In an era of spiraling health care costs, ophthalmologists face increasing pressure to justify the costs of the treatments they prescribe. That pressure is particularly intense when it comes to anti-VEGF drugs as a treatment for diabetic retinopathy and related conditions.

In research so far, anti-VEGF drugs have mostly met standard tests for cost effectiveness. Aside from bevacizumab (Avastin; Genentech), their cost effectiveness often doesn’t fare well compared to that of other treatments for diabetic diseases of the retina. And the advantage for bevacizumab comes entirely from its lower cost; in most trials, it appears somewhat less effective. And because it is compounded for ophthalmic use, rather than manufactured, it is typically not available to clinicians in the preparations used in these trials. While ophthalmologists must consider many other factors in their choice of prescriptions, a solid grounding in the economics of anti-VEGF drugs can enhance their discussions of these treatments with patients, caregivers, and policy makers.

RISING COSTS

Despite recent cost control efforts by both government and private insurers, national health expenditure accounts for 18% of the US gross domestic product and is expected to reach 20% by the year 2025.¹ More than 1 in 10 dollars spent on health care goes to treatment of diabetes. Diabetes is the leading cause of blindness among people ages 20-24,² and care for retinopathy adds significantly to that cost.³ Expenditures for Medicare part B for aflibercept and ranibizumab alone reached $2.5 billion in 2013.⁴

MEASURING COST AND EFFECTIVENESS

Is their cost worth the help these drugs provide? Answering that question requires breaking down measurements of cost and effectiveness into their component parts.

The obvious way to measure cost effectiveness might be the number of dollars spent by a patient or insurer to prevent vision loss. In 1989, Javitt et al estimated that the cost of preserving a patient’s vision using laser panretinal photocoagulation was about $966 per year. This was “only one-seventh of the $6,900 average cost of 1 year of Social Security Disability for those disabled by vision loss.”⁵

But such a calculation leaves out important considerations. On the cost side of the equation, there are “opportunity costs,” such as the time patients spend traveling to and at appointments. And there are the costs of managing treatment complications. Likewise, effectiveness is more than a question of lines of vision. It’s also a matter of how well a patient can function and enjoy life, and for how long.

For a more comprehensive measurement of how valuable a treatment is to a patient, researchers might start by asking patients to imagine having 10 years left to live. They might give patients a hypothetical choice between living with impaired vision for all of those 10 years, or giving up some of their life span in exchange for better vision.

Based on responses and estimates of patients’ life spans, researchers can calculate quality adjusted life years (QALYs). To use a hypothetical example, they might find that an average individual expecting to live 10 years would give up 4 months of life in exchange for an improvement in VA of 1 line. This would render a QALY of about 0.033 per line of vision (4 months is about 0.33 years or 0.033 of 10 years).⁶

To know whether the expense of a treatment is worth the number of QALYs it saves means putting a dollar value on a human life. As repugnant as this may sound in principle, it occurs every day, for example, when individuals decide whether to pay for optional safety features in their cars, when governments decide whether to subsidize vaccinations, or when employers decide whether to provide workplace gyms.

Not surprisingly, it’s hard to reach an agreement on such a value. For years, US economists used $50,000 per QALY as a rough threshold above which a treatment is too expensive. More recently some authors have argued for a cutoff of $100,000 or $150,000 per QALY.⁷

COST EFFECTIVENESS IN MACULAR EDEMA

The 3 anti-VEGF drugs currently being used for diabetic retinopathy and macular edema differ widely in cost. As of 2015, aflibercept (Eylea; Regeneron) cost $1,850 per 2.0 mg dose, ranibizumab (Lucentis; Genentech) cost $1,170 per 0.3mg dose, and bevacizumab cost about $60 when repackaged at compounding pharmacies into syringes for ophthalmic use containing 1.25 mg.⁸

To figure out how these costs relate to the drugs’ effectiveness, Smiddy used data from multiple published clinical trials to compare treatments for DME. He based the analysis on conservative estimates of cost and high assessments of effectiveness for the more expensive treatments, and the opposite for the less expensive ones.⁹

Even with the scale tilted in this way, he found that bevacizumab was the most cost effective treatment and the ranibizumab was the least. Equivalent data on aflibercept were not available at the time of this study. He calculated a cost per QALY of $23,119 for ranibizumab using data from a DRCR.net 2010 trial, and $19,251 using data from the READ trial. These were all within Smiddy’s “acceptable range” of $1,466 to $25,566 per QALY.⁹ Bevacizumab, by contrast, had a cost per QALY of $2,013 using PACORES data and $4,160 using BOLT data, making it less expensive than the $5,862 that Smiddy calculated for grid laser, $6,346 for intravitreal triamcinolone, $9,446 for dexamethasone implant, $16,667 for pegatanib, and $8,706 for vitrectomy.⁹

Smiddy reached similar conclusions about treatments for BRVO, with ranibizumab coming in at $25,566 per QALY, bevacizumab at $824 and everything else in the middle between these extremes. For CRVO, ranibizumab cost $15,867 per QALY, bevacizumab cost $2,613, and again the other treatments fell between the two.

Looking at the problem from another direction, Mitchell et al compared the cost effectiveness of ranibizumab with or without laser to laser alone for DME in the UK. They based their analysis on data from the RESTORE trial showing that patients receiving ranibizumab for 1 year improved by about 5 letters more than patients receiving laser alone and scored higher on a visual function questionnaire.10 Using an average patient age of 63 years, they translated these superior outcomes to a difference of 0.17 QALYs. They calculated that 1 year of treatment with ranibizumab alone cost $7,050 (£4,191) more than a year of laser photocoagulation alone. Dividing the cost by 0.17 gave them a difference in cost of $40,419 (£24,028) per QALY using ranibizumab instead of laser. This was below the “willingness-to-pay threshold” of $50,465 (£30,000) per QALY “that has generally been considered cost effective in the UK.”10 Combination therapy with laser and ranibizumab compared to laser alone fell just outside this threshold at $60,658 (£36,106) per QALY.10 With younger patients, treatment with ranibizumab was more cost effective, the researchers found. The researchers also reported that adverse events did not occur often enough to affect cost effectiveness.¹⁰

On the other hand, Ross et al included adverse events important in a cost-effectiveness model comparing aflibercept, ranibizumab, and bevacizumab as treatments for DME.⁸ The study was based on a DRCR.net trial that showed aflibercept to be more effective than the other 2 drugs, especially in patients starting with the worst VA. But Ross et al found that bevacizumab was so much cheaper that using the other 2 drugs could not be justified on the basis of cost effectiveness.⁸

COST EFFECTIVENESS IN PROLIFERATIVE DIABETIC RETINOPATHY

Anti-VEGF therapy may be more cost effective for proliferative diabetic retinopathy than for DME because it may prevent more vision loss with fewer injections. That was one finding of Lin et al who used data from the DRCR.net Protocol S clinical trial to compare the cost effectiveness of panretinal photocoagulation, ranibizumab, and bevacizumab for proliferative DR.¹¹ Since Protocol S showed that outcomes were similar for the 2 treatments, the relative cost was particularly relevant. Lin et al used Medicare reimbursement data to determine costs and performed separate calculations for services performed in a hospital and for services performed in an office-based clinic. In Protocol S, some patients received photocoagulation and others ranibizumab as the primary treatment, and in each case a portion of the patients subsequently received the other treatment. They based their calculations on 2 years of treatment.¹¹

The authors found that photocoagulation as the primary treatment cost $7,988 in the hospital and $6,297 in the office per QALY. When ranibizumab was the primary therapy, the cost was $19,150 in the hospital and $16,238 in the office per QALY. The difference in cost between the 2 treatments was less than might be expected, Lin et al speculated, in part because anti-VEGF treatments can treat or perhaps even prevent DME at the same time it treats proliferative retinopathy.¹¹ Further, the ranibizumab dose in the study was 0.5 mg, which is more than the 0.3 mg specified by the FDA. Using the lower dose would save $1,860 per QALY in the photocoagulation arm of the study and $5,602 per QALY in the ranibizumab arm.¹¹ Lin et al also estimated considerable savings if bevacizumab were used, assuming it achieved the same outcomes. Using bevacizumab yielded $5,864 in the hospital and $2,952 in the office per QALY.¹¹ Extending photocoagulation treatment beyond the initial 2 years would incur costs of retreatment, office visits and imaging, bringing the cost per QALY up to $24,005 in the hospital or $14,219 in the office.¹¹ In this analysis, only 1 scenario exceeded the threshold of $50,000 to $100,000 per QALY that these authors used for acceptable cost. This scenario assumed that ranibizumab required 2 additional injections per year after the first 2 years. In that case, the cost would reach $164,360 per QALY in the hospital or $138,852 in the office over the patient’s lifetime. Substituting bevacizumab into this scenario yielded a cost of $47,914 in the hospital or $22,407 in the office per QALY.¹¹

They found that treatment with ranibizumab might exceed a $100,000 threshold of acceptable cost per QALY after about 20 years, leading them to speculate that photocoagulation might be more cost effective, especially for younger patients.¹¹

This analysis didn’t include all the adverse events possible with photocoagulation, such as peripheral visual field loss, induced glaucoma, and a possible increased need for vitrectomy. Lin et al considered the statistical effects of these problems to be minor,¹¹ but a wealth of clinical experience suggests these such side effects can place a significant burden on patients. Aflibercept achieved regulatory approval more recently than ranibizumab. Head-to-head comparisons so far have shown mixed results and have mostly focused on macular edema. Adedokun et al found ranibizumab to be more cost effective, calculating a lifetime QALY of $19,091 (£15,273) for ranibizumab and $21,684 (£17,347) for aflibercept as a treatment for macular edema secondary to BRVO in the UK.¹²

In Sweden, Eriksson et al found aflibercept to be more cost effective when compared to ranibizumab for macular edema secondary to CRVO. They cited a greater number of treatments needed with ranibizumab.¹³

Such results might seem to point to bevacizumab as a clear winner in the cost effectiveness contest. But Genentech doesn’t sell it in formulations for ophthalmic use, so clinicians must rely on compounding pharmacies to reformulate it. The products these pharmacies provide is variable and may not be as effective as the formulation used in clinical trials. Using the drug in the United States may become even more problematic. Some compounding pharmacists have warned that new FDA regulations aimed at preventing contamination could make it all but impossible to reformulate for ophthalmic use.¹⁴ It’s worth noting as well that studies of cost effectiveness can’t take into consideration all variables. Results may vary with the health of the population, for example, if one population has a higher average hemoglobin A1c than another. Some studies have measured VA benefits in the worse-seeing eye and others in the better-seeing eye. And few studies attempt to measure office visit costs or indirect costs such as caregiver burden.

Also a meta-analysis by Poku et al showed that the difference between moderate visual impairment (20/50 to 20/100) and legal blindness (worse than 20/200) was more important to patients than the difference between moderate visual impairment and 20/20 vision. But many cost effectiveness studies have treated the value of a line of vision as the same, regardless of baseline VA.¹⁵

Clinicians should bear in mind the limitations of these analyses. Effectiveness measurements are based on clinical trials where patients with significant comorbidities or who have difficulty adhering to treatment are often excluded. And as noted above, in the case of bevacizumab the formulation available to the clinician is often not the one used in trials. Moreover, cost effectiveness is only one consideration in choosing a treatment. More important are the needs of particular patients, which can vary depending on their economic circumstances, their support networks, the treatments they have previously tried, their lifestyles, and their overall health.

In an ideal world, physicians could ignore costs and simply prescribe the most effective treatment for each patient. But as pressure grows to bring down the cost of health care, fewer and fewer enjoy that luxury. RP

REFERENCES

1. US Centers for Medicare and Medicaid Services. NHE fact sheet. Available at: https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/nhe-fact-sheet.html . Accessed May 19, 2017.
2. American Diabetes Association. Economic costs of diabetes in the U.S. in 2012. Diabetes Care. 2013;36(4):1033-1046.
3. Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion. Vision Health Initiative national data. Available at: https://www.cdc.gov/visionhealth/data/national.htm
4. US Government Accountability Office. Medicare Part B: Expenditures for new drugs concentrated among a few drugs, and most were costly for beneficiaries. Washington, DC: US Government Accountability Office; 2015:QAO-16-12. https://www.gao.gov/products/GAO-16-12
5. Javitt JC, Canner JK, Sommer A. Cost effectiveness of current approaches to the control of retinopathy in type I diabetics. Ophthalmology. 1989;96(2):255-264.
6. Whitehead SJ, Ali S. Health outcomes in economic evaluation: the QALY and utilities. Br Med Bull. 2010;96:5-21.
7. Neumann PJ, Cohen JT, Weinstein MC. Updating cost-effectiveness--the curious resilience of the $50,000-per-QALY threshold. N Engl J Med. 2014 Aug 28;371(9):796-797.
8. Ross EL, Hutton DW, Stein JD, et al; Diabetic Retinopathy Clinical Research Network. Cost-effectiveness of aflibercept, bevacizumab, and ranibizumab for diabetic macular edema treatment: analysis from the Diabetic Retinopathy Clinical Research Network Comparative Effectiveness Trial. JAMA Ophthalmol. 2016;134(8):888-896.
9. Smiddy WE. Economic considerations of macular edema therapies. Ophthalmology. 2011;118(9):1827-33.
10. Mitchell P, Annemans L, Gallagher M, et al. Cost-effectiveness of ranibizumab in treatment of diabetic macular oedema (DME) causing visual impairment: evidence from the RESTORE trial. Br J Ophthalmol. 2012;96(5):688-693.
11. Lin J, Chang JS, Smiddy WE. Cost evaluation of panretinal photocoagulation versus intravitreal ranibizumab for proliferative diabetic retinopathy. Ophthalmology. 2016;123(9):1912-1918.
12. Adedokun L, Burke C. Cost-effectiveness of ranibizumab versus aflibercept for macular edema secondary to branch retinal vein occlusion: a UK healthcare perspective. Adv Ther. 2016;33(1):116-128.
13. Eriksson M, Castelo-Branco A, Nilsson J. Cost-effectiveness of aflibercept in the treatment of macular oedema secondary to central retinal vein occlusion in Sweden. Value Health. 2014;17(7):A608.
14. Harrison L. Pharmacists warn of ophthalmic drug shortages. April 29, 2015. Available at: http://www.medscape.com/viewarticle/843991#vp_1
15. Poku E, Brazier J, Carlton J, Ferreira A. Health state utilities in patients with diabetic retinopathy, diabetic macular oedema and age-related macular degeneration: a systematic review. BMC Ophthalmol. 2013;13:74.