Impact on visual acuity and psychological outcomes of ranibizumab and subsequent treatment for diabetic macular oedema in Japan (MERCURY)

Purpose The MERCURY study aimed to evaluate the effects on visual acuity and psychological symptoms, and safety, of ranibizumab and subsequent treatment in patients with diabetic macular oedema (DME) and impaired visual acuity (VA). We report data from the prespecified 12-month interim analysis. Methods This was a 24-month, phase 4, open-label, single-arm, prospective, observational study conducted at 20 specialised retinal centres in Japan. Participants were 209 patients with DME and impaired VA, not previously treated with either intravitreal or systemic anti-vascular endothelial growth factor (anti-VEGF) agents, who initiated ranibizumab 0.5 mg per investigator discretion. Following ranibizumab administration, patients were treated per routine clinical practice. Other treatments were allowed. The main outcome measure was the mean change in best-corrected VA (BCVA) in logarithmic minimum angle of resolution (logMAR) from baseline to month 12. An exploratory objective was to assess patients’ psychological status using the Hospital Anxiety and Depression Scale (HADS). Results The mean ± standard deviation BCVA at baseline was 0.43 ± 0.39 logMAR. The mean number of injections of ranibizumab and anti-VEGF agents from baseline to month 11 was 3.2 ± 2.0 and 3.6 ± 2.4, respectively. The BCVA change from baseline to 12 months was − 0.08 ± 0.34 logMAR (p = 0.011), showing a significant improvement; the HADS-anxiety score also decreased significantly (p = 0.001) and the depression score decreased numerically (p = 0.080). Conclusion MERCURY study data confirm the effectiveness of real-world treatment initiated with ranibizumab in Japanese patients with DME. In addition, treatment was able to positively influence anxiety via VA improvement. Supplementary Information The online version contains supplementary material available at 10.1007/s00417-021-05308-8.


Study design
The MERCURY study is being conducted at 20 specialised retinal centres nationwide, and has enrolled patients with diabetic macular oedema (DME) who initiated ranibizumab 0.5 mg in daily clinical practice. The study is ongoing; it commenced in April 2017 and is expected to be completed in December 2020.
Patients were enrolled between August 2017 and June 2018.
No visit schedule was imposed, and patients were treated per routine medical practice. Data were recorded at each visit in electronic case report format (eCRF) and were regularly checked by field monitors. It was recommended that data should be captured in the eCRF at every visit or a minimum of every 3 months, and investigators were encouraged to follow-up with patients who had not been seen in the centre for at least 6 months in order to capture data. Patients not seen at least once per year were discontinued from the study.

Study outcomes and measures
For assessment of the primary study objective, it was recommended that bestcorrected visual acuity (BCVA) be measured by Landolt C chart in a sitting position at an initial testing distance of 5 m (16.4 feet) at every visit; however, the usual practice of the participating investigator was permitted. For statistical analysis, the decimal BCVA value collected was converted into logarithmic minimum angle of resolution (logMAR) units using the formula: logMAR BCVA = log10 (1/decimal BCVA) = −log10 (decimal BCVA).
Secondary objectives also included assessment of treatment exposure (ranibizumab, anti-vascular endothelial growth factor (VEGF) treatments and DME Study data collected at baseline included patient demographics, non-ocular and ocular characteristics, and medical history/comorbidities. During the study period, information on anti-VEGF treatments and other DME treatments (number, date and reason for treatment) were collected.

Statistical methods
For sample size calculation, prior studies suggest that a mean BCVA change of 0.1 logMAR units from baseline to month 12 (equivalent to five Early Treatment Diabetic Retinopathy Study letters in BCVA) and 0.1 units of standard deviation can be assumed. A sample size of 180 primary treated eyes (PTEs) would yield a 95% confidence interval (CI) with a half-width of approximately 0.015 units. The overall sample size of 200 PTEs was set with the assumption that 10% of participants would not contribute any data at month 12.
The safety set included all enrolled patients who received ≥ 1 dose of ranibizumab during this study and had ≥ 1 safety assessment after the first treatment. The PTE set included all primary treated eyes; the secondary treated eye (STE) set included all secondary treated eyes, and the fellow eye set included all fellow eyes. For effectiveness analyses, only the PTE and STE sets were evaluated.
All data were analysed descriptively and summarised together with estimates and corresponding 95% CIs as appropriate. The paired t-test was performed to evaluate mean changes from baseline to month 3 and month 12, and the chi-square test was used to compare categorical variables at baseline and month 3 or month 12. Pearson's correlation coefficient was used to assess the correlation between continuous variables (i.e. BCVA and HADS scores).
All analyses were conducted based on observed data and no imputation method was applied. All statistical calculations were performed using SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA).