With interest, we read the article by Planté-Bordeneuve et al. published in Expert Opinion on Pharmacotherapy in November 2018 [Citation1]. Transthyretin amyloid polyneuropathy (ATTR-PN) is an inherited, progressive, and irreversible disease [Citation2]. In the absence of direct head-to-head trials comparing the ATTR-PN drugs patisiran and tafamidis, Planté-Bordeneuve and colleagues conducted an indirect, quantitative comparison of the efficacy of these two drugs in patients with ATTR-PN based on the APOLLO and Fx-005 trials, respectively [Citation1]. The authors concluded that patisiran had a greater treatment effect than tafamidis across several outcome measures. However, their findings have important methodological shortcomings that must be addressed.
First, large imbalances in baseline patient characteristics between APOLLO and Fx-005 exist that can impact the quality and interpretability of results obtained in an indirect quantitative comparison. The APOLLO population was, on average, 21.4 years older than the Fx-005 population, with a 28% greater proportion of males and a 265% greater mean Neuropathy Impairment Score of the Lower Limbs (NIS-LL) [Citation1,Citation3,Citation4]. Furthermore, all patients enrolled in Fx-005 exhibited the Val30Met mutation in the TTR gene and were classified as early-stage, and at the time of enrolment, were ATTR-PN pharmacotherapy-naïve [Citation4]. Conversely, there were 39 different mutations represented in the APOLLO study (ie. not limited to Val30Met). The prevalence, pathogenesis and prognosis of ATTR-PN depend on the specific mutation in the TTR gene [Citation2]. Additionally, patients at any disease stage were enrolled in APOLLO, and more than half of the enrolled patients were previously treated with ATTR-PN pharmacotherapy. A checklist of good research practices published by the ISPOR Task Force on indirect treatment comparisons (ITCs) states that differences in baseline characteristics may threaten the validity of indirect comparisons if these characteristics are treatment effect modifiers [Citation5]. Given the focused inclusion criteria, data from the Fx-005 trial cannot be used to determine which of the previously mentioned baseline characteristics are treatment effect modifiers, but Planté-Bordeneuve and colleagues have stated that the patient characteristics beyond age and sex could not be ruled out as effect modifiers [Citation1].
Second, the subgroups assessed by Planté-Bordeneuve and colleagues to account for patient heterogeneity did not fully address the differences in these baseline characteristics between the trials [Citation1]. To compare with the Fx-005 trial (n = 125), the base-case subgroup from the APOLLO trial considered only Stage 1 patients (n = 104). It must be acknowledged that ATTR-PN is a multi-faceted disease that requires adjustment for multiple patient characteristics simultaneously, particularly where there are potential effect modifiers. The few additional subgroup analyses conducted by Planté-Bordeneuve and colleagues only addressed two characteristics at a time. For example, although one subgroup accounted for disease stage and Val30Met mutation status (n = 44), treatment history was not considered. Similarly, another subgroup accounted for disease stage and patients who were ATTR-PN treatment-naïve (n = 48), but mutation status was ignored. The small sample sizes (<50 patients) in the performed subgroup analyses further reduce the credibility of the indirect comparisons.
The authors also incorrectly used a Bucher ITC, which focuses on summary-level data and is reserved for indirectly comparing treatments when trials have similar characteristics [Citation5,Citation6]. Rather, more methodologically rigorous analysis using matching-adjusted indirect comparison (MAIC) techniques could simultaneously consider all sources of heterogeneity to minimize the imbalances between APOLLO and Fx-005 trial populations. MAIC methods leverage individual patient data and utilize reweighting techniques to adjust for observed differences in characteristics among trials that may have different patient populations and limit potential bias [Citation7]. However, if the considerable heterogeneity between trial populations result in low effective sample size after adjusting for heterogeneity through a MAIC, the credibility of drawing an indirect comparison between treatments diminishes and the analysis will likely fail to meet the required standards for conducting a MAIC [Citation6].
In conclusion, the results presented by Planté-Bordeneuve and colleagues have major limitations. The authors did not adhere to best practices for conducting ITCs and failed to adequately adjust for major differences in patient populations across trials. Notably, it was inapproporate to use subgroup analyses to adjust for potential treatment effect modifiers, as the number of patients included in each subgroup was too low for robust ITC analyses. Collectively, these limitations may lead to biased ITC results, and the ITC results presented by Planté-Bordeneuve and colleagues should be interpreted with caution.
Declaration of interest
EM Salvo, C Cameron, D Tran and IA Samjoo are all employeed Cornerstone Research Group Inc with Dr Cameron also a shareholder in Cornerstone Research Group Inc. All authors are hired consultants to Pfizer Inc in connection with the development of this letter to the editor. The authors also declare that as part of Cornerstone Research Group Inc, they have worked with a number of pharmaceutical and medical device companies. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Acknowledgments
Michelle Stewart and Leslie Amass, employees of Pfizer Inc reviewed the final draft of this letter.
Additional information
Funding
References
- Plante-Bordeneuve V, Lin H, Gollob J, et al. An indirect treatment comparison of the efficacy of patisiran and tafamidis for the treatment of hereditary transthyretin-mediated amyloidosis with polyneuropathy. Expert Opin Pharmacother. 2018 Nov 29. PubMed PMID: 30489166; eng. DOI:10.1080/14656566.2018.1554648.
- Parman Y, Adams D, Obici L, et al. Sixty years of transthyretin familial amyloid polyneuropathy (TTR-FAP) in Europe: where are we now? A European network approach to defining the epidemiology and management patterns for TTR-FAP. Curr Opin Neurol. 2016 Feb;29 (Suppl 1):S3–S13. PubMed PMID: 26734951; PubMed Central PMCID: PMCPMC4739317. eng.
- Adams D, Gonzalez-Duarte A, O’Riordan WD, et al. Patisiran, an RNAi therapeutic, for hereditary transthyretin amyloidosis. N Engl J Med. 2018 Jul 5;379(1):11–21. PubMed PMID: 29972753; eng.
- Coelho T, Maia LF, Martins Da Silva A, et al. Tafamidis for transthyretin familial amyloid polyneuropathy: a randomized, controlled trial. Neurology. 2012 Aug 21;79(8):785–792. PubMed PMID: 22843282; PubMed Central PMCID: PMCPMC4098875. eng.
- Hoaglin DC, Hawkins N, Jansen JP, et al. Conducting indirect-treatment-comparison and network-meta-analysis studies: report of the ISPOR task force on indirect treatment comparisons good research practices: part 2. Value Health. 2011 Jun;14(4):429–437. PubMed PMID: 21669367; eng.
- Phillippo D, Ades AE, Dias S, et al. Methods for Population-Adjusted Indirect Comparisons in Submissions to NICE. 2016.
- Signorovitch JE, Sikirica V, Erder MH, et al. Matching-adjusted indirect comparisons: a new tool for timely comparative effectiveness research. Value Health. 2012 Sep 01;15(6):940–947.