Article Text

The growing importance of pharmacoeconomics: the case of age-related macular degeneration
1. A F Smith
1. Correspondence to Andrew F Smith, Medmetrics Inc, 30 Charles Street, Ottawa, Ontario K1M 1R2, Canada; Andrew.Smith{at}Medmetricsinc.com

## Statistics from Altmetric.com

A logical point of departure in any discussion of age-related macular degeneration (ARMD) is an examination of global ageing trends. Currently, the global population stands at around 6.9 billion, and this is expected to grow rapidly to 9.5 billion by 2050, roughly a 1.4-fold increase in a 40-year time span.1 2 Moreover, the percentage of persons aged >60 years, particularly among high-income economies, will experience unparalleled growth. In 2000, for example, it was estimated that there were approximately 606 million people ≥60 years; however, by 2050, this figure will rise to nearly 2.4 billion, a nearly fourfold potential increase.1 2 Such underlying population trends will invariably place tremendous pressures on healthcare systems to provide safe, effective and affordable interventions for those persons with ARMD. In fact, using pooled data from a number of well-conducted population-based eye surveys in high-income countries, it is possible to examine the potential impact of such underlying population projections for the number of persons with potentially treatable neovascular ARMD (NV ARMD) globally.3 This is presented in table 1. As can be seen, the number of persons with NV ARMD aged ≥60 years will rise sharply over the course of the next 40 years, ranging from 23.47 million in 2010 to 80.44 million by 2050. On the obverse side of the coin is the question of who will pay to treat this expected swell in ARMD patients, given that the number of persons engaged in full-time employment whose tax dollars fund healthcare expenditures is likely to decline at the same time as the number of retirees is expected to rise dramatically. As such, the primary raison d'être of the pharmacoeconomics approach lies in being able to guide those involved in healthcare spending decisions to better allocate scarce healthcare resources to meet the ever-growing demand for eye care services now and in the future.

Table 1

Global population projections and number of persons with NV ARMD in both sexes: 2010–2050

In the case of cost-effectiveness analyses, it is typical to consider the incremental cost-effective ratio (ICER) of the existing medical intervention relative to the new intervention of interest. Mathematically, the ICER may be expressed as follows: ICER=ΔCosts (Ca−Cb)/Δ Effects (Ea–Eb); where, ΔCosts represents the change in average costs and ΔEffects represents the change in the average effects between the two healthcare interventions a and b being considered. Typically, therefore, pharmacoeconomic studies attempt to evaluate the cost and health effects of a new drug or medical device relative to an existing drug or medical device and express such findings in cost per quality-adjusted life year (QALY) gained. In addition, various so-called cost-effectiveness thresholds have been advanced to determine at what point a particular healthcare intervention is deemed to be cost-effective. The most widely used benchmark for what is deemed to be a cost-effectiveness intervention was first advanced in 1999 in the UK by the then newly established National Institute for Clinical Excellence that pegged this value at a maximum acceptable ICER value of between £20 000 (€29 500; US$40 000) to £30 000 (€44 250; US$60 000) per QALY gained.4

In recent years, a number of approaches to treating patients with NV ARMD have emerged. Table 2 highlights a few of the main interventions over the course of the last 5 years, and though by no means exhaustive, it serves to illustrate the flurry of activity of new treatments for ARMD. Initially, there was considerable promise for photodynamic therapy with Visudyne and pegaptanib (Macugen) whose cost-effectiveness ratios were generally favourable. With the approval of ranibizumab (Lucentis; Genetech, South San Francisco, California, USA) that obtained Food and Drug Administration approval in June 2006 for the treatment of wet or NV ARMD and the widespread off-label use of bevacizumab (Avastin; Genentech) that began concurrently, considerable debate has arisen over the role of insufficient clinical trial data and the economic imperatives driving the adoption of Avastin relative to Lucentis for NV ARMD. Although there is persistent anecdotal evidence that many ophthalmologists using Avastin obtain the same clinical results as do those using Lucentis to treat patients with NV ARMD, no comparative head-to-head clinical trials data exist. It is against this backdrop that the National Eye Institute launched the Comparison of AMD Treatments Trials in 2008 and is expected to report in the near future on the effectiveness of Lucentis versus Avastin for NV ARMD. In the meantime, economic pressures are favouring the widespread off-label use of Avastin because in many jurisdictions, Avastin is 40 to 100 times cheaper than Lucentis. For comparative purposes, Raftery et al5 calculated that given that there are roughly 25 000 new cases of NV ARMD in the UK annually, the cost of treating these patients with ranibizumab (Lucentis) would amount to £300 million, whereas if bevacizumab (Avastin) was substituted for Lucentis, then the National Health Service in the UK would save £292 million annually. This immediately begs the question as to whether Lucentis and Avastin are in fact perfect substitutes or, in pharmacoeconomics terms, effectively me-too drugs. Essentially, a me-too drug is a drug that is structurally very similar to an existing drug. The term me-too, however, often carries with it negative implications as me-too drugs tend to foster competition and drive down the price of drugs over the long term. In the case of Avastin and Lucentis, such a scenario is complicated by the fact that both drugs are manufactured by the same manufacturer, namely, Genentech. An equally important parallel question is whether the existing cost-effectiveness analyses of Avastin and Lucentis have appropriately taken into account the full costs and full health effects associated with possible severe systemic adverse effects due to their long-term use in patients with NV ARMD.

Table 2

Cost-effectiveness of interventions for NV ARMD

In the case of Avastin, which was originally developed for the treatment of metastatic colorectal cancer, it is know that Avastin, owing to its higher half life (20 days; range 11–50 days), takes up to 100 times longer to clear from the circulatory system than Lucentis. Indeed, pharmacologic studies have highlighted that the direct ocular injection of bevacizumab does not prevent it from entering the systemic circulation. Indeed, after the intravitreal injection of the 1.25 mg human dose into a rabbit model, serum bevacizumab concentrations peaked at 8 days at a concentration of 3.3 μg/ml and only fell below 1 μg/ml 29 days after injection.13 Because ranibizumab was not detected in the serum after intravitreal injection in the same rabbit model, the rates of systemic adverse events observed with this drug may not be applicable to bevacizumab, although they both have the same mechanism of action.14 Moreover, given the continuing absence of any published phase III trials using bevacizumab for the treatment of ocular disease, large safety surveys and clinical studies form the foundation of the systemic safety data that exists to date. Even large prospective clinical trials such as the National Eye Institute's Comparison of AMD Treatments Trial may not possess adequate statistical power to detect low rates of systemic adverse events and differences between treatment arms.15 As intravitreal bevacizumab continues to be used for an increasing number of ocular conditions, the systemic risk of treatment is not well established. The absence of evidence for such systemic events does not necessarily indicate the evidence of absence of such systemic effects. Rather, it may be a case of planning larger studies powered to detect such potentially significant systemic adverse events.

Finally, it is important to reflect on the fact that traditional cost-effectiveness studies typically only measure the costs and health effects associated with a particular indication and do not generally include the overall costs and health effects of systemic adverse events. Given the unknown systemic safety signals surrounding Avastin, future pharmacoeconomic analyses of Avastin and Lucentis for ARMD and other ocular indications should ideally be designed to take into account the overall patient safety profile and not be limited strictly to ocular health considerations. An isolated ocular approach would be unwise, as this may fail to adequately capture the severe adverse effects, such as life-threatening hypertension and hypertensive events due to repeated intravitreal injections. Such systemic events are not routinely detected by ophthalmologists, whose remit is ocular as opposed to general patient health. In sum, both the ocular and overall health of the patient should be considered in an integrated fashion in future pharmacoeconomic studies.

## Footnotes

• Competing interests None.

• Provenance and peer review Commissioned; not externally peer reviewed.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.