Response to the letter to the editor on “an indirect treatment comparison of the efficacy of patisiran and tafamidis for the treatment of hereditary transthyretin-mediated amyloidosis with polyneuropathy”

KEYWORDS:

• Transthyretin amyloid polyneuropathy
• indirect treatment comparison

We thank the authors of the recent letter to the editor ^[1] for their thoughtful review of our article, ‘An indirect treatment comparison of the efficacy of patisiran and tafamidis for the treatment of hereditary transthyretin-mediated amyloidosis with polyneuropathy’ [1,2].

Given the growing interest in hereditary transthyretin-mediated (hATTR) amyloidosis, it is expected that questions will be raised by health care professionals, payers, and regulators about the comparative efficacy of patisiran and tafamidis. As described in the original article, the objective of the analysis was to conduct an indirect, quantitative comparison of patisiran and tafamidis in patients with hATTR amyloidosis with polyneuropathy, since head-to-head trials comparing these two therapies are not available.

As reported in the original article and noted in the letter to the editor [1], there are differences in the baseline characteristics of patients enrolled in the APOLLO and Fx-005 studies (e.g., age, baseline Neuropathy Impairment Score in the Lower Limbs [NIS-LL]) [3,4]. However, differences in baseline patient characteristics do not detract from the methodological rigor of the analysis unless there is evidence that such characteristics are effect modifiers. For example, age would be considered an effect modifier for treatment if older patients received less treatment benefit relative to placebo than younger patients (or vice versa). It is important to note that age would only be an effect modifier if it impacted the relative difference between treatment and placebo. If older and younger patients had similar differences between treatment and placebo, even if older patients experienced poorer results than younger patients, there is no effect modification.

Our systematic review of the published literature on the tafamidis Fx-005 study and subgroup analyses from the patisiran APOLLO study yielded no evidence of effect modifiers that would introduce bias to this indirect comparison. The indirect comparison methodology used in our analysis adheres to preferred practices outlined by health technology assessment agencies and was developed to minimize bias in estimates of the treatment effect between groups, despite differences in baseline characteristics between studies [5–7].

We acknowledge, as suggested in the letter to the editor [1], that a matching-adjusted indirect comparison could be considered as an alternative methodology to conduct this analysis. However, we did not implement this methodology in our analysis in keeping with guidance from the National Institute for Health Excellence (NICE) on indirect treatment comparisons (ITCs) [6], which states that such adjustments for baseline characteristics must be supported by clear, quantitative evidence showing that this approach would produce less biased estimates than standard approaches. There is no available evidence to suggest that implementing such an approach would produce less biased estimates; in fact, to the contrary, patisiran demonstrated consistent treatment effects in the APOLLO study, regardless of patient baseline characteristics [3].

In addition, as described in the original article, we followed preferred practices in conducting sensitivity analyses to explore the effect of potential effect modifiers in different patient populations (disease stage, mutation status, treatment history) to ensure the robustness of our findings [8]. As noted, all analyses across populations showed consistent trends, with statistical significance observed in a number of analyses despite small subgroups – all of which further supports the reliability of our findings.

In the design of our analysis, we considered that potential effect modifiers could impact the interpretation of an indirect comparison. However, our systematic literature review and the results of our sensitivity analyses did not find any evidence of effect modification. Therefore, we believe that the methodology in the original article is valid and adheres to established standards for conducting indirect comparison. While noting the limitations intrinsic to ITCs, we believe that our analysis provides a reliable and important addition to the existing literature that may facilitate more informed decisions in the absence of data from a head-to-head study. Indeed, the analysis has already been used by regulatory bodies and payers to facilitate decision-making [9,10].

Declaration of interest

H Lin and S Agarwal are employees of Alnylam Pharmaceuticals. M Betts, K Fahrbach and M Chitnis are employees of Evidera, who received funding from Alnylam Pharmaceuticals to conduct this research. Finally, M Polydefkis declares personal fees from Alnylam Pharmaceuticals, grants and consultancy fees from Pfizer, expenses from Ionis Pharmaceuticals, and consultancy fees from Vertex Pharmaceuticals, Biogen, and Chromocell, all outside the submitted work. 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.

Additional information

Funding

This work was sponsored by Alnylam Pharmaceuticals, Cambridge, MA, USA.