Aflibercept in the real world – a clinical audit of people with diabetic macula oedema from 21 UK hospitals
by Irene Stratton
To describe the treatment protocols and outcomes of intra-vitreal injections of aflibercept to patients with diabetic macular oedema.
Anonymised data of patients undergoing anti-VEGF injections for diabetic macular oedema was exported from a dedicated ophthalmology electronic patient record systems. Those who had aflibercept injections only were included. For those who had both eyes treated, only the first eye to be treated was included and for those who had both eyes treatment started simultaneously, the better eye was included. Linear regression was used to relaterelated change in visual acuity (VA) over 12 months from the first injection with age (grouped <65 years, 65 to 74, 75 and above) , baseline letter score (<50 letters, 50 to 59, 60 to 69, 70 to 79, 80 or more) and number of injections.
Data was available for 1566 patients who could have had 12 months follow-up from first injection. Of these 142 had no baseline VA in the 8 weeks prior to the first injection, 181 had no 12-month anniversary VA within 8 weeks of the 12-month anniversary, 40 had no baseline or 12-month data. Thus Because of missing data 363 patients (23%) were excluded from the analysis. The 1153 patients with complete data were of age 64 (57 to 73) years (median (25th to 75th centile)), with the baseline VA of 64 (54 to 72) ETDRS letters and had received 6 (5 to 8) injections. Over 12 months the VA improved by 5 (0 to 12) letters.
Worse baseline VA was associated with greatest improvement (p<0.001), those with fewer than 50 letters at baseline having improvement of 15 (18.7) (mean (s.d.)) letters and those with 80 letters or more losing 2.2 (5.7) letters. Older patients had smaller improvement (p=0.0002)), patients aged 75 years and above gaining 4.6 (13.8) letters and those under 65 gaining 6.6 (14.3) letters.
The number of injections was of borderline significance (p=0.051) with each additional injection giving improvement of 0.4 letters.
Despite the high level of missing data the results concur with previous reports. The efficacy of antiVEGF demonstrating demonstrates a 5 letter gain similar to previous real life studies, the ceiling and floor effect on VA outcomes, with somewhat lower injection count (n=6).