iERM: Immediate Versus Deferred Surgery

March 2023

Joshua Uhr, MD

Location/Affiliation: Retinal and Ophthalmic Consultants, PC

Northfield, NJ

Al-Khersan H, Shaheen AR, Flynn HW Jr, Smiddy WE. Natural History and Surgical Timing for Idiopathic Epiretinal Membrane. Ophthalmol Retina. 2022 Nov;6(11):978-984. doi: 10.1016/j.oret.2022.02.014. Epub 2022 Mar 8. PMID: 35272083.

The authors studied the natural history of idiopathic epiretinal membrane (iERM) and compared outcomes in patients with iERM who underwent surgery within 6 months of initial presentation (immediate surgery group) to patients who later progressed and underwent surgery more than 6 months after presentation (deferred surgery group).

Prior research has shown that better preoperative visual acuity (VA) is associated with better post-operative VA. Therefore, some surgeons argue that surgery at earlier stages – while VA is still good – may help prevent future vision loss. However, this rationale for early surgery may lead to unnecessary surgery and subject patients to unnecessary complications. This study aimed to clarify the rate of progressive vision loss from iERM and to determine if long-term visual outcomes are worse if patients defer surgery until VA worsens.

A retrospective chart review of patients presenting with iERM was performed. Patients with a past ocular history that could be contributory to ERM formation were excluded. The decision to proceed with surgery was individualized, and took into consideration vision and the perceived severity of the patient’s symptoms.

The authors studied the natural history of iERM in patients for whom immediate surgery was not performed. To that end, they collected data regarding the incidence of undergoing surgery at a later date, various demographic variables, ocular history, and VA at multiple time points. The authors also performed a nested case-control study comparing eyes that underwent deferred surgery (6 months or longer after presentation; the cases) with eyes that underwent immediate surgery (within 6 months of presentation; the controls).

During the study period from 2014 to 2019, 693 eyes met inclusion criteria. Of these, 369 did not undergo surgery during a mean follow up of 37.16 months (unoperated group). Forty-four eyes were initially observed but underwent surgery greater than 6 months after diagnosis (deferred surgery group; cases in the case-control study). These 413 eyes (unoperated group plus deferred surgery group) made up the natural history study arm. The remaining 280 eyes underwent surgery within 6 months of diagnosis (immediate surgery group), of which 44 were selected using a frequency matching technique to serve as the controls in the nested case-control portion of the study. These controls were matched for preoperative vision, age, and sex.

In the 413 eyes included in the natural history study arm, 176 (42.6%) reported visual symptoms at initial presentation, including blurred vision, metamorphopsia, general vision loss, or loss of function. Forty-four (10.7%) underwent vitrectomy with membrane peeling at a mean of 18.1 months after presentation. At first presentation, 38% of the unoperated group was at least mildly symptomatic, compared to 84% of the deferred surgery group (P<0.0001). In the logistic regression model, only the presence of initial symptoms was associated with eventual deferred surgery (P<0.0001). Median vision decreased in unoperated pseudophakic eyes from 20/30 at baseline to 20/40 at final follow up. Median vision in pseudophakic eyes with deferred surgery was 20/40 at baseline, decreased to 20/66 immediately preceding their surgery, and improved to 20/40 at final follow up.

In the nested case-control study, median preoperative vision was 20/59 in the deferred surgery group and 20/30 in the immediate surgery group (p=0.395). One-year postoperative vision was 20/50 in the deferred surgery group and 20/33 in the immediate surgery group (p=0.0288).  However, lens status is a confounding variable in any vitrectomy series. A subanalysis of pseudophakic eyes was therefore performed. Psuedophakic eyes in the deferred surgery group (19 of the 44 cases) had a median vision of 20/40 at the initial presentation, 20/66 at the preoperative visit, and 20/40 at final follow up. Pseudophakic eyes in the immediate surgery group (22 of the 44 controls) had a median vision of 20/55 at the initial and preoperative visits and 20/34 at final follow up. Eyes in the deferred surgery group had worsening vision after presentation (which was a major contributor to undergoing eventual deferred surgery), but there was no difference in the magnitude of vision change from the preoperative visit to the final visit in pseudophakic eyes that underwent deferred surgery (median logMAR change: 0.22) versus immediate surgery (median logMAR change: 0.21). The final vision was similar for pseudophakic eyes in the deferred versus immediate surgery groups (20/40 vs 20/34, respectively; p=0.4550).

The authors discuss that in their study, deferred surgery was not associated with worse visual outcomes (when lens status was corrected for). Since the rate of eventual surgery in eyes that were initially observed was only around 10%, and since deferring surgery was not associated with poorer outcomes in these eyes, the authors conclude that performing surgery early with the goal of avoiding future vision loss involves unnecessary surgery in approximately 90% of such cases.

There are several limitations related to the retrospective design of the study. Data regarding presenting symptoms – which are important considerations in terms of recommending immediate surgery versus observation – were not able to be standardized. There was also subjectivity in patients’ symptoms, patients’ perceived visual needs, and in criteria for surgical intervention among the two surgeons in the study. The two surgeons were generally biased against immediate surgery for patients with relatively good vision and mild symptoms. Statistical analyses for pseudophakic eyes may also be limited by smaller sample sizes.

In conclusion, the authors note that their findings support the safety of deferring surgery for patients with iERM with relatively good vision and mild symptoms until symptoms worsen. Doing so does not appear to compromise final visual outcomes and minimizes risk to patients.

Obtain CME credit: