Comparison of the visual outcome between macula-on and macula-off rhegmatogenous retinal detachment based on the duration of macular detachment

Introduction

Rhegmatogenous retinal detachment (RRD) occurs when structural changes in the vitreous cause retinal tears and passage of a liquefied vitreous through the break, resulting in separation of the neurosensory layer of the retina from the retinal pigment epithelium.1 Previous studies have reported that the involvement of the macula was a poor prognostic factor in RRD.2–7 Several preoperative factors that affect the visual outcomes in macula-off RRD have been identified, including the patients’ age, preoperative visual acuity (VA), the height of detachment and the duration of macular detachment (DMD).3 8

A 7-to-10-day window was considered standard practice for repairing macula-off RRD; however, patients included in these prior studies were primarily managed by scleral buckling (SB), and more recent evidence suggests that earlier interventions may be associated with better outcomes in macula-off RRD.9–14 Yorston et al reported that better visual outcomes were observed in macula-off RRD repaired with vitrectomy when the duration of foveal detachment was less than 3 days.13 A recent study by Malosse et al described that outer retinal layer damage was more prevalent in spectral domain optical coherence tomography in patients with macula-off RRD with longer DMD.15 In addition, previous studies on animal models revealed that the hypoxia induced by retinal detachment could cause molecular and cellular changes in the photoreceptors as early as 1–3 days.16 17

Reattachment of the macula does not guarantee significant improvements in vision. Some patients may experience metamorphopsia and dysmetropsia after the repair, especially when the fovea is involved.6 18 Other factors, including cystic outer retinal degeneration, retinal folds, cystoid macular oedema, epiretinal membrane and proliferative vitreoretinopathy (PVR), have been found to impact the degree of vision improvement.3 19–22

This study aims to evaluate the effects of DMD on the final VA in macula-off RRD compared with the visual outcomes in macula-on RRD. We hypothesised that the visual outcomes in macula-off RRD would be closer to those of macula-on RRD if the macular attachment was achieved earlier. We also evaluated if phakic status, preoperative VA and the presence of PVR grade C or worse could help predict the postoperative outcomes.

Patients/materials and methods

This retrospective study included 96 eyes in 96 patients diagnosed and treated for primary RRD at a tertiary referral centre from June 2012 to March 2020. The approval of the West Virginia University Institutional Review Board was obtained. It was not appropriate or possible to involve patients or the public in the design, or conduct, or reporting or dissemination plans of our research. An initial search was conducted using the West Virginia Clinical and Translational Science Institute’s Integrated Data Repository and ICD-10 diagnosis and procedure codes. We used strict exclusion criteria, and the list of exclusion criteria is summarised in table 1. Ninety-six eyes met the inclusion criteria, which were: RRD without other ophthalmic comorbidity, single operation success, follow-up of at least 1 month and known DMD.

RRD was categorised as either macula-on (macula-sparing) or macula-off (macula-involving). Thirty-four patients had macula-on RRD (macula-sparing), and 62 patients had macula-off RRD (macula-involving). Individuals with macula-on RRD were included in group A. We further divided patients with macula-off RRD into three groups based on their DMD: group B (DMD≤3 days; n=22), group C (DMD between 4–7 days; n=22) and group D (DMD≥8 days; n=18). In macula-off RRD cases, DMD was recorded as the number of days from the time of central vision loss to the time of intervention.

We recorded age, gender, phakic status, laterality, the presence of intraoperative PVR grade C or worse, types of procedures, preoperative best-corrected visual acuity (BCVA) and follow-up duration. Postoperative final BCVA was measured at the most recent visit. Snellen VA was converted into the logarithm of the minimum angle of resolution (logMAR) scale for statistical analysis. We assigned logMAR values of 2.0 and 2.3 to represent counting fingers and hand motion VA, respectively.23 There were no patients who demonstrated VA worse or equal to light perception in our study.

Our patients underwent pars plana vitrectomy (PPV), SB or pneumatic retinopexy (PnR) and the types of procedures performed in each group are summarised in table 2. PPV was the primary procedure in 44 eyes without other interventions, while SB was performed in two eyes as the primary procedure. Combined PPV and SB were done in 10 eyes. Twenty-five eyes received PnR as the primary procedure without additional surgeries. We included 15 patients who received PnR prior to PPV (11 eyes), SB (2 eyes) and combined PPV and SB (2 eyes). These cases were not classified as recurrent RRDs, because the retinal attachment was never successfully achieved with the initial PnR alone. In these cases, PPV and SB were performed immediately after the retinal attachment was found to be unsuccessful, and the mean duration between the PnR and PPV and SB was 3.6±3.0 days. The DMD was measured from the day of central vision loss to the day of PPV and SB.

The correlation between the preoperative VA and the final outcome of patients with RRD was assessed and compared between patients with preoperative VA, better or worse than Snellen VA 20/400. Wilcoxon rank-sum test was used in the analysis of continuous outcomes between two groups, and Fisher exact test was used in the analysis of categorical outcomes. Spearman’s correlation was estimated between continuous variables. Analysis of variance (ANOVA) analysis was used to assess the differences of VA among three groups based on their DMD. A multivariable linear regression model was used to assess the final postoperative BCVA, when adjusted for age, gender, laterality, phakic status, the presence of PVR grade C or worse, preoperative BCVA, the number of detached quadrants and the types of interventions, including PPV, SB and PnR. A stepwise variable selection was carried out to identify the best subgroup of independent variables based on Akaike Information Criterion (AIC) statistics. All statistical tests were two-sided, and p<0.05 was considered statistically significant. Statistical analyses were carried out using SAS V.9.1 (SAS Institute, Cary, North Carolina, USA) and R software (V.3.6.3, R Foundation, Vienna, Austria).

Results

Ninety-six eyes (49 left eyes) in 96 patients (63 men) were included in the study. The demographic data are summarised in table 2. The mean age was 59.0 years (SD=10.9 years). Forty-nine phakic patients and 47 pseudophakic patients were identified. The median DMD was 4 days (range: 1–21 days). The median follow-up duration was 15.8 months (range: 1 month to 7 years).

The demographic information (age, gender), phakic status, types of interventions and the preoperative and postoperative BCVA in each group are summarised in table 2. The mean number of detached quadrants in patients with macula-off RRD is recorded in table 2 as well. As judged by multiple linear models, no statistically significant differences in the mean final VA were found between group A (logMAR 0.04±0.07) and group B (logMAR 0.05±0.06) (p=0.78). However, there were significant differences in the final VA between group A and group C (logMAR 0.15±0.15) (p=0.017) and between group A and group D (logMAR 0.36±0.30) (p<0.001). When the final VA values of group B and C were compared, there were no statistically significant differences in mean final VA (p=0.33).

The mean final VA of patients with macula-off RRD was stratified by their DMD with a 1-day interval up to DMD of 10 days, and the results are summarised in table 3. The mean final VA was comparable among patients with the DMD of 1–3 days; however, there was a tendency of worsening final VA starting at DMD of 4 days.

A scatter-plot was used to illustrate the association between the DMD and the final VA in patients with macula-off RRD (figure 1). There was a positive correlation between them (Spearman correlation coefficient=0.60, p<0.0001), demonstrating that earlier interventions may be associated with overall better VA at the final visit.

Request permission

In patients with macula-off RRD, there was a statistically significant difference in the final VA between patients with preoperative VA ≥20/400 and those with <20/400 using univariate analysis (p=0.042) (table 4). There was no significant association between phakic status and the final VA (p=0.62) (table 5). However, there were statistically significant differences in the final VA when the patients were divided based on the presence of PVR grade C or worse (p=0.016) (table 6). In a multivariable linear regression model, DMD≥8 days significantly correlated with worse final VA (coefficient=0.132, SD=0.04, p=0.0005, AIC=2.12) in macula-off RRD, when adjusted for potential confounding variables including age, gender, laterality, phakic status, the presence of PVR grade C or worse, preoperative BCVA, the number of detached quadrants and the types of interventions, including PPV, SB and PnR. After a stepwise variable selection based on AIC statistic, the final model (AIC=1.86) included the phakic status (p=0.08), SB (p=0.012), PnR (p=0.08), preoperative BCVA (p=0.02) and the group of DMD (p<0.0001).

Discussion

Our study aimed to determine the effect of the DMD on the visual outcomes in macula-off RRD. We hypothesised that earlier repair of the macular detachment in macula-off RRD would result in similar visual outcomes to those of macula-on RRD.

Burton et al reported that patients with a macular detachment less than 5 days had better visual outcomes, and approximately 1 line of vision was lost for 1 additional week until 27 days.24 Ross and Kozy reported no statistically significant differences in anatomic or visual outcomes if the repair was performed within the first 7 days of macular detachment and suggested that nonemergent treatments may be more cost-effective.11 Studies from Hassan et al and Diederen et al demonstrated that good postoperative outcomes could be obtained in patients treated within 10 days of macular detachment.10 25 However, these previous studies only included patients who underwent SB and did not investigate the effect of repair within the first 3 days of macular detachment.

In contrast, Van Bussel published a meta-analysis that reported better visual outcomes in patients with DMD less than 3 days compared with those with DMD of 4–7 days.26 Furthermore, Greven et al described that patients with DMD less than 3 days had better outcomes than those with DMD of 4–7 days, and they included patients who underwent PPV with or without SB to ensure that the macula was attached after the drainage of SRF.12 Yorston et al performed a large database study involving 2074 eyes undergoing vitrectomy for macula-off RRD and reported that there was a higher probability of achieving postoperative VA of ≤0.30 logMAR when the repair was performed within 3 days of vision loss.13

Our results demonstrated no differences in preoperative VA among patients with macula-off RRD when grouped by the DMD (table 2). In fact, the patients with macula-off RRD whose DMD was less than 3 days had worse preoperative VA than those with longer DMD. Group A (macula-on) had the mean final VA of logMAR 0.04±0.07 (Snellen 20/22), which was comparable to the 12-month postoperative VA of logMAR 0.06±0.1 (Snellen 20/23) in patients with pseudophakic macula-on RRD as reported by Rezai et al.27 Their 12-month postoperative VA in macula-off patients was logMAR 0.2±0.3 (Snellen 20/32), which was similar to the final VA value of logMAR 0.175±0.22 (Snellen 20/30) in our macula-off RRD group.

The mean final VA of group A (macula-on) and B (macula-off, DMD≤3 days) was not significantly different. This indicates that the final visual outcomes in macula-off RD were similar to that of macula-on RRD when the macular attachment was achieved within 3 days of central vision loss. In contrast, there were significant differences in the final VA between group A and group C (macula-off; DMD 4–7 days) and between group A and group D (macula-off; DMD≥8 days). Although there was a small difference in the mean final VA between group B (macula-off, DMD≤3 days) and C (macula-off, DMD 4–7 days), it did not reach statistical significance.

Williamson et al reported that patients with fovea-on RRD achieved significantly better visual outcomes than fovea-off RRD regardless of the duration of vision loss.6 On the contrary, our results showed that the final VA in macula-off RRD might be similar to that of macula-on RRD if macular detachment repair was performed within 3 days of central vision loss. Furthermore, Williamson et al demonstrated that patients with a shorter duration of vision loss (less than 3 days) had significantly better visual outcomes than those with 4–6 days of vision loss. However, their definition of vision loss was different from that of our study. Our study attempted to specify the number of days of central vision loss based on the patients’ history, while they did not differentiate the central and peripheral vision loss, which may have resulted in variation in the estimation of the DMD.

There were no differences in the final VA among patients with macula-off RRD with DMD of 1 day, 2 days or 3 days, while patients with DMD of 4 days or greater had worse final VA (table 3). This suggests that the visual outcomes might not differ up to DMD of 3 days, but it might be negatively affected by DMD starting on day 4. This supports Henrich’s study, which demonstrated a statistically significant decrease in VA gain in patients whose DMD was ≥4 days.28 However, our results were inconsistent with a previous paper reporting that patients with shorter DMD were associated with the better final visual outcome even among patients with less than 3 days of DMD.12

Preoperative VA and the presence of PVR grade C or worse have been proposed as statistically significant predictors for visual outcome. Based on our analysis, a better preoperative VA and the absence of PVR grade C or worse were identified as potential positive prognostic factors consistent with the previous studies.3 6 11 12 29 However, the phakic status of our patients was not found to be a significant factor in predicting the visual outcomes in our study. This might be because 19 out of 26 patients (73%) who were initially labelled as phakic and had macula-off RRD later underwent cataract extraction before the final visit. Furthermore, we included patients with relatively short follow-up duration, which might be too early for post-PPV cataracts to develop. The mean duration of follow-up was 161±96 days in patients with macula-off RRD who were still phakic at the final visit.

Our study’s limitations include the small sample size, the retrospective nature of the study, and the fact that we had to rely on elicited patients’ history, which may not be accurate for estimating the exact DMD.30 A randomised prospective study might minimise any bias, which could not be done due to ethical reasons.9 We have excluded patients with trauma and those who had recurrent RRD; therefore, extrapolating the data to these patients may not be applicable. We also did not assess the presence of metamorphopsia or dysmetropsia, which could affect the quality of life of postoperative patients.

In conclusion, the mean final VA of patients with macula-on RRD was comparable to that of the macula-off patients with a DMD≤3 days. The clinical implication based on the data cautiously suggests that if macula-on RRD cannot be immediately repaired, a repair completed within 72 hours appears to carry comparable outcomes despite macular detachment within that time frame. However, we did not observe any statistically significant differences in the final visual outcomes for repairs done within 7 days after macular detachment. Larger studies are warranted to further investigate the effect of DMD on the long-term VA outcome.