Scleral buckling and vitrectomy are the most popular techniques by which to treat non-complicated rhegmatogenous retinal detachment (RD). In recent years, there have been many reports trying to demonstrate the superiority of one over the other,1^–3 but published data show no significant differences.1 Thus, there is no evidence that either surgical approach is better than the other, as reattachment rates are similar using either of these techniques.2 4 5

We have the impression that management of RD has experienced considerable changes in recent years despite the lack of clear evidence to support that variation. Probably because surgeons have greatly increased their experience and skills to perform vitrectomy, this option is gaining popularity as a first approach to RD,6^–8 although it is more expensive and sometimes more invasive.

The aim of this study was to assess variations in the management and characteristics of two series of RD, operated on by the same surgeons—the second series collected 4 years after the first one—and to compare anatomical results between them.

METHODS

A prospective, multicentric study involving five Spanish hospitals was carried out. Surgeons from these hospitals are currently participating in a multicentric study to identify risk factors in proliferative vitreoretinopathy (PVR) (Retina 1 project). Clinical characteristics of all consecutive RD treated by them since 1999 are recorded in a database. All surgeons in Retina 1 project are highly experienced in RD treatment and are allowed to treat patients following their personal criteria.

Only RDs without preoperative PVR or with PVR grade A or B are included and post-traumatic RDs are excluded from the study. In order to evaluate the anatomical results, cases without at least 3 months of follow-up are also excluded.

To assess the variations in the management of RD in recent years, cases were obtained from Retina 1 database and were divided into two groups: Group 1 (G1) (n = 186) includes consecutive cases operated on in the participating hospitals from March 1999 to April 2001; Group 2 (G2) (n = 153) includes cases from November 2004 to June 2006.

Eighty-three variables related to preoperative characteristics, surgical management and postoperative evolution were recorded (table 1 shows the main variables studied).

A comparison between preoperative characteristics, surgical management and anatomical results was carried out. Continuous variables were expressed as mean (SD). Qualitative variables were expressed as percentages. Continuous variables were compared by Mann–Whitney U test, and qualitative variables were compared with the χ2 test. The Fisher exact test was used on a sparse contingency table. In order to evaluate the importance of RD characteristics for the surgical management and anatomical success in each group, a multivariate analysis was carried out by logistic regression analysis with progressive backward elimination of variables. Statistical significance was established at the 0.05 confidence level.

RESULTS

Table 2 shows the main preoperative characteristics in both groups. No significant differences in the rate of some relevant variables such as vitreous haemorrhage, pseudophakia/aphakia or previous failed attempts to repair the RD were observed. However, G1 showed a significantly longer delay in performing surgery since the first symptoms appeared (G1: 29 (SD 50) days; G2: 22 (55); p<0.001) and more RD without visible retinal break than G2 (G1: 17.4%; G2: 9.2%; p = 0.028). In G2, cases with multiple retinal breaks (G1: 31.6%; G2: 44.6%) were more frequent (p = .022).

In aphakic/pseudophakic patients, G1 also showed a longer delay in performing surgery since the first symptoms appeared (G1: 25.24 (24.24) days; G2: 14.23 (18.83); p = 0.004). No significant differences in other preoperative variables were observed in aphakic/pseudophakic cases.

Differences in the surgical management of RD were observed. As a primary surgical approach, a vitrectomy was performed in 56 out of 186 cases (30.1%) in G1, and a vitrectomy in G2 was performed in 120/153 (78.4%) (p<0.001). In G1, when a vitrectomy was performed, it was associated with a scleral band in 83.9% of cases, whereas in G2 only 41.2% was associated with a band (p<0.001). The use of cryotherapy in association with vitrectomy was also more frequent in G2 (G1: 5.4%; G2: 20.3%; p = 0.029). In vitrectomy cases, no differences in laser use were observed (G1: 98.2%; G2: 93.3%; p = 0.275).

Regardless of the RD characteristics, the rate of vitrectomy was higher in G2. Except in circumstances such as previous failed attempts of reattachment or RD associated with vitreous haemorrhage, the rate of vitrectomy was similar in both groups (G1: 54.4%, 6/11; G2: 85.7%, 12/14; Fisher exact p value = 0.177).

In G1 vitrectomy was mainly associated with non-visible retinal breaks (OR: 0.110; 95% CI: 0.029 to 0.424) and vitreous haemorrhage (OR: 3.410; 95% CI: 1.104 to 10.538). In G2, cases with PVR grade A or B (OR: 6.198; 95% CI: 1.344 to 28.582) and pseudophakia/aphakia (OR: 28.694; 95% CI: 3.762 to 218.864) underwent vitrectomy.

No significant differences in anatomical results were observed. A total of 174 out of 185 cases (94.1%) in G1 and 140 of 148 cases (94.6%) in G2 showed retinal reattachment after 3 months of follow-up (p = 0.833). One patient in G1 and five in G2 were missed.

The statistical analysis showed differences only in patients with pseudophakic RD: pseudophakic patients showed better anatomical results in G2 than in G1 (G1: 83.9%; G2: 96.4%; p = 0.028).

The rate of reattachment in cases with undetected breaks was 90.6% in G1 (n = 32) and 100% in G2 (n = 4). This difference was not significant (p = 0.543)

DISCUSSION

Important differences in preoperative characteristics of RD in Spain were observed between 1984 and 1998.9 The most relevant variations between those years were the great increase in RD in patients with pseudophakia (11.6% in 1984 and 36.3% in 1998) and the higher rate of cases with concomitant vitreous haemorrhage (2.3–6.1%).9 To some extent, those changes were related to the refinement of cataract surgery and the huge increase in cataract operations in that period of time in Spain. Similar changes in clinical presentation of RD were reported in UK.10

In contrast with those variations of RD observed during the 1990s, a comparison of both series of RD included in this study (G1 and G2) demonstrates that RD characteristics have not clearly changed in recent years. Although there are reports showing differences in preoperative characteristics of RD related to racial factors and habits,11 G1 and G2 can be considered quite similar to other RD series reported in western countries in recent years.10 12

According to the statistical analysis, there are slight differences between both series that do not explain the changes in management. The efforts developed in the last years to improve results of RD treatment in Spain have probably contributed to a reduction in the delay between first symptoms and surgical procedure. It is difficult to reduce the time between the appearance of initial symptoms of RD and the patient’s first consultation, but a better knowledge of RD symptoms by General Practitioners that refer patients to hospitals involved in Retina 1 project can be accounted for the reduction of that time. The National Survey on RD management in Spain could also have contributed to this variation by making the general ophthalmologist more sensitive towards the importance of time in RD treatment.13 Further improvements in this field are necessary to achieve better results in the future.

Identification of retinal breaks was performed both pre- and intraoperatively. There is no clear explanation as to why fewer cases without visible breaks and more frequency of multiple breaks were found in G2. Surgeons participating in Retina 1 project are encouraged to carry out a thorough examination of patients before they complete the database. A more meticulous examination could help explain these differences. The higher rate of vitrectomy could also explain this fact, because during vitrectomy it is not unusual to find small holes that were missed in the preoperative examination, mainly in pseudophakic cases with peripheral opacities.

Surgeons in the Retina 1 project are free to select the best treatment option for their patients according to their experience, and a substantial change in their attitude has been observed in the last 4 years. This change was not justified by a better anatomical outcome—over 90% in G1—and no significant improvement has been achieved in G2.

In G1, vitrectomy was reserved to circumstances where a scleral buckle was difficult or impossible to perform. However, in G2, vitrectomy was the first approach in most cases of RD. Only in cases of dense vitreous haemorrhage or failure of a previous reattachment attempt was the rate of vitrectomy similar in both series. In all other circumstances, vitrectomy is significantly more frequent in G2 than in G1 (even in cases with non-visible or multiple breaks). However, the results do not show that this preference for vitrectomy has improved the anatomical success, except in pseudophakic cases.

The optimal approach to pseudophakic/aphakic RD is still controversial, and some randomised studies provide contradictory conclusions.14 15 However, in recent years, there has been a growing tendency to treat these cases with a vitrectomy.16 17 The impression that vitrectomy could provide better results in pseudophakic/aphakic RD is supported by most surgeons involved in the Retina 1 project,8 18 and this explains the fact that in G2, all pseudophakic/aphakic RD were treated with vitrectomy. This study is not designed to clarify the best management of pseudophakic/aphakic RD, but our data seem to justify the trend to use vitrectomy in such patients.

In recent years, some surgeons participating in this study—and other authors as well19 20—have published good results in cases treated with vitrectomy without associated scleral buckle.21 This fact can explain the tendency to perform vitrectomy without scleral buckle in G2. There is no doubt that some cases could benefit from this approach, as it reduces the surgery time and potential complications, but further data are necessary in order to compare its results with other techniques.

It is remarkable that pneumatic retinopexia has not been used either in G1 or in G2. In addition, vitrectomy using 25G or 23G has not been performed, although this technology is easily available in Spanish hospitals.

Unfortunately, postoperative visual acuity was not recorded in G1, so a comparison of functional results cannot be provided.

In summary, there is an increasing tendency to treat RD with primary vitrectomy in the last years, not related to a higher complexity of cases or to better anatomical results. Only pseudophakic RDs seem to benefit from this change.

This study is not designed to obtain evidence of the best management of RD, but it does provide a real picture of the variations that are taking place to treat it in the last few years. Further research is necessary to establish whether these variations are justified.