the matched fingolimod and natalizumab cohorts, respectively, required an MS-related inpatient admission (Figure 2a). In the post-index persistence period, the corresponding proportions were 2.7% and 1.1%, respectively (fingolimod: 71% reduction versus pre-index period, p ¼ 0.0027; natalizumab: 87% reduction versus pre-index period, p ¼ 0.0002; Figure 2a), and there was no significant difference between the treatment cohorts in the postindex persistence period (p ¼ 0.2620). Similar results were seen for the unmatched fingolimod and natalizumab cohorts (Figure 2a), with the corresponding reductions in admissions in pre-index versus post-index persistence periods being 82% (p50.0001) and 84% (p50.0001), respectively. In the matched analyses, a higher proportion of patients in each treatment cohort visited an emergency department Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1465
Current Medical Research & Opinion Volume 30, Number 8
August 2014
Table 1. Demographic and clinical characteristics of patients in matched and unmatched fingolimod and natalizumab cohorts in the pre-index period.
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Characteristic
Age, years Mean SD Median Female, n (%) Previous use of dalfampridine, n (%) Prior use of non-index DMT, n (%) Glatiramer acetate Interferon Charlson comorbidity index score, n (%) 0 1 2–3 4 Symptoms affecting 10% of patients, n (%) Fatigue Walking (gait), balance, and coordination problems Numbness Headache Visual symptoms Muscle weakness/spasm/spasticity Bladder dysfunction Dizziness and vertigo Respiration and breathing problems Bowel dysfunction Comorbidities affecting 5% of patients, n (%) Dyslipidemia Depression Tobacco use (including disorder) Diabetes mellitus History of CVD No. of pre-index claims-based relapses, mean SD Patients experiencing an outpatient claims-based relapse in the pre-index perioda, n (%) Patients experiencing an inpatient claims-based relapse in the pre-index perioda, n (%) Healthcare costs, US$ Total, mean SD Median
Matched Fingolimod (n ¼ 185)
Natalizumab (n ¼ 185)
44.0 9.8 44.0 140 (75.7) 30 (16.2) 185 (100) 87 (47.0) 106 (57.3)
43.5 9.9 44.0 139 (75.1) 33 (17.8) 185 (100) 76 (41.1) 124 (67.0)
116 (62.7) 30 (16.2) 38 (20.5) 1 (0.5)
135 (73.0) 24 (13.0) 22 (11.9) 4 (2.2)
71 (38.4) 53 (28.6) 52 (28.1) 46 (24.9) 33 (17.8) 30 (16.2) 27 (14.6) 26 (14.1) 22 (11.9) 18 (9.7)
Unmatched p Value
Fingolimod (n ¼ 299)
Natalizumab (n ¼ 324)
44.4 9.7 45.0 231 (77.3) 50 (16.7) 299 (100) 134 (44.8) 180 (60.2)
42.4 9.8 43.0 229 (70.7) 40 (12.3) 324 (100) 142 (43.8) 210 (64.8)
0.0426
190 (63.5) 51 (17.1) 56 (18.7) 2 (0.7)
236 (72.8) 39 (12.0) 40 (12.3) 9 (2.8)
0.0053
66 (35.7) 46 (24.9) 55 (29.7) 51 (27.6) 28 (15.1) 35 (18.9) 31 (16.8) 22 (11.9) 22 (11.9) 26 (14.1)
0.5904 0.4111 0.7308 0.5545 0.4836 0.4946 0.5673 0.5360 1.0000 0.1988
100 (33.4) 66 (22.1) 74 (24.7) 76 (25.4) 56 (18.7) 50 (16.7) 39 (13.0) 44 (14.7) 39 (13.0) 30 (10.0)
117 (36.1) 108 (33.3) 107 (33.0) 100 (30.9) 67 (20.7) 62 (19.1) 61 (18.8) 52 (16.0) 32 (9.9) 44 (13.6)
0.4853 0.0018 0.0230 0.1314 0.5413 0.4332 0.0494 0.6451 0.2140 0.1716
44 (23.8) 43 (23.2) 18 (9.7) 14 (7.6) 12 (6.5) 1.56 0.93 171 (92.4)
36 (19.5) 44 (23.8) 16 (8.6) 10 (5.4) 12 (6.5) 1.52 0.85 175 (94.6)
0.3124 0.9024 0.7189 0.3985 1.0000
68 (21.0) 82 (25.3) 28 (8.6) 20 (6.2) 24 (7.4) 1.65 0.91 307 (94.8)
0.2245 0.3965 0.5504 0.9255 0.7721
0.3985
75 (25.1) 67 (22.4) 30 (10.0) 19 (6.4) 24 (8.0) 1.52 0.91 278 (93.0)
17 (9.2)
16 (8.6)
0.8553
28 (9.4)
32 (9.9)
0.8287
44,158 19,401 42,753
47,325 28,072 45,293
0.3240
42,128 19,375 40,360
47,079 25,618 45,146
0.0067
0.7376 0.9039 0.6782
p Value
0.0176 0.0620 0.1206
0.3546
a Among those patients who had a claims-based relapse; percentages do not add up to 100% as some patients experienced both inpatient and outpatient visits. CVD, cardiovascular disease; DMT, disease-modifying therapy; SD, standard deviation.
in the pre-index period than in the post-index period (fingolimod: 58% reduction versus pre-index period, p ¼ 0.1573; natalizumab: 69% reduction versus pre-index period, p ¼ 0.0201; Figure 2b), and there was no significant difference between the treatment cohorts in the postindex persistence period (p ¼ 0.7215). A similar reduction in the number of visits in the pre-index versus post-index persistence periods was seen for the unmatched cohorts (fingolimod: 70% reduction, p ¼ 0.0027; natalizumab: 66% reduction, p ¼ 0.0007; Figure 2b).
Corticosteroid days’ supply During the pre-index period, the number of days’ supply of all corticosteroids (Figure 2c), oral corticosteroids (Figure 2d) and intravenous corticosteroids (Figure 2e) was similar for the matched fingolimod and natalizumab 1466
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
treatment cohorts. In the post-index persistence period, the annualized number of days’ supply was significantly reduced from the pre-index period for all, oral and intravenous corticosteroids for the matched cohorts (p50.0001 for all corticosteroid parameters; Figures 2c–e). Similar results were also seen for the unmatched fingolimod and natalizumab treatment cohorts (Figures 2c–e).
Other resource use For the matched cohorts, the number of MS-related outpatient claims was similar in the pre-index period (11 claims in each treatment cohort). In the post-index persistence period, the annualized median number of MS-related outpatient claims in the matched fingolimod and natalizumab cohorts was 8 and 19 claims, respectively. www.cmrojournal.com ! 2014 Informa UK Ltd
Current Medical Research & Opinion Volume 30, Number 8
Pre-index period
Patients with MS-related inpatient stays (%)
p=0.0027 10
9.2
Patients with MS-related emergency department visits (%) Mean number of days’ supply for all corticosteroids
p
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
0300-7995 doi:10.1185/03007995.2014.915802
Vol. 30, No. 8, 2014, 1461–1471
Article FT-0123.R1/915802 All rights reserved: reproduction in whole or part not permitted
Original article Healthcare resource use and relapses with fingolimod versus natalizumab for treating multiple sclerosis: a retrospective US claims database analysis
Niklas Bergvall Raquel Lahoz Tracy Reynolds Novartis Pharma AG, Basel, Switzerland
Jonathan R. Korn
Abstract Objective: Healthcare resource utilization in patients with multiple sclerosis (MS) is linked to relapses and disease progression. This retrospective cohort database analysis compared healthcare resource use and proxy measures of relapse outcomes in patients with active disease who switched to fingolimod or natalizumab.
IMS Health, Waltham, MA, USA Address for correspondence: Dr. Niklas Bergvall, Novartis Pharma AG, Postfach, CH-4002 Basel, Switzerland. Tel.: +41 61 324 03 84; Fax: +41 61 696 80 19; [email protected] Keywords: Administrative claims – Effectiveness – Fingolimod – Healthcare resource – Multiple sclerosis – Natalizumab – Relapses Accepted: 14 April 2014; published online: 2 May 2014 Citation: Curr Med Res Opin 2014; 30:1461–71
Methods: Using administrative claims data from the US PharMetrics Plus database, we identified patients with an MS diagnosis and a claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012 (index period) who had experienced a relapse (identified using a claims-based algorithm) and used other diseasemodifying therapies (DMTs) in the previous year. Patients in the fingolimod and natalizumab cohorts were propensity score matched (1:1). MS-related inpatient stays, corticosteroid use and the proportion of patients experiencing claims-based relapses were assessed in the pre-index and post-index persistence periods. Time to first claims-based relapse in the post-index persistence period was assessed using a Kaplan–Meier curve. Results: The study included 623 unmatched patients (299 and 324 patients in the fingolimod and natalizumab cohorts, respectively) and 370 matched patients (185 in each cohort). In the matched analysis, MS-related inpatient stays and corticosteroid use were similar in the fingolimod and natalizumab cohorts during the post-index persistence period, and were significantly reduced versus the pre-index period (p50.01). A similar proportion of patients in the fingolimod and natalizumab cohorts were free from claims-based relapses in the persistence period (68.1% and 68.6%, respectively). There was no significant difference in the likelihood of experiencing a claims-based relapse (p ¼ 0.8696). Limitation: Identification of relapses is based on database claims rather than on clinical assessment. Conclusions: In analyses of patients with MS with a history of relapse and DMT use, fingolimod and natalizumab reduce healthcare resource utilization and have similar effectiveness in a real-world setting.
Introduction Several new therapies for the treatment of multiple sclerosis (MS) have been approved for use in the past 10 years1. All disease-modifying therapies (DMTs) ! 2014 Informa UK Ltd www.cmrojournal.com
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1461
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
August 2014
for this chronic, neurodegenerative disease aim to reduce the number of relapses (acute attacks of neurological dysfunction) and delay the progression to irreversible disability. The new therapies have different mechanisms of action and modes of administration compared with conventional therapies1, such as interferons and glatiramer acetate2–4, which have been in routine use for over two decades. Interferons and glatiramer acetate, which are administered by either subcutaneous or intramuscular injection, are recommended as first-line treatments for MS, but these therapies are increasingly perceived to have partial effectiveness and a suboptimal tolerability profile1. Fingolimod (Gilenya*), a sphingosine 1-phosphate receptor modulator, was the first oral therapy approved for the treatment of MS. It was approved by the US Food and Drug Administration (FDA)5 in September 2010, and was approved in the EU in March 2011 for the treatment of patients with high disease activity despite previous treatment with interferon beta or with rapidly evolving severe relapsing–remitting MS (RRMS)6. Natalizumab (Tysabriy), a recombinant alpha 4 integrin antagonist administered by monthly infusions in an outpatient setting, received initial approval in the USA in 2004 and was approved in the EU in 2006, and has the same licensed indications as fingolimod7,8. There have been no head-to-head clinical studies comparing the efficacy of fingolimod and natalizumab. Data from phase 3, randomized, controlled trials (RCTs) have demonstrated the efficacy of these therapies over placebo and interferon across clinical and magnetic resonance imaging (MRI; used to detect central nervous system lesions) endpoints9–13. Fingolimod was assessed in three phase 3 trials: the 24-month FREEDOMS (FTY720 Research Evaluating Effects of Daily Oral Therapy in Multiple Sclerosis)11 and FREEDOMS II9 trials, and the 12-month, active-controlled TRANSFORMS (Trial Assessing Injectable Interferon versus FTY720 Oral in Relapsing–Remitting MS)10. These trials showed that fingolimod 0.5 mg delays confirmed disability progression, as well as significantly reduces relapses and improves MRI and brain atrophy outcomes compared with placebo or intramuscular interferon beta-1a (Avonexz)9–11. The efficacy of natalizumab for reducing disability progression, relapse rates and MRI endpoints relative to placebo was demonstrated in the 24-month AFFIRM (Natalizumab Safety and Efficacy in Relapsing–Remitting Multiple Sclerosis) trial12. The 24-month SENTINEL (Safety and Efficacy of Natalizumab in Combination with Interferon beta-1a in Patients with Relapsing–Remitting MS) trial demonstrated the efficacy of natalizumab in combination *Gilenya is a registered trade name of Novartis Pharma AG, Basel, Switzerland. yTysabri is a registered trade name of Biogen Idec, Cambridge, MA, USA. zAvonex is a registered trade name of Biogen Idec, Cambridge, MA, USA.
1462
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
with intramuscular interferon beta-1a (Avonex) compared with interferon beta-1a monotherapy13. The disabling nature and lifelong impact of MS imposes a considerable burden on healthcare systems. Burdenof-illness studies conducted in Europe and North America have demonstrated that disease progression to advanced states of disability14–18 and relapses14,15 are associated with increased healthcare resource utilization, and direct and indirect costs. The symptoms of mild relapses can typically be treated with corticosteroids (administered either orally or intravenously), while treatment of severe relapses might involve admission to hospital19. Relapses requiring hospitalization and/or corticosteroid treatment put a greater burden on healthcare resources than relapses not requiring these interventions. Several studies have shown that relapse severity is a factor influencing the resource utilization associated with MS14,15. Patients with high relapse activity (two or more relapses per year) also had increased resource utilization20. DMTs that reduce the frequency of relapses regardless of disease severity and slow the progression to disability could reduce healthcare resource use and the socioeconomic burden of the disease. Administrative claims databases provide a robust source of evidence on healthcare resource use and real-world insights into the comparative effectiveness of therapies for MS. Evidence from these databases, which reflect real-world clinical experiences, can be used to complement data from clinical trials. MS relapses are not recorded directly in administrative claims databases, but can be identified using diagnosis codes and claims that have been shown to correlate with the occurrence of clinical relapses21–24. Retrospective database studies assessing these proxy measures of relapses have been valuable in confirming that switching to more efficacious therapies in patients experiencing breakthrough disease is clinically superior to cycling between first-line injectable DMTs in a real-world clinical practice setting24,25. One 12-month outpatient observational study has compared the realworld comparative effectiveness of fingolimod and natalizumab as second-line therapies, and results suggest that the clinical effectiveness of both treatments is similar with regard to stabilizing disability progression and improving relapse rates26. We have conducted a retrospective realworld analysis using a US health insurance claims database to compare healthcare resource use and proxy measures of relapses in patients with active disease who switched to fingolimod or natalizumab.
Patients and methods Data source This study was a retrospective cohort analysis of administrative health plan claims data from the PharMetrics Plus database. This database, which contains adjudicated www.cmrojournal.com ! 2014 Informa UK Ltd
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
medical and pharmacy claims for over 87 million health plan members, has been described in detail elsewhere23,24. By having integrated claims data (i.e. data from several places of services [inpatient, outpatient and pharmacy]) for over 100,000 patients with MS in 2011, the database is considered to be a valuable source of real-world data on a commercially insured population of patients with MS in the USA. The database recorded inpatient and outpatient diagnoses and procedures (using International Classification of Diseases, Ninth Revision, Clinical Modification [ICD9-CM] codes), patient demographics, dates of the first and last healthcare plan enrollment, product type and payer type, dates of service, retail/mail-order prescription records, detailed information on pharmacy and medical benefit, and inpatient stay and provider details. The dates of claims were also available for services provided, along with the amount charged by the provider and the amounts allowed and paid for by healthcare plans. For hospital stays, detailed diagnosis and procedural ICD-9-CM codes were recorded but not the names and descriptions of medications, including DMTs. For outpatient visits, medications and DMTs prescribed were recorded. For full details, see Bergvall et al. 2013 and 201423,24. The PharMetrics Plus database meets all Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliance standards regarding the protection of humans. As patient information was de-identified to maintain anonymity, approval from the Institutional Review Board and patient-informed consent were not necessary. The study was designed, implemented and reported in line with the Guidelines for Good Pharmacoepidemiology Practices (GPP) of the International Society for Pharmacoepidemiology27, the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines28, and with the ethical principles described in the Declaration of Helsinki. Access to the PharMetrics Plus database requires a licensing agreement under which open access to the data used in this study is not permitted.
Patient selection Patients with a diagnosis of MS (ICD-9-CM code 340), who had at least one prescription or administration claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012 (index period) were identified in the database. Claims for fingolimod or natalizumab were identified using National Drug Codes (NDCs) (Supplementary table 1) for prescriptions or through procedural codes (Supplementary table 2) for DMTs administered to patients in a clinical setting. The index date was the date of the first prescription fill or DMT administration. For inclusion in the study, patients were required to meet all of the following criteria: at least 360 days of continuous ! 2014 Informa UK Ltd www.cmrojournal.com
August 2014
Patients with at least one prescription/ administration claim for fingolimod or natalizumab between 1 October 2010 and 30 June 2012, and 24 months of continuous enrollment (12 months pre- and post-index) (n=3523)
Excluded (n=2900) • No evidence of at least one pre- or post-index MS diagnosis (n=72) • Less than 18 years of age at index (n=7) • Receipt of more than one DMT at index (n=3) • Receipt of index DMT in the pre-index period (n=1375) • Data quality issues (invalid enrollment dates, age, gender, coverage information) (n=7) • Missing data for index therapy (days’ supply on NDC claim) (n=8) • No pre-index claims-based relapse (n=1141) • No pre-index DMT use (n=287)
Unmatched fingolimod cohort (n=299)
Unmatched natalizumab cohort (n=324)
Patients dropped during propensity score matching (n=253)
Matched fingolimod cohort (n=185)
Matched natalizumab cohort (n=185)
Figure 1. Attrition of the study sample, by reason. Patients were propensity-score matched within strata (number of claims-based relapses and the presence of non-index DMT in the pre-index period) based on criteria such as age, gender, prescribing physician specialty, prior use of dalfampridine, presence of a pre-index claims-based relapse within 90 days of the index date, presence and timing of MRI use, and the number of pre-index corticosteroid claims. DMT, disease-modifying therapy; MS, multiple sclerosis; NDC, National Drug Code.
pre-index date health plan enrollment; at least 360 days of continuous post-index date enrollment; at least one MS diagnosis code (ICD-9-CM code 340) in either the preindex or post-index period; and at least 18 years of age at index date. Additionally, to ensure inclusion of those with active disease, eligible patients had to have experienced at least one MS-related claims-based relapse (see definition below) and used DMTs (Supplementary table 1) in the pre-index period (Figure 1). Patients were excluded from the study if they met any of the following criteria: receipt of more than one DMT on the index date; receipt of fingolimod or natalizumab in the 360-day pre-index period; data quality issues (invalid enrollment date, incomplete claims Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1463
Current Medical Research & Opinion Volume 30, Number 8
August 2014
data, missing/invalid age and/or gender data); or had missing days’ supply in claims data in the post-index period for the index therapy. Patients meeting the inclusion criteria were assigned to one of two unmatched cohorts depending on their index DMT: fingolimod or natalizumab cohort.
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Propensity score matching Analyses were also performed on matched cohorts of patients. Patients in the fingolimod cohort were randomly matched (1:1) to patients in the natalizumab cohort using a propensity score matching approach similar to that used by Bergvall et al.23,29. To ensure that patients receiving fingolimod were directly matched on specific characteristics to those receiving natalizumab, a logistic regression model was created to match patients within strata (number of claims-based relapses in the pre-index period [1 versus 2þ claimsbased relapses]). Use of fingolimod therapy was the dependent variable. The independent variables, established as of the index date and during the pre-index period, were entered stepwise into the model and included characteristics such as age, gender, prescribing physician specialty, prior use of dalfampridine, presence of a preindex claims-based relapse within 90 days of the index date, presence and timing of MRI use, and the number of corticosteroid claims. Pairs of patients with similar propensity scores were matched using the ‘nearest neighbor’ approach, discarding cases defined by a minimal difference (e.g. 0.01) from the fitted probability of DMT use.
Study measures Patient demographics and baseline clinical characteristics, which were evaluated during the pre-index period, were recorded for matched and unmatched cohorts of patients. The baseline characteristics that were assessed included prescribing physician specialty for index DMT, Charlson comorbidity index (CCI) score, number of unique medications, MS-related symptoms, previous use of dalfampridine and comorbidities of interest (e.g. dyslipidemia, depression, diabetes mellitus, history of cardiovascular disease [CVD]). Healthcare resource use was evaluated during the preindex and post-index period. Parameters assessed included the proportion of patients requiring MS-related inpatient stays or MS-related emergency department visits, and the number of MRI claims or MS-related outpatient claims (which included natalizumab infusion). MS-related resource use required a primary MS diagnosis or a claim for a DMT of interest. The mean and median number of days’ supply for corticosteroid use (all, oral and intravenous) was calculated for the total number of patients in each cohort, not just those patients receiving corticosteroids. Resource use in the post-index persistence period was 1464
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
annualized to enable direct comparison with resource use in the 360 day pre-index period. As there are no specific codes to directly record relapses in administrative databases, an algorithm using diagnosis codes and claims that has been shown to correlate with the occurrence of clinical relapses23,26,30,31 was used to identify relapses in this study. Relapses identified using this algorithm can be considered a proxy measure of relapses and are termed claims-based relapses to distinguish them from relapses reported in clinical trials. The algorithm defined an inpatient claims-based relapse as an inpatient visit with an ICD-9-CM code 340 in the primary position. An outpatient claims-based relapse was defined as an outpatient visit with an ICD-9-CM code 340 and oral or intravenous corticosteroid use within 7 days of the outpatient visit. Only outpatient visits without a primary exclusionary diagnosis code (e.g. asthma, uveitis, gout and rheumatoid arthritis) were included in the analyses. Claims-based relapse events that occurred within the same 30-day period were treated as a single claims-based relapse. To determine the effectiveness of fingolimod and natalizumab, the proportion of patients experiencing claimsbased relapses and the time to claims-based relapse in the post-index period while persistent with medication (termed the persistence period) was assessed. Persistence was measured as the time in consecutive days from initiation of the index DMT until discontinuation, receipt of another DMT or the end of the study, whichever occurred first. Discontinuation was defined as a gap in exposure to the index DMT of at least 60 days following the date when the index DMT should next have been dispensed or administered.
Statistical analyses For categorical variables, data are presented as number of cases and proportion of total patients observed in each category. For continuous variables, data were summarized using the mean (95% confidence interval [CI]), standard deviation (SD) and median. For healthcare resource utilization, the mean and median number of claims/visits/days’ supply was measured. Statistical significance (p50.05) of the differences in the distribution of these variables was calculated using the McNemar test and the Wilcoxon signed-rank test for categorical and continuous variables, respectively. The statistical significance of differences in healthcare resource use between the pre-index and postindex persistence period (annualized) was also calculated for each healthcare resource parameter for both matched and unmatched cohorts. To calculate the statistical significance of differences in healthcare resources utilization between treatment cohorts in the post-index persistence period generalized estimating equations (GEEs) were used with a binary distribution and logit link for inpatient and www.cmrojournal.com ! 2014 Informa UK Ltd
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Current Medical Research & Opinion Volume 30, Number 8
emergency department visits, and a negative binomial distribution and log link for corticosteroid days’ supply. To calculate the likelihood of experiencing a claimsbased relapse while persistent with index DMT a GEE with a binary distribution and logit link function was used for the matched cohorts. The dependent variable was the presence of relapse while persistent and the offset variable was the log of the number of years on therapy. Time to first claims-based relapse (in days) while persistent with the index DMT in the matched cohorts was estimated during the post-index persistence period using a Kaplan–Meier analysis, with separate survival curves for the fingolimod and natalizumab cohorts. The time to claims-based relapse among those with a relapse was also assessed. The probability of experiencing a claims-based relapse over time was calculated based on the number of patients still being followed through the post-index period. Patients were followed until claims-based relapse, discontinuation of index therapy or the end of the available data period (360 days post-index), whichever occurred first. The statistical significance of the differences between curves in the matched analyses was assessed using a logrank test.
August 2014
included in the study were female and the median age of patients at index date was 44 years in both cohorts. Owing to the matched nature of the cohorts, frequencies of MS symptoms and comorbidities in the pre-index period were similar in both treatment groups. This was with the exception of swallowing problems, which were more common in the natalizumab cohort than in the fingolimod cohort (p ¼ 0.0064) although this affected less than 5% of the total patients. Pre-index use of dalfampridine was similar in both cohorts, but CCI scores, which are a general measure of clinical comorbidity, were significantly higher in the fingolimod cohort than in the natalizumab cohort; 37% of patients in the fingolimod cohort had a CCI score of at least 1 compared with 27% in the natalizumab cohort (p ¼ 0.0426; mean scores: 0.65 and 0.49, respectively). The characteristics of patients in the unmatched cohorts are also provided in Table 1; there were significant differences between the treatment cohorts for median age, healthcare costs, CCI scores and for several symptoms and comorbidities, including numbness, walking, balance and coordination, and bladder dysfunction (all p50.05).
Persistence with fingolimod and natalizumab
Results Study attrition In total, 3523 patients were initially identified as having at least one claim for the prescription or administration of fingolimod or natalizumab between 1 October 2010 and 30 June 2012, with 24 months continuous enrollment (12 months pre-index and 12 months post-index). To include patients with active disease who switched to fingolimod or natalizumab, our study excluded patients with no evidence of previous DMT use (n ¼ 287; 3.6%) or no claims-based relapses (n ¼ 1141; 14.4%) in the pre-index period. Overall, 623 patients met all the inclusion/exclusion criteria, with 299 and 324 patients in the unmatched fingolimod and natalizumab cohorts, respectively. During propensity score matching, a further 253 patients were dropped due to no available match with a similar propensity score, with some of these being excluded to allow for an equal number of patients in each cohort (Figure 1). The final matched analyses included 370 patients, with 185 patients in each of the fingolimod and natalizumab cohorts.
Pre-index patient demographics and clinical characteristics Pre-index demographics and clinical characteristics of patients in the matched fingolimod and natalizumab cohorts are shown in Table 1. Three-quarters of patients ! 2014 Informa UK Ltd www.cmrojournal.com
During the entire 360-day post-index period, a similar proportion of patients in both the fingolimod and natalizumab cohorts were persistent with their medication (matched: 71.9% and 76.2%, respectively, p ¼ 0.3427; unmatched: 72.9% and 76.9%, respectively, p ¼ 0.2565). The length of the persistence period was also similar between the treatment cohorts (matched: mean SD, 297 112 days and 301 113 days, respectively; unmatched: mean SD, 297 113 days and 304 109 days, respectively).
Healthcare resource utilization Inpatient stays and emergency department visits During the pre-index period, 9.2% and 8.6% of patients in the matched fingolimod and natalizumab cohorts, respectively, required an MS-related inpatient admission (Figure 2a). In the post-index persistence period, the corresponding proportions were 2.7% and 1.1%, respectively (fingolimod: 71% reduction versus pre-index period, p ¼ 0.0027; natalizumab: 87% reduction versus pre-index period, p ¼ 0.0002; Figure 2a), and there was no significant difference between the treatment cohorts in the postindex persistence period (p ¼ 0.2620). Similar results were seen for the unmatched fingolimod and natalizumab cohorts (Figure 2a), with the corresponding reductions in admissions in pre-index versus post-index persistence periods being 82% (p50.0001) and 84% (p50.0001), respectively. In the matched analyses, a higher proportion of patients in each treatment cohort visited an emergency department Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
1465
Current Medical Research & Opinion Volume 30, Number 8
August 2014
Table 1. Demographic and clinical characteristics of patients in matched and unmatched fingolimod and natalizumab cohorts in the pre-index period.
Curr Med Res Opin Downloaded from informahealthcare.com by Georgian Court University on 04/24/15 For personal use only.
Characteristic
Age, years Mean SD Median Female, n (%) Previous use of dalfampridine, n (%) Prior use of non-index DMT, n (%) Glatiramer acetate Interferon Charlson comorbidity index score, n (%) 0 1 2–3 4 Symptoms affecting 10% of patients, n (%) Fatigue Walking (gait), balance, and coordination problems Numbness Headache Visual symptoms Muscle weakness/spasm/spasticity Bladder dysfunction Dizziness and vertigo Respiration and breathing problems Bowel dysfunction Comorbidities affecting 5% of patients, n (%) Dyslipidemia Depression Tobacco use (including disorder) Diabetes mellitus History of CVD No. of pre-index claims-based relapses, mean SD Patients experiencing an outpatient claims-based relapse in the pre-index perioda, n (%) Patients experiencing an inpatient claims-based relapse in the pre-index perioda, n (%) Healthcare costs, US$ Total, mean SD Median
Matched Fingolimod (n ¼ 185)
Natalizumab (n ¼ 185)
44.0 9.8 44.0 140 (75.7) 30 (16.2) 185 (100) 87 (47.0) 106 (57.3)
43.5 9.9 44.0 139 (75.1) 33 (17.8) 185 (100) 76 (41.1) 124 (67.0)
116 (62.7) 30 (16.2) 38 (20.5) 1 (0.5)
135 (73.0) 24 (13.0) 22 (11.9) 4 (2.2)
71 (38.4) 53 (28.6) 52 (28.1) 46 (24.9) 33 (17.8) 30 (16.2) 27 (14.6) 26 (14.1) 22 (11.9) 18 (9.7)
Unmatched p Value
Fingolimod (n ¼ 299)
Natalizumab (n ¼ 324)
44.4 9.7 45.0 231 (77.3) 50 (16.7) 299 (100) 134 (44.8) 180 (60.2)
42.4 9.8 43.0 229 (70.7) 40 (12.3) 324 (100) 142 (43.8) 210 (64.8)
0.0426
190 (63.5) 51 (17.1) 56 (18.7) 2 (0.7)
236 (72.8) 39 (12.0) 40 (12.3) 9 (2.8)
0.0053
66 (35.7) 46 (24.9) 55 (29.7) 51 (27.6) 28 (15.1) 35 (18.9) 31 (16.8) 22 (11.9) 22 (11.9) 26 (14.1)
0.5904 0.4111 0.7308 0.5545 0.4836 0.4946 0.5673 0.5360 1.0000 0.1988
100 (33.4) 66 (22.1) 74 (24.7) 76 (25.4) 56 (18.7) 50 (16.7) 39 (13.0) 44 (14.7) 39 (13.0) 30 (10.0)
117 (36.1) 108 (33.3) 107 (33.0) 100 (30.9) 67 (20.7) 62 (19.1) 61 (18.8) 52 (16.0) 32 (9.9) 44 (13.6)
0.4853 0.0018 0.0230 0.1314 0.5413 0.4332 0.0494 0.6451 0.2140 0.1716
44 (23.8) 43 (23.2) 18 (9.7) 14 (7.6) 12 (6.5) 1.56 0.93 171 (92.4)
36 (19.5) 44 (23.8) 16 (8.6) 10 (5.4) 12 (6.5) 1.52 0.85 175 (94.6)
0.3124 0.9024 0.7189 0.3985 1.0000
68 (21.0) 82 (25.3) 28 (8.6) 20 (6.2) 24 (7.4) 1.65 0.91 307 (94.8)
0.2245 0.3965 0.5504 0.9255 0.7721
0.3985
75 (25.1) 67 (22.4) 30 (10.0) 19 (6.4) 24 (8.0) 1.52 0.91 278 (93.0)
17 (9.2)
16 (8.6)
0.8553
28 (9.4)
32 (9.9)
0.8287
44,158 19,401 42,753
47,325 28,072 45,293
0.3240
42,128 19,375 40,360
47,079 25,618 45,146
0.0067
0.7376 0.9039 0.6782
p Value
0.0176 0.0620 0.1206
0.3546
a Among those patients who had a claims-based relapse; percentages do not add up to 100% as some patients experienced both inpatient and outpatient visits. CVD, cardiovascular disease; DMT, disease-modifying therapy; SD, standard deviation.
in the pre-index period than in the post-index period (fingolimod: 58% reduction versus pre-index period, p ¼ 0.1573; natalizumab: 69% reduction versus pre-index period, p ¼ 0.0201; Figure 2b), and there was no significant difference between the treatment cohorts in the postindex persistence period (p ¼ 0.7215). A similar reduction in the number of visits in the pre-index versus post-index persistence periods was seen for the unmatched cohorts (fingolimod: 70% reduction, p ¼ 0.0027; natalizumab: 66% reduction, p ¼ 0.0007; Figure 2b).
Corticosteroid days’ supply During the pre-index period, the number of days’ supply of all corticosteroids (Figure 2c), oral corticosteroids (Figure 2d) and intravenous corticosteroids (Figure 2e) was similar for the matched fingolimod and natalizumab 1466
Resource use and relapses with fingolimod versus natalizumab Bergvall et al.
treatment cohorts. In the post-index persistence period, the annualized number of days’ supply was significantly reduced from the pre-index period for all, oral and intravenous corticosteroids for the matched cohorts (p50.0001 for all corticosteroid parameters; Figures 2c–e). Similar results were also seen for the unmatched fingolimod and natalizumab treatment cohorts (Figures 2c–e).
Other resource use For the matched cohorts, the number of MS-related outpatient claims was similar in the pre-index period (11 claims in each treatment cohort). In the post-index persistence period, the annualized median number of MS-related outpatient claims in the matched fingolimod and natalizumab cohorts was 8 and 19 claims, respectively. www.cmrojournal.com ! 2014 Informa UK Ltd
Current Medical Research & Opinion Volume 30, Number 8
Pre-index period
Patients with MS-related inpatient stays (%)
p=0.0027 10
9.2
Patients with MS-related emergency department visits (%) Mean number of days’ supply for all corticosteroids
p
