Estimating the Yearly Number of Eyes with Treatable Neovascular Age-Related Macular Degeneration Using a Direct Standardization Method and a Markov Model

Abstract

purpose. To estimate the number of treatable eyes with neovascular subfoveal age-related macular degeneration (ARMD) in France.

methods. A literature search for studies documenting neovascular ARMD incidence rates and direct standardization according to age and gender were performed. Projection to the year 2025 was based on OECD (Organization for Economic and Co-operation Development) data. A cohort of patients aged 75 years was simulated by a seven-state Markov model. The mean treatment duration was fixed arbitrarily at 2 years. The probability of ARMD in the second eye was fixed at 30% at 5 years. Monthly mortality incidence was modeled from INSEE (Institut National de la Statistique et des Etudes Economiques) mortality tables. The time horizon of the model was 25 years. Sensitivity analyses were performed.

results. Based on the Rotterdam Study, 30,192 citizens per year will develop ARMD in one eye. Among them, 17,585 will be neovascular and 13,805 neovascular subfoveal ARMD. Taking into account the second eye, mortality, and a 2-year treatment duration, the number of neovascular subfoveal treatable eyes yearly would be 37,019 by 2025. Treatment duration was the most sensitive parameter. The number of eyes would be 18,899, 53,204, 67,535, and 80,162, for treatment lasting 1, 3, 4, and 5 years, respectively. A 2% yearly increase is expected up to 2025, due to population aging and the 1950s baby boom.

conclusions. According to the study model, the yearly number of subfoveal neovascular ARMD treatable eyes in France will be 37,019 by 2025. Average treatment duration was the most sensitive parameter.

Age-related macular degeneration (ARMD) is the most common cause of blindness in adults in Western developed countries. ^{1 2} ARMD significantly impairs vision- and health-related quality of life and functional independence. ^{3 4 5 6 7} The burden of ocular morbidity and visual disability due to ARMD will increase further with an expanding older population if there is no reduction of its incidence or improvement in treatment. Already, a steady increase in the number of people registering as blind in most Western countries suggests that the incidence of ARMD is growing. ⁸ Therefore, ARMD is becoming an increasing public health issue for decision makers when allocating resources. 

Two drugs are registered for the treatment of neovascular ARMD, verteporfin (Visudyne; Novartis, Basel, Switzerland) and pegaptanib (Macugen; OSI Pharmaceuticals, Melville, NY). Market authorization for these drugs was based on the results of placebo-controlled clinical trials. ^{9 10 11 12 13 14 15 16} Other treatments, with other mechanisms of action, are currently in development. Phase II randomized clinical trial results suggest that anecortave acetate is better than placebo with a similar efficacy to verteporfin, ^{17 18} with respect to preserving visual acuity, at least during the first year of treatment. Last, ranibizumab has shown efficacy superior to that of placebo. ¹⁹  

Cost-effectiveness analysis is the standard procedure for obtaining reimbursement from third-party payers in Western developed countries. ²⁰ Population size is a major parameter when estimating the additional budget necessary to finance an innovation. Several health technology assessment (HTA) reports have been published since verteporfin was marketed. For a comprehensive review, we attempted to collect all verteporfin HTA reports on treatment of neovascular AMD published in 15 European Economic Community (EEC) countries, knowing that most reports are not made public. We located one accessible report in each of the following countries: Germany, ²¹ Sweden, ²² France, ²³ and the United Kingdom. ²⁴ However, the methods used to delineate the target population differed. Some estimates were based on incidence rates and others on prevalence rates. Some countries favored local surveys, whereas others used foreign data. Also, definitions of the targeted populations were not identical. Some reports restricted the use of verteporfin to patients with visual acuity from 1/10 to 5/10, based on the visual acuity range for the patient eligibility in the clinical trials. Last, the ratio of classic to occult CNV in neovascular ARMD was based on different information sources. It was therefore not surprising to find significant target population differences between countries. In the England and Wales (52.8 million inhabitants) 5,000 new patients per year would require photodynamic therapy to treat predominantly classic neovascular ARMD; in Germany (82.5 million inhabitants) the number of new patients per year was between 9,000 and 60,000, in France (59.5 million inhabitants) it was 20,000, and in Sweden (8.9 million inhabitants) 700 to 1,000. 

Several factors should be taken into account to estimate the number of patients with neovascular ARMD who would be treated each year: number of new cases per annum, treatment duration, the probability of the fellow eye’s being affected by ARMD, the increasing death rate with advancing age. Last, population aging due to the 1950s baby-boom and longer life expectancy should be included. The present study was designed to estimate the total number of eyes affected by classic subfoveal neovascular ARMD that will be treated annually in France. 

Materials and Methods

The study was performed in three steps: (1) identification of epidemiologic studies on ARMD; (2) first treated eye incidence rate direct standardization to include effect of age (the major risk factor of ARMD); and (3) construction of a Markov model to estimate the total number of neovascular ARMD treatments in France per annum, taking into account the mortality rate, the occurrence of treatable disease in the second eye, and treatment duration. 

Literature Search

A comprehensive literature review was conducted by searching the PubMed database. ²⁵ Publications were searched from 1965 to June 2004. Search terms included age-related maculopathy (ARM), age-related macular degeneration (ARMD), neovascular ARMD, dry ARMD, atrophic ARMD, nonexudative ARMD, late ARMD, exudative ARMD, wet ARMD, soft drusen, large drusen, small drusen, pigmentary abnormalities, hyperpigmentation, hypopigmentation, increased retinal pigment, retinal pigment epithelial hypopigmentation, extrafoveal, juxtafoveal, subfoveal, incidence, prevalence, names of the relevant countries, predominantly classic lesions, minimally classic lesions, occult lesions, second-eye involvement, and progression from wet- to dry-form ARMD, along with epidemiologic terms and disease etiologies. 

All articles retrieved by this process were screened for relevance to the present study. Case reports and animal studies were excluded. Any article that provided the incidence and/or prevalence of ARMD was further reviewed in detail. Priority was given to well-conducted, population-based epidemiologic studies that provided valid, reliable and generalizable estimates of ARMD. The completeness of this search was checked against citations in published articles and by reviewing the published ARMD epidemiology literature in books. 

A thorough Internet search was also conducted for further information on ARMD epidemiology. These were conducted via large search engines (e.g., Yahoo [http://www.yahoo.com], Google [http://www.google.com]) targeting both public and subscription-based medical Web sites (e.g., Medscape [http://www.medscape.com], MD Consult [http://www.mdconsult.com]). 

Articles that provided estimates on any of the searched terms were included for review and analysis for this study. Priority was given to well-conducted population-based epidemiologic studies which provided valid, reliable, and generalizable estimates. Studies were focused on France, but other countries were admitted. Articles had to report incidence or prevalence rates of neovascular ARMD by age groups, to permit direct standardization. Papers reporting incidence rates were favored, since current neovascular ARMD treatments should be initiated soon after the onset of disease. Reports of incidence or prevalence rates from visual impairment registries were ignored, because about half of all visually impaired patients do not register. ^{4 5 26 27} Last, ARMD had to be medically confirmed, preferably according to the International Age-Related Maculopathy Epidemiologic Study Group rules. ²⁸  

Direct Standardization

Demographic data by age and gender, projected to 2025, was derived from United Nations data ²⁹ Because gender is not a confounding factor for ARMD’s incidence or prevalence, direct standardization was conducted on age only. This allowed adjustment for age, when data concerned foreign countries, and took into account population aging when projections were made to 2025. 

The Markov Model

A Markov model was developed (Tree Age Pro 2004; TreeAge Software Inc., Williamstown, MA), counting treatment months from diagnosis of ARMD in the first eye to the patient’s death. This method made it possible to estimate the annual number of treatable eyes, while taking into account treatment duration, death, and ARMD development in the second eye. A cohort of 1000 hypothetical patients entered the model with a new diagnosis of neovascular subfoveal ARMD in the first eye. Age was set at 75 years (the mean age of diagnosis of neovascular ARMD). ^{9 10 11 12 13 14 15 17} The sex ratio was modeled as a 4th-order polynomial function of age, estimated from national demographic data. ³⁰ Hence the sex ratio of this cohort was country specific. The cohort was processed by the Markov model and followed up throughout 25 years, with a cycle duration of 1 month and death as the absorption state. Patients entered the model with treated neovascular ARMD in one eye and could change state at the end of each month. Three new states were possible: (1) treated disease in the second eye, (2) treatment stopped in one eye, and (3) death. Figure 1presents a schematic design of the model. 

Many studies evaluated the risk of development of an exudative lesion in the second eye. Some studies performed in the 1970s and 1980s ^{31 32 33 34 35 36} reported frequency of development of neovascular ARMD in the second eye ranging from 3% at 1 year ³⁵ to 48% at 4 years. ³³ Annual incidence, estimated using a time-independent probabilistic model, varied from 3% ³⁵ to 15%. ³³ Other studies published later ^{37 38} reported calculated (same method as above) annual incidence rates varying from 8% to 12%. We used a 7% annual incidence rate for our central scenario. Sensitivity analyses were performed to account for the uncertainty associated with this estimate. 

A reference scenario fixed the average treatment duration at 2 years Monthly probabilities were estimated using an exponential model. It was hypothesized that the duration of treatment for the second eye would be the same as the first. The monthly mortality incidence function is already published. ³⁹ Mortality tables per age and gender were collected from INSEE (Institut National de la Statistique et des Etudes Economiques). Yearly mortality was modeled as a function of intercept (−5.233), age-square (in year, 4.958 10⁻⁴), gender (1: male, 2: female; −1.385), and the age × square by gender interaction (1.271 10⁻⁴). Hence the mortality incidence rate of the cohort was country specific. 

The Markov reward was the number of months under treatment. Reward was fixed to 1 month when one eye (either the first or the second) was treated and to 2 months when both eyes were treated during the same month. The number of treatable eyes/1000 patients was defined as the asymptotic value of the cumulative function of rewards over time. 

Results

As mentioned, a comprehensive literature review was conducted by searching the PubMed database. These searches yielded 3671 articles. The abstracts of all these articles were reviewed for relevance, resulting in 304 relevant articles. Of these, we reviewed all articles published in English. For articles published in other languages, we reviewed their findings as reported in their abstracts. After thoroughly reviewing these 304 articles, 48 studies were considered relevant for inclusion in the analysis of this study. 

Two studies were identified in France. A national study ⁴⁰ was performed in 1995 by the IPSEN Institute, to assess low vision. The objectives of this study were to estimate the incidence of low vision and related eye diseases presenting at ophthalmologists’ consultations. This study was not retained because neovascular ARMD was not clearly specified. 

Delcourt et al. ⁴¹ conducted the first large prospective population based study (POLA) in France, evaluating the association of cardiovascular disease and its risk factors with ARMD, and included both cataract and ARMD with their respective risk factors. These investigators predicted age- and gender-specific prevalence estimates of ARMD and its subtypes, in a sample of 2584 community residents aged 60 to 95 years. Definitions and diagnostic criteria were based on the International Classification of Diseases. However, the cross-sectional design of the survey did not permit an estimation of incidence rates. 

Three studies were identified outside France that satisfied the selection criteria: the Rotterdam Study, ⁴² The Blue Mountains Eye Study, ⁴³ and the Beaver Dam Eye Study. ⁴⁴ The Rotterdam Study was selected because, regarding the ARMD risk factors, French citizens are supposed to be more comparable (e.g., sun exposure, eating habits) to the Dutch than to Americans or Australians from the two other studies. 

The Rotterdam Study reported neither the location (extrafoveal, juxtafoveal, or subfoveal) nor the type of neovascular lesion ARMD (classic versus occult). This information was available neither for French nor European patients. The results reported by Moissieiev ⁴⁵ were not used, since data generated on the basis of the Macular Photocoagulation Study guideline did not fit our needs (e.g., the minimally/classic distinction and the type of lesions per localization were not addressed in detail). Olsen et al. ⁴⁶ and Margerhio et al. ⁴⁷ reported this information in two cross-sectional surveys. There was good agreement on localization—that is, 82.7% of neovascular ARMD lesions were subfoveal according to Margherio et al., and 78.5% according to Olsen et al. However, discrepancies arose concerning the nature of the condition. According to Margherio et al. 54.0% of neovascular lesions were subfoveal and classic, whereas the proportion was 19.5% according to Olsen et al. Results will be presented according to both findings. Finally, it was hypothesized that these estimates were the same for the first and second eye. 

Table 1presents the yearly incidence per age group reported in the Rotterdam Study. Depending on the type of lesion, dry or neovascular, incidence rates increased exponentially with age. 

Table 2shows the population projection for France estimated by the United Nations. In the next 20 years, the population aged >55 years will increase by 29.2%. 

Table 3indicates the number of single-eye ARMD cases according to direct standardization, showing that 30,192 new cases should have appeared in 2005 (17,585 neovascular and 12,607 dry). The number of neovascular subfoveal ARMD eyes will lie between 13,805 and 14,543, and the number of classic subfoveal neovascular ARMD eyes between 3,429 and 9,498. In the next 20 years the number of neovascular ARMD will increase by 33.2%, irrespective of location and type. 

Figure 2illustrates Markov’s stated probabilities for a patient having the first neovascular eye (FE) treatment at age 75. The probability of dying before 100 years of age was close to 1 (97.5%). After an average treatment duration fixed at 2 years, the probability of treatment continuing beyond 5 years was 4.2%, and the probability of treatment for neovascular ARMD in the second eye (SE) by the fourth year peaked at 8.8%. Under the hypothesis that a neovascular eye (whether the first or second eye) treatment is stopped when its visual acuity (VA) is >1.0 logMAR (average patient VA included in TAP was close to 0.7 logMAR and a failure was defined by a loss of 3 lines ⁴⁸ ), patients belonging to the “FE not treated+SE not treated” state had poor visual acuity in both eyes and were legally blind. Their curve reaches a plateau 10 to 12 years after the onset of disease and 25.8% of them were legally blind. Among 1000 ARMD patients starting a treatment for a subfoveal CNV in the first eye at 75 and followed-up for 5 years, 192 would die, 15 would be legally blind due to ARMD, 61 would still be treated for the first eye, 386 would have the disease in the fellow eye, and 83 would be treated for the second eye. 

Table 4shows the cumulative number of treatable eyes over 5 years, among 1000 new patients per year with one eye already treated, according to age at diagnosis, treatment duration, and different incidence rates of neovascular ARMD in the second eye during the same period, as determined by our Markov model. With respect to our reference scenario (based on Fig. 2 : average treatment duration 2 years, age at diagnosis 75 years, second-eye incidence rate 30% in 5 years) the projections in Table 3should be multiplied by 2.682. 

The Rotterdam Study provided first-eye incidence rates (Table 1) . Table 4gives the number of additional eyes that would be treated if the treatment duration is fixed at one year. For example, a population of 1000 persons diagnosed with neovascular ARMD at age 75 and subject to a 5-year second-eye incidence rate of 30%, followed up over their full life, would generate 369 second eyes needing treatment. 

Table 5presents results for France in 2005, projected to the end of the year. After 2 years of prior treatment and a second-eye 5-year incidence of 30%, the number of neovascular subfoveal eyes lies between 37,019 and 39,000, and the number of neovascular classic subfoveal eyes between 9,196 and 25,469. 

Discussion

Direct standardization based on the Rotterdam Study, ⁴² associated with a Markov model, was used to estimate the number of neovascular subfoveal ARMD treatable eyes in France, accounting for mortality, treatment duration, age at diagnosis, and the probability of ARMD in the second eye. We used the United Nations’ ²⁹ demographic projection for 2025 to identify future needs. According to our reference scenario (average treatment duration of 2 years, age at diagnosis 75 years, and second-eye incidence rate 30% in 5 years) neovascular subfoveal ARMD eyes totaling 37,019 to 39,000 should have needed treatment in 2005, of which 9,196 to 25,469 should have been classic. The cases will increase by >1% per annum until 2025. 

A comparison with the verteporfin HTA reports ^{21 22 23 24} is not straightforward because we did not use the same methods, and our results are expressed as number of treatable eyes, whereas the HTA reports describe number of patients. We believe that the use of incidence rates, instead of dividing prevalence by average disease duration to derive a proxy for the incidence rate, is a more appropriate method, especially with neovascular ARMD where age plays a predominant role. Also, the second eye contributes significantly to the number of treated eyes. If the bilateral nature of the disease is overlooked, the resources needed to treat patients adequately are underestimated by about one third. 

The Olsen et al. ⁴⁶ paper was published after the publication of the HTA reports. If this article has confirmed the frequency of subfoveal localization, reported by Margherio et al., ⁴⁷ a dramatic discrepancy exists in the rate of classic neovascular lesions. Olsen et al. ⁴⁶ themselves discussed a variety of selection biases to explain the acknowledged difference. More data are needed to resolve this uncertainty about the number of treatable subfoveal classic eyes. 

Our approach has several limitations. We used Dutch data as a proxy for French data so that we could use incidence estimates rather than prevalence estimates. The French POLA survey, ⁴¹ projecting a life expectancy of 10 years after the diagnosis of ARMD, would have given similar results. In our attempt to evaluate treatment needs in 2025 we hypothesized a constant incidence of neovascular ARMD per age group, which somewhat contradicts an RNIB (Royal National Institute of the Blind) report. ⁸ Some strong hypotheses were built into our Markov model: (1) equal treatment duration for both eyes; (2) treatment probability independent of age; (3) treatment failure probability, an exponential function; (4) independent disease evolution in the two eyes; (5) independence of visual impairment and death; and (6) no bilateral disease at entry into the model. Some refinements could be added to the model should it be combined with a full stochastic approach, but the sensitivity analyses in Table 4has taken into account the three major variables contributing to outcome: age at diagnosis, treatment duration, and 5-year second-eye incidence rate. 

Apart from sex ratio and mortality incidence rates, no other data in our model were country specific. Life expectancy and the age structure are very similar across 15 of the EEC countries ⁴⁹ (excluding eastern European Union countries). Therefore, the estimations in Table 4may be used for countries other than France, with a fair approximation. We also tried to provide sufficient estimates to permit a reasonable linear extrapolation. However, we advise use of an exponential extrapolation, which would give less biased results as our estimates came from a multiplicative model. 

New drug classes for neovascular ARMD will become available after verteporfin. It may be that drug combinations and cyclical treatment will be used, as with other chronic diseases, such as cancer. These developments may preserve visual acuity for a longer time, but lengthy treatment would call for additional resources. Table 4can be used to estimate the incremental resources required. For example, a new drug necessitating administration for more than 3 years instead of 2 years, would increase the number of treatable eyes by 43.7% (line 3 of Table 4 : [3854 – 2682]/2682), according to our reference scenario. 

In conclusion, the allocation of resources to neovascular ARMD should cover the bilateral extension of the disease. New drugs will have an impact on treatment duration, and this should be anticipated by public health decision makers. To neglect now the long-term trend of increasing incidence will incur serious consequences for patient care delivery 20 years hence. 

Finally, much uncertainty remains as to the incidence rate of classic neovascular subfoveal ARMD. If public health decision makers want to consider this subgroup of patients as a potential target population, additional epidemiologic work must be done. 

 Supported by an unrestricted grant from Alcon France SA. The literature research and calculation of age-gender incidence estimates by type of ARMD was contracted out to the Mattson Jack Group, St. Louis, MO.

 Submitted for publication November 16, 2005; revised March 16, 2006; accepted August 18, 2006.

 Disclosure: J.F. Korobelnik, None; N. Moore, Alcon Laboratories (F); P. Blin, Alcon Laboratories (F); C. Dharmani, Alcon Laboratories (F); G. Berdeaux, Alcon Laboratories (E)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

The Markov model structure. Patients enter the model with neovascular ARMD in the first eye. At the end of each month, the disease can appear in the second eye (SE treated), a treatment (FE treated, SE treated) can be stopped (FE not treated, SE not treated), or the patient can die (death). All second eyes developing the disease are treated. Death probability is described in the text. Treatment duration models used exponential functions. Five-year to 1-month incidence rates were extrapolated using an independent probabilistic calculation.

Table 1.

 

Incidence (per thousand) of Dry and Neovascular ARMD According to the Rotterdam Study Results after 6.5 Years of Follow-Up

Age (y)	Follow-Up in Person-Years	Dry			Neovascular			Dry + Neovascular
				Number	Incidence	Number	Incidence	Number	Incidence
55–59	2,240	0	0.000	0	0.000	0	0.000
60–64	6,218	0	0.000	1	0.161	1	0.161
65–69	6,602	3	0.454	2	0.303	5	0.757
70–74	5,460	3	0.549	7	1.282	10	1.832
75–79	3,578	5	1.397	9	2.515	14	3.913
80+	2,494	8	3.208	9	3.609	17	6.816
Total	26,592	19	0.715	28	1.053	47	1.767

 The Rotterdam Eye Study participants were classified according to the most advanced lesion in either eye, which explains why the numbers in the Dry + Neovascular columns correspond exactly to the sum of the Dry and Neovascular columns (i.e., no one was classified as having neovascular AMD in one eye and dry AMD in the other).

 Reprinted, with permission, from van Leeuwen R, Klaver CCW, Vingerling JR, Hofman A, de Jong PTVM. The risk and natural course of age-related maculopathy: follow-up at 6½ years in the Rotterdam study. Arch Ophthalmol. 2003;121:519–526. Copyright © 2003, American Medical Association. All rights reserved.

Table 2.

 

French Population Demographics in 2005 with Projection to 2025 in Spans of 5 Years

Age (y)	Current year 2005	2006-2025 Demographic Projection
				2006	2007	2008	2009	2010	2015	2020	2025
55–59	3,968,000	3,971,600	3,975,200	3,978,800	3,982,400	3,986,000	3,934,000	4,022,000	4,003,000
60–64	2,659,000	2,888,200	3,117,400	3,346,600	3,575,800	3,805,000	3,830,000	3,788,000	3,879,000
65–69	2,498,000	2,497,200	2,496,400	2,495,600	2,494,800	2,494,000	3,579,000	3,612,000	3,580,000
70–74	2,404,000	2,374,400	2,344,800	2,315,200	2,285,600	2,256,000	2,259,000	3,251,000	3,293,000
75–79	2,014,000	2,016,000	2,018,000	2,020,000	2,022,000	2,024,000	1,909,000	1,918,000	2,775,000
80–84	1,060,000	1,072,000	1,084,000	1,096,000	1,108,000	1,120,000	1,117,000	1,130,000	1,161,788
85+	1,227,000	1,270,000	1,313,000	1,356,000	1,399,000	1,442,000	1,645,000	1,707,000	1,753,212
Total	15,830,000	16,089,400	16,348,800	16,608,200	16,867,600	17,127,000	18,273,000	19,428,000	20,445,000

 Source: United Nations. ²⁹

Table 3.

 

Yearly Incident Cases Using a Direct Standardization Approach, Applying French Demographics to Incidence Rates in van Leeuven et al. ⁴²

Yearly Incident Cases	2005	2006	2007	2008	2009	2010	2015	2020	2025
Neovascular + dry (van Leeuwen et al. 42 )	30,192	30,557	30,921	31,286	31,651	32,016	33,760	36,142	40,094
Dry (van Leeuwen et al. 42 )	12,607	12,769	12,931	13,094	13,257	13,419	14,395	15,208	16,664
Neovascular (van Leeuwen et al. 42 )	17,585	17,788	17,990	18,192	18,394	18,597	19,365	20,934	23,430
Neovascular subfoveal (Olsen et al. 46 )	13,805	13,963	14,122	14,281	14,439	14,598	15,202	16,433	18,392
Neovascular subfoveal classic (Olsen et al. 46 )	3,429	3,469	3,508	3,547	3,587	3,626	3,776	4,082	4,569
Neovascular subfoveal (Margherio et al. 47 )	14,543	14,710	14,878	15,045	15,212	15,379	16,015	17,312	19,376
Neovascular subfoveal classic (Margherio et al. 47 )	9,498	9,607	9,716	9,825	9,934	10,044	10,459	11,306	12,654

 Subfoveal and classic subfoveal population sizes were estimated according to Olsen et al. ⁴⁶ and Margherio et al. ⁴⁷ Rates presented in the text should be applied from the neovascular population (from neovascular to neovascular subfoveal, from neovascular to neovascular subfoveal classic).

Figure 2.

 

Treatment of subfoveal CNV in AMD: Markov’s stated probabilities. Probabilities presented are not conditional to death. Age at diagnosis: 75 years. INSEE mortality tables. At 5 years, disease will develop in the second eye of 30% of the patients. Two Markov states (SE treated+FE treated; SE not treated+FE treated) are not shown because their probabilities were always inferior to 5%. Average treatment duration: 2 years.

Table 4.

 

Number of Treatable Eyes among 1000 New Neovascular ARMD Cases Per Annum in France

Fellow Eye 5-Year Incidence Rate (%)	65 Years									75 Years									85 Years
		1	2	3	4	5	1	2	3	4	5	1	2	3	4	5
10	1193	2401	3536	4596	5577	1082	2141	3091	3937	4685	960	1828	2539	3120	3596
20	1395	2792	4101	5319	6444	1243	2441	3513	4463	5302	1067	2018	2794	3425	3941
30	1533	3063	4497	5832	7065	1369	2682	3854	4892	5807	1163	2191	3027	3706	4260
40	1627	3249	4774	6195	7510	1468	2872	4128	5240	6220	1250	2348	3241	3965	4556
50	1690	3377	4968	6453	7830	1545	3023	4347	5521	6556	1327	2491	3438	4204	4830

 Data are shown according to age at diagnosis, treatment duration (1–5 years), and second-eye 5-year incidence rate. The estimation was performed by a Markov model (see Figs. 1 2 ).

Table 5.

 

Number of Treatable Eyes in France in 2005

Treatment Duration (y)	Second-Eye 5-Year Incidence Rate (%)	Type of Lesion
				Neovascular Subfoveal 46	Neovascular Subfoveal Classic 46	Neovascular Subfoveal 47	Neovascular Subfoveal Classic 47
1	10	14,938	3,711	15,738	10,278
	20	17,153	4,261	18,071	11,801
	30	18,899	4,695	19,911	13,003
	40	20,265	5,034	21,350	13,943
	50	21,327	5,298	22,468	14,673
2	10	29,550	7,341	31,131	20,331
	20	33,703	8,372	35,506	23,188
	30	37,019	9,196	39,000	25,469
	40	39,650	9,849	41,772	27,280
	50	41,730	10,366	43,963	28,710
3	10	42,670	10,600	44,953	29,357
	20	48,500	12,048	51,095	33,368
	30	53,204	13,216	56,051	36,605
	40	56,981	14,155	60,030	39,203
	50	60,006	14,906	63,217	41,284
4	10	54,343	13,499	57,250	37,388
	20	61,616	15,306	64,913	42,392
	30	67,535	16,776	71,148	46,464
	40	72,331	17,968	76,202	49,764
	50	76,213	18,932	80,291	52,435
5	10	64,672	16,065	68,132	44,495
	20	73,187	18,180	77,103	50,353
	30	80,162	19,913	84,452	55,152
	40	85,858	21,328	90,452	59,071
	50	90,506	22,482	95,349	62,269

 Data are shown according to treatment duration, second-eye 5-year incidence rate, and type of neovascular ARMD lesions. Age at diagnosis: 75 years. The results are extrapolated from Tables 3 and 4 . According to Olsen et al. ⁴⁶ 13,805 first neovascular subfoveal ARMD eyes should be treated in 2005 (Table 3) . If age at diagnosis is 75 years old on average, if the fellow eye 5-year incidence rate is 10%, and if treatment duration is 1 year, the multiplying factor is 1,082/1,000 (Table 4) . 13,805 × 1,082/1,000 makes 14,938 as reported in Table 5 .