Abstract
Purpose:
Retinal detachment surgery is usually successful, but 10 to 20% of initial operations will fail, and it is difficult to predict which eyes will require re-operation, or to establish risk-adjusted normal success rates for primary surgery. We performed a prospective observational clinical study to determine the primary success rate of retinal detachment surgery in the United Kingdom, and to identify risk factors associated with failure.
Methods:
A national retinal detachment dataset was used to design a database. In order to maximise participation, the database is online, and can be accessed from any computer connected to the internet. When data is entered at the time of surgery, an operation note printout is generated. Outcomes are entered either at discharge or at the time of primary failure. Users are presented with a graph showing their own outcomes compared to all outcomes in the database. The online data is downloaded to a MS Access database for analysis.
Results:
1,698 operations were entered by UK consultant retinal surgeons, of which 1,474 were performed before 1 April 2014. At presentation, 52% were macula off, and 28% were pseudophakic. In 75% of eyes the detachment was due to a horseshoe tear. 1,315 (89%) had at least two months’ follow-up. 1,109 (87.5%) were re-attached with one operation. 159 (12.5%) required further surgery. Logistic regression identified three risk factors for failure: PVR grade C (O.R. 3.71, 95%c.i. 2.3-5.98); four quadrants involved (O.R. 1.95, 95%c.i. 1.22-3.11); and inferior breaks (O.R. 1.76 95%c.i. 1.22-2.55). The risk factors were used to develop a complexity score. Although eyes with a score >0 were more likely to require further surgery (Fisher’s exact test, p<0.0001), 40% of failures occurred in eyes with no identifiable risk factors. Eyes with a macula off detachment treated less than four days after losing vision, were more likely to regain driving vision (LogMAR 0.3) than those operated later (Fisher’s exact test, p=0.027)
Conclusions:
Our study shows that large amounts of data can be obtained prospectively using online data collection. This innovative tool is useful for developing models to identify retinal detachments at high risk of failure, and to allow adjustment of outcomes for complexity. As data collection is ongoing, we anticipate refinements to our model that will improve its accuracy.