Aims: To compare anatomical and functional outcomes for 546 phakic and pseudophakic primary rhegmatogenous retinal detachments (RDs) treated by pars plana vitrectomy or scleral buckling.
Methods: Prospective, non-randomised, interventional study in 15 centres in Spain and Portugal, with data from RDs consecutively treated from January 2005 to May 2007. Cases with preoperative proliferative vitreoretinopathy grade C-1 or higher and perforating trauma were excluded. Minimum follow-up was 3 months. Twenty-seven pre-, intra- and post-surgical variables for each patient were analysed. Multivariate analysis was carried out by logistic regression analysis with stepwise selection of variables.
Results: Data from 546 patients were analysed. Global anatomical success was 94.7%. Logistic regression analysis showed that only the development of postoperative proliferative vitreoretinopathy was associated with a poor anatomical outcome. The poorest functional results were associated with macular involvement, extension of RD, previous RD surgery, time of evolution of RD, and age of patient. Hierarchical log-linear analysis showed no effect of the lens status (phakic versus pseudophakic) on the functional results. However, pars plana vitrectomy was most often performed in pseudophakic eyes and resulted in a worse final visual acuity (p<0.001).
Conclusions: No differences in anatomical success between phakic and pseudophakic eyes were found in this series. Pars plana vitrectomy was most often performed in pseudophakic eyes and had a greater probability of a worse final visual acuity than scleral buckling.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The Retina 1 Project is a prospective multicentric study designed to improve a published formula1 for predicting the risk of developing postoperative proliferative vitreoretinopathy (PVR) in surgical patients with rhegmatogenous retinal detachments (RDs). The basis for making the prediction relies on pre-, intra- and post-surgical factors. The purpose of the project is twofold: to identify patients with higher risk of developing post-surgical PVR, and to propose different surgical approaches or complementary medical treatments to reduce the risk of PVR development. Retina 1 is based on the assessment of 83 variables obtained before, during and after surgery.1 To achieve this objective, 15 centres in Spain and Portugal have collected data on RD operations since January 2005. Thus, prospective and highly valuable information on the characteristics of RD and its management and outcomes is now available.
During the last few years, several papers have compared the anatomical and functional outcomes for RDs treated with scleral buckling (SB) versus vitrectomy2 with special emphasis on pseudophakic RDs.3–6 The data suggest that there are no significant differences in the anatomical outcome between the two techniques. However, some controversy exists about the functional outcome. Few studies have focused on functional results, and most of them are retrospective studies from a single centre with few surgeons involved.7 8 Thus, the purpose of this report is to compare the anatomical and functional outcomes in 546 primary RD phakic and pseudophakic cases obtained from the centres participating in the Retina 1 Project. These data can be used to improve treatment strategies on the basis of pre-, intra- and post-surgical information for each patient.
Fifteen institutions in Spain and Portugal participated in this prospective, non-randomised, interventional, multicentric study. The study protocol was approved by the research committee of the coordinating institution (IOBA, University of Valladolid, Spain). Informed consent was obtained from each patient before inclusion of his/her data in a common database.
Inclusion and exclusion criteria
All patients seen between January 2005 and May 2007 with primary rhegmatogenous RD with a follow-up of 3 months were considered for inclusion. Cases with preoperative PVR grade C-1 or higher according to the Retina Society classification were excluded.9 Cases in which the RD was due to perforating injury were also excluded. Each centre was allowed to establish its own criteria for offering the SB procedure or pars plana vitrectomy (PPV) with or without SB to their patients according to the experience of the surgeon.
Only 27 out of the total 83 characteristics gathered in the Retina 1 Project1 were used for analysis in this report. Those that were constant, non-informative or provided little information for predicting the development of PVR were eliminated. Other characteristics were regrouped as unique variables. Anatomical success and final visual acuity (VA) were the dependent variables. According to our experience, 16 pre-surgical, eight intra-surgical and three post-surgical variables were considered to be the presumed causes of changes in these outcomes. Anatomical success was defined as a reattached retina at the last postoperative visit at 3 months. Functional outcome was determined in the final visit when the patient’s VA was measured as the best corrected Snellen distance. The data were then grouped into three ranges: VA ⩽20/40, VA between 20/50 and 20/100, and VA <20/100.
Descriptive statistics were determined for demographic and ocular characteristics. Quantitative results are expressed as mean (SD), and qualitative variables are expressed as percentages. χ2 analyses were performed to evaluate the association between categorical RD characteristics and anatomical and functional outcomes. The Fisher exact test was used with sparse contingency tables. To study differences in continuous variables, the Student t test for two independent samples and analysis of variance methods were used. Hierarchical log-linear models were used to test relationships between visual outcomes, phakic and pseudophakic eyes, and surgical technique. Multivariate analysis was carried out by logistic regression analysis with stepwise selection of variables. All variables with p<0.1 in the univariate analysis were entered into the multivariate model. Statistical significance was established at the 0.05 confidence level. Statistical analyses were performed using SPSS V13 (SPSS, Chicago, Illinois, USA) and proc LOGISTIC in SAS statistical software V8 (SAS Institute, Cary, North Carolina, USA).
Information from 546 patients with a mean (SD) age of 56.7 (15.2) years was analysed. The time between onset of RD (defined by the beginning of early signs and symptoms including flashes and floaters) and surgery was 15.5 (34) days. Preoperative intraocular pressure (IOP) of the affected eye was 13.6 (4.1) mm Hg. Table 1 shows the most relevant pre-, intra- and post-operative factors associated with a higher risk of developing post-surgical PVR.
Global anatomical success was achieved in 94.7% of the cases. Univariate analysis did not show any statistically significant differences between anatomical success and the following preoperative variables: age, presence of preoperative PVR grades A or B, type of retinal break, previous RD surgery, existence of vitreous haemorrhage, status of the macula, time between onset of symptoms and surgery for RD, aphakia or pseudophakia, existence of posterior vitreous detachment, preoperative uveitis, previous ocular trauma, or myopia. Regarding the extension of the RD, the lowest reattachment rate occurred when the RD extended over four quadrants (table 2). A lower reattachment rate was also found in cases with previous intraocular surgery other than RD surgery (table 2).
On the basis of univariate analysis, preoperative IOP was significantly higher in cases in which the retina was reattached (13.7 (4.1) mm Hg) than in those with reattachment failure (11.6 (4.4) mm Hg) (p = 0.009). However, this relationship was not supported by multivariate analysis. IOP was related to the extension of the RD. For cases with up to three preoperative quadrants affected by the RD, the IOP was 13.9 (4.1) mm Hg, and for cases with four affected quadrants, it was 11.9 (4.6) mm Hg (p = 0.001).
None of the analysed intraoperative factors were associated with different reattachment rates; thus, anatomical success was not different between eyes undergoing PPV and those submitted to SB alone. RD in aphakic and pseudophakic eyes (n = 185) was treated almost exclusively with PPV alone or in combination with SB (n = 182). Of the postoperative factors, only the development of a PVR was associated with a lower anatomical success rate (table 2). This association was supported by logistic regression analysis (p<0.001)*.
For these analyses, only cases with reattached retinas at 3 months after surgery were considered (n = 517). Of these, 42.7% reached a final vision better than or equal to 20/40, and 36% achieved a VA between 20/50 and 20/100. Finally, 21.3% had a VA lower than 20/100. The type of retinal break, existence of vitreous haemorrhage, aphakia or pseudophakia, existence of a posterior vitreous detachment, values of preoperative IOP, uveitis, myopia and trauma did not significantly affect the final VA.
Age correlated inversely with the final functional result (p<0.001; table 3). The mean age of patients with a final VA ⩾ 20/40 was 53.2 (15.6) years. For a final VA between 20/50 and 20/100, the mean age was 58.5 (13.8) years, and for a final VA lower than 20/100, the mean age was 59.6 (15.9) years.
The time between onset of symptoms and surgery for RD was also associated with the functional result (p<0.001; table 3). The mean duration was 10.8 (25.9) days in patients with VA ⩾ 20/40, 13.9 (15.3) days in patients with a final VA between 20/50 and 20/100, and 27.9 (61.2) days in those with final VA lower than 20/100. Preoperative PVR grades A or B, extension of RD, previous RD surgery, macular involvement, and previous intraocular surgery other than RD were also significantly associated with worse final VAs (table 3).
Regarding the intraoperative factors, the use of encircling bands and/or scleral sponges (both as independent variables), drainage of subretinal fluid, and endotamponade with gas were not associated with functional results. SB surgery was associated with better VA results, whereas PPV, intraoperative use of perfluorocarbon liquids, and the execution of drainage retinotomies were associated with worse final VAs (table 3). Post-surgery factors such as a severe intraocular inflammatory response, and peaks of ocular hypertension were not associated with functional results.
On the basis of logistic regression analysis, characteristics associated with a worse functional result were: RD with macula involved (macula-off), RD involving more than three quadrants, preoperative PVR (grade A or B), unsuccessful previous attempts to reattach the RD, and long lapse of time between the development of RD and the surgical procedure. Also, older patients have the worst visual prognosis (table 4). A hierarchical log-linear analysis was performed to determine the influence of status of the lens (phakia/pseudophakia), PPV and SB surgery on the functional results. The lens status was not significantly associated with the visual outcome (p = 0.098). However, the technique chosen by the surgeon was related to the lens status. Thus PPV was most often performed in pseudophakic eyes (p<0.001), and SB surgery was most often performed in phakic eyes (p<0.001). In a global analysis, the surgical technique influenced the final VA. Thus PPV was more strongly associated with a poor visual outcome (p<0.001) than was SB surgery (p = 0.022).
This study was specifically designed to improve the accuracy of our predictive formula for PVR in patients with RD,1 but prospective data obtained from different centres in two countries with public health services provided a general idea on the anatomical and functional results obtained in the routine management of RD with a variety of pre-, intra- and post-surgical conditions.
The anatomical success rate obtained with either PPV or SB surgery was slightly higher than some series2 10–12 and lower than others.8 13 Our results did not show the significant differences for anatomical outcome that were obtained by five of the participating centres 5 years ago.14 In the series reported here, there were no differences in anatomical reattachment rate between cases treated by PPV (93.5%) and those treated by SB surgery alone (97.8%), confirming the results of a previous series.15
Regarding the predictive factors for anatomical failure, our results agree with those obtained by Ahmadieh et al12 in a retrospective study. Extension of RD and preoperative PVR are common crucial factors in both series. The presence of multiple breaks, a risk factor found by Afrashi et al,13 was not a predictive factor in our series or that of Ahmadieh et al.12
Lower IOP is a negative predictive factor for reattachment of the retina. This relationship has been reported previously14 16 and seems to be related to the extension of RD. As expected, the factor most strongly related to a higher anatomical failure rate was postoperative development of PVR, confirming that this complication continues to be the major cause of failure after RD surgery.11 17
More than 42% of eyes with reattached retinas had a VA ⩾ 20/40, and more than 21% had a VA <20/100. These values are similar to some series8 11 18 and better than others.15 The time between the onset of symptoms and surgery for RD8 11 12 19 may explain some of these differences between series. In our series, the mean time was 15.5 days, which is similar to other recently published ones.20
Some factors, most of which have previously been identified, have a significant association with the final functional results. These include the age of patients,12 21 22 preoperative PVR,8 18 20 extension of RD,8 11 previous RD or other ocular surgery,18 and the status of the macula.12 20 22 In our series, only 35% of the RDs had macula-on at the time of surgery.
Among the intraoperative characteristics, the type of surgery, the use of perflurocarbon liquids during PPV, and the performance of a drainage retinotomy were significantly associated with the existence of more complex RDs and consequently poorer functional results. Conversely, less complicated RDs can be treated with simpler techniques and usually result in excellent visual outcomes.23 24
The association of PPV with a worse functional outcome than SB surgery has been a matter of controversy.2 5 6 15 25–27 This topic was recently approached through a meta-analysis4 of papers published between 1966 and 2004. The conclusion was that, for uncomplicated pseudophakic RD, better anatomical and functional outcomes were achieved with PPV than with SB. However, the authors pointed out the limitations of the study, suggesting the need for a well-designed prospective, randomised, controlled clinical trial.4 Our results do not support a better anatomical outcome of one technique over the other. Rather, they suggest that PPV is associated with a higher probability for the final VA to be <20/100.
Using a hierarchical log-linear model, we determined that the presence or absence of a lens per se was not related to the visual outcome. However, the surgical technique chosen to resolve the RD was associated with the lens status. PPV was more often performed in aphakic or pseudophakic eyes (98.7%) than in phakic eyes (73.6%). Thus the final VA was associated with the lens status through the type of surgery, with PPV having a worse functional outcome than SB.
Whereas univariate analysis revealed some significant associations of variables with the anatomical outcome, multivariate analysis showed that only the postoperative development of PVR affected that result. For functional outcomes, multivariate analysis revealed that the age of the patients, preoperative PVR, extension of the RD, status of the macula, and time of evolution of the RD were important factors.
In conclusion, we have described a large sample of RDs from a multicentric prospective study, showing pre-, intra- and post-operative characteristics that can determine the anatomical and functional results. No differences in anatomical success between phakic and pseudophakic eyes were found; however, patients who had PPV had a greater probability of a worse final VA. These results can be used as a reference for hospitals with general ophthalmology units such as those working in the European National Public Health Systems.9
Clinica Barraquer (Barcelona, Spain), Fundación Oftalmológica del Mediterraneo (Valencia, Spain), Hospital Clínico Universitario (Valladolid, Spain), Hospital de Donostia (San Sebastián, Spain), Hospital Nuestra Señora de Sonsoles (Avila, Spain), Hospital Ramón y Cajal (Madrid, Spain), Hospital Reina Sofia (Córdoba, Spain), Hospital San Millán (Logroño, Spain), Hospital Sao João (Porto, Portugal), Hospital Universitario de Canarias (Santa Cruz de Tenerife, Spain), Hospital Universitario de Coimbra (Coimbra, Portugal), Hospital Universitario del Rio Hortega (Valladolid, Spain), Hospital Universitari Vall de Hebrón (Barcelona, Spain), Instituto Oftalmológico de Alicante (Alicante, Spain), and Instituto Universitario de Oftalmobiología Aplicada (Valladolid, Spain).
Competing interests: None declared.