Aims To analyse the incidence and baseline clinical characteristics of patients with neovascular age-related macular degeneration (nAMD) treated with intravitreal anti-vascular endothelial growth factor (VEGF) injections in a defined UK region.
Methods A standardised dataset was collected prospectively using an electronic medical record (EMR) system from 1 January 2008 to 21 June 2012 for all patients living in Gloucestershire who received intravitreal anti-VEGF injections for nAMD.
Results Over the study period, 1207 eyes from 1033 patients began intravitreal anti-VEGF injections for nAMD. The annual incidence in the years after National Institute for Health and Care Excellence (NICE) technology appraisal 155 implementation was stable at 120 (95% CI 110 to 138) eyes or 100 (89 to 115) people per 100 000 population. The most common indication was occult choroidal neovascularisation (51%). Median baseline visual acuity (VA) was significantly higher for second treated than first treated eyes (66 and 56 letters, respectively; p<0.0001). Median baseline VA of fellow eyes increased from 47 (2008) to 67 letters (2012; p<0.005). The proportion of patients with baseline VA in the better eye ≥70 letters increased from 27.6% (2008) to 51.4% (2012; p<0.0001), while the proportion eligible at baseline for full or partial certificate of visual impairment decreased from 13.8% (2008) to 7.1% (2012; p<0.05).
Conclusions The incidence of patients undergoing anti-VEGF therapy for nAMD increased substantially following NICE approval of ranibizumab (August 2008), and has been stable since 2009. This equates to an annual UK incidence of 26 850 (21 320 to 32 440) eyes, similar to NICE estimates. Patients eligible for blindness certification before treatment decreased by half from 2008–2012. Prospective data collection using an EMR system is invaluable for efficient monitoring of real-world clinical care.
- Treatment Surgery
- Public health
Statistics from Altmetric.com
Age-related macular degeneration (AMD) is the most common cause of blindness and visual impairment in developed countries including the USA and the UK.1 ,2 Advanced AMD has neovascular and atrophic forms, but neovascular AMD (nAMD) is responsible for over 90% of cases of severe visual loss.3–5 The prevalence of nAMD is predicted to continue increasing over the next decades, from 415 000 cases in 2010 to 516 000 in 2020 in the UK,6 and from 1 million in 2010 to over 2 million in 2050 in the USA.7
The management of patients with nAMD forms a large and increasing proportion of hospital eye service activity in the UK, particularly since the introduction of intravitreal anti-vascular endothelial growth factor (VEGF) therapy.8 The delivery of anti-VEGF therapy has created substantial service pressures on National Health Service (NHS) ophthalmology departments.9 ,10 Delays in access to timely treatment and the required frequency of follow-up assessment are common.11
There is uncertainty regarding the annual incidence of patients with nAMD requiring anti-VEGF therapy in the UK, with recent estimates ranging from 21 000 to 26 000 cases.9 ,12 ,13 The incidence figure of 26 000 patients per annum is widely quoted, and was used by National Institute for Health and Care Excellence (NICE) in technology appraisal 155 (TA155) of ranibizumab (Lucentis; Novartis (Basle, Switzerland) and Genentech (San Francisco, California, USA)) for the treatment of AMD, though no reference source was cited.12 The annual incidence of nAMD itself (ie, treated and untreated) in the UK has been estimated at 39 700 (95% credible interval 23 700 to 68 300) by Owen et al.14 However, the number actually requiring anti-VEGF therapy may be much lower, as many of these patients may not present to the hospital eye service, and others may be ineligible or refuse treatment. Accurate epidemiological data are vital for commissioning, planning service delivery and assessing the cost-effectiveness of treatment.
The primary aim of this study was to provide estimates of the annual incidence of intravitreal anti-VEGF therapy for nAMD by age, sex and baseline clinical characteristics in a well-defined region of the UK. The secondary aim was to define incidence figures for all treated cases and for those falling within NICE TA155 guidance on eligible levels of visual acuity (VA).
Data were collected prospectively on all patients (both NHS and private) who received intravitreal injections of ranibizumab or bevacizumab (Avastin; Roche (Basle, Switzerland) and Genentech) in the ophthalmology department that serves Gloucestershire (GL), a geographically well-defined county of the UK where very few, if any, patients seek care from neighbouring centres. Prospective collection of a standardised dataset was carried out as part of routine clinical care using a single electronic medical record (EMR) system (Medisoft Ophthalmology, Medisoft Limited, Leeds, UK). This system is used in the context of paperless AMD clinics where all staff enter all data directly into the EMR, guaranteeing high levels of data quality and completeness.15
Refracted VA is measured at the baseline visit only; at subsequent visits, VA is measured with the patient's habitual spectacle correction. Early Treatment Diabetic Retinopathy Study (ETDRS) logMAR VA scoring charts are included in the EMR; nurses record which letters are seen and the EMR system calculates the ETDRS letter score. At each follow-up assessment visit, optometrists or nurse practitioners working to protocols under the supervision of a consultant ophthalmologist perform fundal examinations and optical coherence tomography imaging. The central retinal thickness and presence/absence/change in re-treatment criteria are recorded at every visit.15
The EMR captured all anti-VEGF intravitreal therapy for this population before and after NICE TA155 was published in August 2008.12 For NHS care, the EMR recorded episode data at all visits, but only data taken as part of injection episodes were recorded for private patient care, as follow-up assessments occurred at sites that do not have access to the EMR.
NICE TA155 recommends treatment for nAMD with ranibizumab if the corrected VA at baseline is between Snellen 6/12 and 6/96 (70–25 ETDRS letters).12 This guidance was ignored in GL for patients with good vision (>70 letters) at baseline, who were treated when clinically indicated.
Data extraction from the EMR system included all eyes undergoing first intravitreal injection between 1 January 2008 and 21 June 2012. Patients were only included in this study if their primary residence had a GL postcode, but excluding the GL9 postcode as this area lies outside the Office for National Statistics (ONS) boundary for GL, and where nAMD was the indication for treatment. Data were collected on 1448 eyes. Of these, 147 eyes were excluded through having a non-GL postcode and one further eye through having a GL9 postcode, leaving 1300 eyes. Of these, 93 eyes were excluded as they had indications other than nAMD for anti-VEGF injection, leaving 1207 eyes from 1033 patients for inclusion. Of these, 1069 eyes (88.6%) had baseline VA data for the eye undergoing treatment.
The incidence rates of nAMD undergoing anti-VEGF therapy were calculated by eyes and by patients for the calendar years 2008–2012. Age and sex-specific incident treatment rates were calculated for the period 2009–2012 (ie, after full implementation of NICE TA155). Population denominators for the GL region were obtained from the ONS as the mid-year population estimates from the 2011 national census.16 Statistical analysis was performed by Student t test (ETDRS VA data) and χ2 test (proportions of patients by VA).
Over the 4.5-year study period, 1207 eyes from 1033 patients with GL postcodes were treated by intravitreal anti-VEGF injection for nAMD. These consisted of 408 eyes from 360 male patients, and 799 eyes from 673 female patients.
Trends over time are shown in figure 1. The number of new cases per annum treated increased from 224 eyes of 217 patients in 2008 to 276 eyes of 240 patients in 2009, reaching a plateau of 288 eyes of 238 patients in 2011. Similarly, the number of new cases treated meeting NICE criteria increased from 142 eyes of 135 patients in 2008 to a peak of 217 eyes of 191 patients in 2009, before reaching a plateau of 204 eyes of 173 patients in 2011.
The incident treatment rate of all new cases per year increased from 101 (95% CI 88 to 115) eyes per 100 000 population (aged ≥50years) in 2008 to a plateau around 120 (110 to 138) in 2009–2011. This equates to an increase from 98 (85 to 111) people per 100 000 population (aged ≥50 years) in 2008 to a peak of 107 (93 to 121) people in 2009, followed by a small decline to 102 (89 to 115) people in 2011. Incident treatment rates for only those cases meeting NICE VA criteria generally mirrored the pattern for all cases.
Age and sex-specific incident treatment data (for 2009–2012, when there was full access to treatment) are shown in figure 2. The annual incidence of treated nAMD increased approximately exponentially with increasing age: in men from 0 people per 100 000 population per year in the 50–54-year age group to 533 (427 to 639) people in the 85+ year age group; the equivalent increase in women was from 3 (0 to 6) to 511 (439 to 583). Overall, the incidence was 65% higher (ratio of 3 : 2) for women than for men. The difference in incident treatment rates between sexes peaked at 90% in the 60–64-year age group, and reversed in the group aged ≥85 years such that incidence was 4% higher in men.
In terms of numbers of new patients treated, mean annual numbers increased from 0 male and 1 female patient in the 50–54-year age group to a peak of 22 (male) and 42 (female) patients in the 80–84-year age group. The majority (53.5%) of new patients were aged 80–89 years. In age groups from 80 to 99 years, the raw number of new females treated was approximately twice that of males, predominantly owing to the higher number of women than men in the population of this age group.
In approximately half the number of eyes (51.3%), the indication for treatment was occult choroidal neovascularisation. The proportions of eyes treated for other subtypes of nAMD are shown in table 1.
Over the study period, the median VA of the affected eye at baseline was 58 letters (mean 54.9, range 0–89), increasing significantly from 53 (50.9, 0–83) in 2008 to 58 (55.2, 0–85) in 2012 (p=0.01) (table 2). The median VA in the fellow eye at baseline over the study period was 65 letters (mean 54.2, range 0–94) but this increased significantly from 47 (44.7, 0–90) in 2008 to 67 (mean 56.5, range 0–90) in 2012 (p=0.0002). The median VA of the better-seeing eye at baseline increased significantly from 58 letters (mean 58.3, range 4–90) in 2008 to 70 (66.1, 17–90) in 2012 (p=0.00002). Additionally, the proportion of patients with vision suitable for driving (VA in the better eye ≥70 ETDRS letters, equivalent to Snellen 6/12) (figure 3) at baseline increased from 27.6% in 2008 to 51.4 in 2012 (p<0.0001). Similarly, the proportion of patients eligible at baseline for a full (0–24 letters) or partial (25–39 letters) certificate of visual impairment (CVI) decreased from 13.8% in 2008 to 7.1% in 2012 (p=0.04).
This study describes the epidemiology of all patients undergoing intravitreal anti-VEGF treatment for nAMD in a defined region of the UK with a population of 598 300 (including 230 700 people aged ≥50 years). Our data demonstrate that the delivery of anti-VEGF therapy has been approximately stable since the Technology Assessment by NICE in August 2008,12 which recommended provision of ranibizumab treatment for NHS patients with nAMD. Prior to that, from December 2007 to August 2008, NHS funding for ranibizumab treatment in GL was limited to second affected eyes. Prior to December 2007, anti-VEGF therapy was not funded by the NHS in this area and was only available privately.
The annual incidence rate since NICE TA155 approval (2009–2012) of treated nAMD in GL has stabilised at 102 (89 to 115) people per 100 000 population, or 78 (66 to 90) per 100 000 if analysis is restricted to eyes meeting NICE TA155 VA criteria.
By applying our age and sex-specific incident treatment rates for GL to the ONS UK population data for 2011,16 and assuming that the population of GL is representative of the UK (see below), this would equate to an annual treatment incidence for the UK of 26 850 (21 320 to 32 440) eyes from 22 180 (17 120 to 27 280) people (all treated cases), or 19 660 (14 980 to 24 400) eyes from 16 960 (12 590 to 21 380) people (only cases meeting NICE TA155 VA criteria). These estimates are broadly similar to the estimate suggested by NICE of 26 000 new cases requiring treatment per annum for the UK, although the source of their estimate is not referenced in their forecasts.12 Using ONS population projections,16 the equivalent figures for 2020 would be 32 380 (25 760 to 39 060) eyes from 26 710 (20 670 to 32 810) people (all treated cases), or 23 740 (18 130 to 29 410) eyes from 20 490 (15 260 to 25 790) people (only cases meeting current NICE VA criteria).
Our estimates of treated nAMD are considerably lower than the estimates of all nAMD suggested by Owen et al.14 These were generated by applying age-specific prevalence data from previous observational population studies to the UK population, though 28 of the 31 population studies included were from non-UK populations. They calculated that the total number of new cases of nAMD per year would be 39 700 (95% credible interval 23 700 to 68 300). However, our incidence rate for nAMD undergoing intravitreal treatment is expected to be lower than the incidence of nAMD itself, as some patients with nAMD may not present to hospital eye services or proceed to treatment, and others may present too late or be ineligible for treatment. For these reasons, our estimate is a more meaningful and relevant measure for planning and commissioning AMD services in similar ethnic and aged populations, both in the UK and globally.
It is of interest that the median VA of the fellow eye (table 2) increased from 47 letters in 2008 to 63 letters in 2009, and has been approximately stable since then. This suggests that many eyes whose anti-VEGF therapy began in 2008 were second affected eyes, that is, the other eye had previously lost vision from untreated nAMD (or other diseases). By contrast, many eyes whose anti-VEGF therapy began after 2009 appear to be either first eyes diagnosed with nAMD, or these patients may have had previous successful treatment for the fellow eye. As blindness is a binocular concept, this study suggests that the incidence of severe visual loss in both eyes may have decreased in this region in recent years. For example, the proportion of patients in our study eligible at baseline for full or partial CVI (figure 3) decreased from 13.8% in 2008 to 7.1% in 2012 (p=0.04). While we have not analysed age-specific VA data on all new patients with nAMD (ie, treated or untreated), this 49% reduction chimes with a recent analysis of CVI returns elsewhere, which reported a decrease of 21% (from 2005 to 2010) in incident blindness in Leeds, largely as a result of a 39% decrease in certification from nAMD.17 Similar decreases in blindness at population levels have been reported from Denmark, Israel and Scotland.18 ,19 ,20 Furthermore, nAMD clinical trials have demonstrated improvements in quality of life (QoL) following anti-VEGF therapy.21 While anecdotal evidence from UK patients suggests that QoL has improved substantially following anti-VEGF therapy, such outcomes are not routinely captured in clinical practice.
The main strength of this paper is the completeness of the dataset, as data were collected on all patients receiving intravitreal injections in GL (NHS or private), rather than a sample of patients. This completeness of data capture is made possible by routine use of an EMR,15 ,22 which records data conforming to a standardised dataset as a byproduct of routine clinical care. Further strengths include the reporting of incidence data and clinical characteristics over several years from a ‘real life’ medical retina service of patients actually being treated,23 rather than from population studies of disease (where patients may or may not have progressed to treatment) or from clinical trials (which have restricted inclusion criteria). Hence, these data will be highly applicable to other UK ophthalmology departments and NHS commissioners, and of wider interest in similar health economies. Additionally, we have provided analyses of nAMD incidence by eyes and by people, as well as by cases treated for all levels of vision and for only cases meeting the current vision criteria dictated by NICE.
A limitation in applying data from this region to calculate UK incidence figures is the extent to which the ethnic and socioeconomic profiles in GL are similar to those of the UK. For example, the proportions of individuals in GL of Asian or Afro–Caribbean ethnicity are approximately 0.5% and 0.2%, respectively, compared with 4.0% and 2.0% for Great Britain16; however, age-specific ethnicity data are not readily available from the ONS, and any differences in the elderly population may be smaller. Similarly, the index of multiple deprivation (IMD 2010) in GL was 14, compared with 19 for England.24 The IMD is the standard UK measure of socioeconomic deprivation, calculated by combining deprivation indices in seven domains (eg, income, employment and health). The index ranges in England from 1 (least deprived area) to 88 (most deprived), such that the level for GL is similar to that for England. The extrapolation is also limited by the absence of study data on AMD risk factors such as cigarette smoking and phakic status. Finally, although GL is geographically well defined, and the vast majority of patients are managed by local ophthalmologists, it is possible that a small number of patients may have travelled to other departments for treatment. Any such patients would not be included in our analysis, hence, our incidence figures might be a slight underestimate of the true situation.
In conclusion, this study provides epidemiological data on the incidence of nAMD treated with anti-VEGF therapy in a defined region of the UK. These suggest that the estimates provided by NICE in TA155 are of merit for service planning. Our report adds to the mounting evidence that treatment of nAMD with anti-VEGF therapy is reducing rates of blindness at population levels. Prospective data collection using an EMR system is recommended for ongoing anti-VEGF therapy.
Contributors TDLK, SPK and RLJ had the idea for the study. AS, QM and RLJ were involved in data collection. TDLK performed data analysis (assisted by RLJ and AT) and drafted the paper. All authors revised the paper.
Funding TDLK is funded by Fight For Sight through a Clinical Fellowship. The views expressed are those of the authors and not necessarily the funding body. The researchers are independent of the funders, and the researchers had access to all of the data.
Competing interests RLJ is the Medical Director of Medisoft Limited, which developed the electronic medical record from which data were extracted.
Ethics approval The local NHS Office for Research Ethics Committee determined that ethical approval was not required for this study.
Provenance and peer review Not commissioned; externally peer reviewed.
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.