Primary vitrectomy versus conventional retinal detachment surgery in phakic rhegmatogenous retinal detachment


Dr Raj Vardhan Azad
All India Institute of medical Sciences
Ansari Nagar
New Delhi 110029
Tel: + 91 11 2956 1541
Fax: + 91 11 2658 8919


Purpose:  This study aimed to compare the results of primary vitrectomy and conventional scleral buckling procedures (conventional retinal detachment surgery) in phakic rhegmatogenous retinal detachment (RRD).

Methods:  We carried out a randomized, prospective, clinical controlled trial of 61 consecutive phakic eyes with primary RRD, not complicated by proliferative vitreoretinopathy ≥ grade C. Subjects were randomized to either scleral buckling (group 1) or pars plana vitrectomy (group 2).

Results:  At 6 months follow-up, the primary reattachment rate was 80% (24/30 cases) in group 2 and 80.6% (25/31 cases) in group 1; the difference between the two groups was not statistically significant (p = 0.213). Best corrected visual acuity improved significantly from a preoperative median of 1.78 (1/60) (mean 1.73 ± 0.91, range 0.3–3) to a median of 0.6 (6/24) (mean 0.689 ± 0.35, range 0.18–1.48) in group 2 and from a preoperative median of 1.48 (2/60) (mean 1.43 ± 0.92, range 0–3) to a median of 0.6 (6/24) (mean 0.608 ± 0.36, range 0–1.78) in group 1; the difference between the two groups was not statistically significant (p = 0.376). Cataract developed in five cases (17%) in the vitrectomy group (group 2), with a statistically significant difference of p = 0.018.

Conclusions:  Although primary vitrectomy can achieve anatomical and functional success rates comparable with those achieved by scleral buckling in uncomplicated forms of phakic RRD, the major drawback of the procedure is the high incidence of postoperative cataract formation. Moreover, visual rehabilitation takes place earlier with scleral buckling than with vitrectomy. Scleral buckling should thus be used as the primary surgical modality in the treatment of uncomplicated RRD where the media are sufficiently clear.


Although scleral buckling (SB) is an effective surgical procedure for treating uncomplicated rhegmatogenous retinal detachment (RRD), its intraoperative and postoperative complications, which ultimately impair visual function, cannot be ignored. Extrusion, intrusion or infection of exoplant material, motility problems (Arruga 1977), refractive changes (Rubin 1975), anterior segment ischaemia, distortion of the macula, and postoperative cystoid macular oedema (Lobes & Grand 1980) are associated with SB. Moreover, external drainage of subretinal fluid may result in subretinal haemorrhage, retinal incarceration or a retinal break.

Primary vitrectomy offers an attractive alternative in these cases because it affords a direct approach to vitreous traction using microsurgical techniques (Kloeti et al. 1983; Escoffery et al. 1985). The results, however, are equivocal, although a few studies have reported good results with pars plana vitrectomy (PPV) (Van Effenterre et al. 1987; Gartry et al. 1993; Hakin et al. 1993; Woon et al. 1995; Heimann et al. 1996; Oshima et al. 1999, 2000; Miki et al. 2000).

We undertook this study to evaluate the role of primary PPV in the management of RRD in phakic eyes that were not complicated by proliferative vitreoretinopathy (PVR) ≥ grade C. The study represents a randomized, prospective comparison of conventional SB techniques and primary vitrectomy, with respect to final anatomical and functional success rates and rates of complications.

Materials and Methods

In this study of primary vitrectomy versus conventional retinal detachment surgery, a randomized, prospective comparison of SB techniques and primary vitrectomy was carried out in phakic patients with RRD. The clinical trial was aimed at improving the surgical therapy of phakic RRD in terms of both functional and anatomical outcome.

A total of 61 consecutive phakic eyes with primary RRD, not complicated by PVR ≥ grade C and with a clear lens, were included in the study. The patients were selected from among users of the vitreo-retina services at our institute from October 2003 to March 2005. All cases with significant media opacities such as vitreous haemorrhage which made SB impossible, PVR ≥ grade C, no-break retinal detachment, aphakia or pseudophakia, history of any previous intraocular surgery, traumatic retinal detachment, and any retinal detachment deemed to be unmanageable by conventional SB surgery were excluded from the study.

Prior approval was obtained from the Ethics Committee of the All India Institute of Medical Sciences, New Delhi. Informed consent was obtained from all patients prior to the surgical procedure.

All patients were randomly assigned to one of two treatment arms. Each patient was given one of a set of sequentially numbered randomization envelopes containing one of the two surgical options; the patient then received the treatment detailed in the envelope.

All surgical procedures were performed under local peribulbar anaesthesia by a single surgeon (RVA). When SB was performed, chorioretinal adhesions were achieved with cryopexy around the retinal breaks. An exoplant (segmental 7-mm silicone tire in combination with an encircling 2.5-mm silicone band) was used to support peripheral retinal breaks. External modified needle drainage (Azad et al. 1997) of subretinal fluid was performed in all eyes and, if necessary, intravitreal gas (air or 18% sulphur hexafluoride [SF6] or 14% perfluoropropane [C3F8]) was used to maintain intraocular pressure (IOP).

All eyes in the primary vitrectomy group underwent a three-port PPV, using an operating microscope with a standard set of lenses. A complete vitrectomy was performed, relieving the vitreous traction around the break and allowing for removal of the anterior vitreous gel using wide-angle lenses and scleral indentation. Posterior vitreous detachment (PVD) was induced when not already present and vitrectomy completed. This was followed by a fluid−air exchange with internal drainage of subretinal fluid through either break or retinotomy. Retinopexy was achieved by diode laser endophotocoagulation (360-degree along with delineation of the break and retinotomy). An air−C3F8 gas exchange was carried out using a non-expansile mixture of 14% C3F8 gas to achieve a complete fill of the vitreous cavity. C3F8 gas was used because it provides a longterm tamponade compared with SF6. Patient compliance could not be relied upon, and hence, to be on the safe side, a longer-acting gas was used.

In both treatment groups, patients injected with gas were instructed on head positioning to encourage tamponade of the retinal breaks during the first 2 weeks postoperatively. Topical antibiotics and anti-inflammatory and cycloplegic drops were applied postoperatively for 1 month, according to our department's protocol. All eyes were evaluated on day 7, and at 1, 3 and 6 months of follow-up. The parameters evaluated included best corrected visual acuity (BCVA), IOP (always measured by applanation tonometry), lens status, and fundus examination. Postoperative complications were also recorded. Only patients who completed 6 months of follow-up were included in the final analyses.

Statistical analysis

Contingency tables were drawn for the variables in the two groups and analysed using Pearson's chi-square test. For continuous variables such as VA in logMAR and IOP, Student's t-test for statistical significance was used. The statistical analysis was performed using spss Version 10 (SPSS, Chicago, IL, USA). For statistical significance, p was set at 0.05.

Visual acuity was analysed for study purposes by means of the logarithm of the minimal angle of resolution (logMAR) score, developed by Ferris et al. (1982). Counting-fingers and hand-motion VAs were converted to Snellen equivalents according to the method proposed by Holladay (1997) and were given values of 0.01 (+ 2.0 logMAR) and 0.001 (+ 3.0 logMAR), respectively.


The SB and PPV groups were similar with respect to the preoperative factors of age, sex, laterality of the eye, duration of retinal detachment, configuration of retinal detachment, macular status, presence of risk factors, PVR status, number of breaks, preoperative VA and IOP; there were no statistically significant differences between the two groups (Table 1). All cases were followed up for a minimum of 6 months.

Table 1.   Basic characteristics of patients in the two treatment groups at the beginning of study.
 Buckle group
(n = 31)
Vitrectomy group
(n = 30)
  1. SD = standard deviation; RD = retinal detachment BCVA = best corrected visual acuity; PVR = proliferative vitreoretinopathy; IOP = intraocular pressure; PVD = posterior vitreous detachment.

Age, years (mean ± SD)36 ± 16 years41 ± 15 years0.219
Sex, % (n) Male
74% (n = 23)
26% (n = 8)
57% (n = 17)
43% (n = 13)
Laterality of eye, right eye, % (n)58% (n = 18)50% (n = 15)0.527
Duration of RD, mean (range)19 days (2–90 days)20 days (2–60 days)0.996
Configuration, total RD, % (n)10% (n = 3)23% (n = 7)0.150
Macula-on, % (n)16% (n = 5)13% (n = 4)0.758
LogMAR, BCVA mean ± SD, median1.43 ± 0.92, 1.481.73 ± 0.91, 1.780.208
PVR status, % (n) Absent
          Stage A/B
87% (n = 27)
13% (n = 4)
90% (n = 27)
10% (n = 3)
Risk factors present, % (n)51% (n = 16)50% (n = 15)0.90
 Myopia (> − 6 D)22.5% (n = 7)20% (n = 6) 
 Peripheral degeneration22.5% (n = 7)13.3% (n = 4) 
 Both6% (n = 2)16.7% (n = 5) 
Preoperative IOP, mmHg, mean ± SD11.7 ± 2.511.7 ± 4.00.990
Number of breaks 171% (n = 22)73% (n = 22)0.085
                   23% (n = 1)17% (n = 5) 
                  ≥  326% (n = 8)10% (n = 3) 
Total PVD68% (n = 21)63% (n = 19)0.717

A primary success rate (Table 2) of 80% (24/30 cases) was achieved in the PPV group compared with 80.6% (25/31 cases) in the SB group; the difference between the two groups was not statistically significant (p = 0.213). An assessment of the causes of surgical failure in the PPV group showed open breaks (new or missed breaks) in three cases, PVR in two cases and both PVR and open breaks in one case. In the SB group surgical failures resulted from open breaks (new or missed breaks) in four cases and improper buckle positioning in two cases. The causative open breaks in both groups were found during a subsequent PPV procedure. The final anatomical success rate at the end of 6 months was 100% in both groups.

Table 2.   Anatomical and functional outcomes observed in the two groups.
VariableVitrectomy groupBuckling group
  1. PVR = proliferative vitreoretinopathy.

% surgical primary success80% (24/30)80.6% (25/31)
(p = 0.213)
Cause of failure3 cases: open break4 cases:open break
2 cases: PVR2 cases: improper
1 case: open break + PVRbuckle position
Final visual acuityMedian 0.6 (6/24)Median 0.6 (6/24)
Range 1.48–0.18
Range 1.78–0.0

There was no statistically significant difference between the two groups in terms of final BCVA (p = 0.376). However, BCVA in the SB group at 1 week (median 0.78, mean 0.84 ±0.45, range 0.18–1.78) and 1 month (median 0.78, mean 0.699 ± 0.41, range 0–1.78) post-surgery was found to be significantly better than in the PPV group at 1 week (median 2, mean 2.0 ± 0.69, range 0.78–3) and 1 month (median 1, mean 1.14 ± 0.74, range 0.3–3) (p = 0.000 at 1 week; p = 0.006 at 1 month) (Fig. 1).

Figure 1.

 Visual acuity trends in both groups. SB = scleral buckling; PPV = pars plana vitrectomy.

The final BCVA (Table 3) in the PPV group improved from a preoperative median of 1.78 (1/60) (mean 1.73 ± 0.91, range 0.3–3) to a median of 0.6 (6/24) (mean 0.689 ± 0.35, range 0.18–1.48). A total of 17 eyes (57%) achieved vision > 0.6 (6/24), whereas only one eye showed a drop of ≥ 1 line of Snellen acuity. Two eyes (7%) maintained their preoperative VA. In the SB group, final BCVA improved from a preoperative median of 1.48 (2/60) (mean 1.43 ± 0.92, range 0–3) to a median of 0.6 (6/24) (mean 0.608 ± 0.36, range 0–1.78). Twenty eyes (65%) achieved vision > 0.6 (6/24), whereas two eyes showed a drop of ≥ 1 line of Snellen acuity. Four eyes (13%) maintained their preoperative VA.

Table 3.   Comparative evaluation of pre- and postoperative best corrected visual acuity in scleral buckling and pars plana vitrectomy groups.
DurationBest corrected visual acuity (logMar)
Buckle group
(n = 31)
Vitrectomy group
n = 30)
Mean ± SDMedianMean ± SDMedian
  • *  

    p-value significant.

  • SD = standard deviation.

Preoperative1.43 ± 0.921.481.73 ± 0.911.780.208
1 week0.84 ± 0.450.782.0 ± 0.6920.000*
1 month0.699 ± 0.410.781.14 ± 0.7410.006*
3 months0.676 ± 0.440.60.773 ± 0.440.60.395
6 months0.608 ± 0.360.60.689 ± 0.350.60.376

The major intra- and postoperative complications encountered during the study are summarized in Table 4. Intraoperatively, there were three (10%) cases of hypotony and fish-mouthing during scleral buckling, for which air or gas was injected. None of the cases had scleral perforation, vitreous loss or retinal incarceration. Iatrogenic breaks occurred in three cases (10%) in the PPV group while completing the vitreous base dissection; the breaks were delimited by laser photocoagulation. Postoperatively, in the SB group there were five cases (16%) with residual fluid, two of which progressed and required further surgery. Two cases (6%) had raised IOP (defined as IOP > 21 mmHg on at least two occasions in the first postoperative week), which was controlled on medical management. One case (3%) developed a persistent epithelial defect that healed with the use of a bandage contact lens. Buckle infection was seen in one case (3%) and one case (3%) developed epiretinal membrane (ERM). In the PPV group, five cases (17%) developed cataract, with a statistically significant difference (p = 0.018). Raised IOP was present in two cases (7%) and controlled on medical management and one case (3%) developed ERM.

Table 4.   Complications observed in the two groups.
ComplicationsNo. of cases
Vitrectomy group
 Iatrogenic breaks3 (10%)
 Cataract5 (17%)
 Raised intraocular pressure2 (7%)
 Epiretinal membrane1 (3%)
Buckling group
 Fish-mouthing/hypotony3 (10%)
 Residual fluid5 (16%)
 Raised intraocular pressure2 (6%)
 Persistent epithelial defect1 (3%)
 Buckle infection1 (3%)
 Epiretinal membrane1 (3%)


We conducted a randomized, prospective, controlled clinical trial comparing primary vitrectomy and conventional retinal detachment surgery with respect to anatomical and functional outcomes in primary RRD complicated by PVR < grade C in phakic eyes. A total of 61 cases of primary RRD, complicated by PVR < grade C, with a clear media, were selected from users of the vitreo-retina services at our institute between October 2003 and March 2005. All eyes were meticulously screened with indirect ophthalmoscopy and fundus biomicroscopy using scleral indentation preoperatively. Postoperatively, all cases were followed up for a minimum period of 6 months as per the study protocol. All analyses and results pertain to the outcome at 6 months post-surgery. Although retrospective studies showing good results with primary vitrectomy have been carried out (Oshima et al. 2000; Miki et al. 2001), this is the first randomized prospective study to compare PPV with SB for uncomplicated RRD in phakic eyes.

In our series, the difference in the primary and final anatomical success rates between the two groups at the last follow-up was not statistically significant. Some studies in the past have had final anatomical success rates of 82−100% (Hakin et al. 1993; Oshima et al. 1999) but the indications for surgery do not match those for SB. Our results are comparable with those of some of the recently published retrospective studies (Oshima et al. 2000; Miki et al. 2001). No significant difference in single procedure reattachment incidence (91%), final success incidence (100%) and incidence of postoperative PVR development (4%) was reported in these studies.

Open breaks (new or missed breaks) were found to be the cause of surgical failure in 3/6 cases in the PPV group and in 4/6 cases in the SB group. Although open retinal breaks were found more frequently in the SB group during the subsequent PPV, it was very difficult to ascertain whether these were fresh retinal breaks or missed ones.

Advanced PVR changes were an important cause of failure in the primary vitrectomy group. Proliferative vitreoretinopathy has been reported to occur in 8–20% of cases undergoing vitrectomy (Escoffery et al. 1985; Gartry et al. 1993; Hoeing et al. 1995; Heimann et al. 1996; Oshima et al. 1998; Miki et al. 2001) and in 5–10% of cases undergoing SB (Girard et al. 1994). In our study, however, PVR was present in none of the cases in the SB group but was detected in three of the vitrectomy eyes. The incidence of PVR is affected by the type of eyes treated. In our study, eyes at high risk of developing PVR were excluded and only eyes with indications for SB were included, which may explain the lower incidence of PVR.

A possible explanation for the occurrence of PVR in the vitrectomy group may be that, following vitrectomy, there is no longer a vitreous tamponade effect; if tears reopen by residual vitreous traction, eyes following vitrectomy are more likely to develop total retinal detachment very early. Because retinal detachment and proliferative changes occur together, PVR may be immediately induced.

The VA results in this series were also comparable with those reported in literature (Wong et al. 1992; Miki et al. 2001; Sharma et al. 2004). In our study, we observed no statistically significant difference in final VA in the two groups. Seventeen eyes (57%) in the PPV group achieved a final VA ≥ 0.6 (6/24), as did 20 eyes (65%) in the SB group.

One of the arguments in favour of primary vitrectomy over SB has been based on the earlier visual recovery it affords (Oshima et al. 2000). However, in the present study, although the difference in final VA achieved in the two groups were not statistically significant, VA at 1 week and 1 month follow-up was significantly better in the SB group. This was to be expected, considering the optical problems created by the C3F8 gas bubble. Thus, visual rehabilitation in eyes in the SB group occurred significantly earlier than in eyes in the PPV group.

Three cases in the SB group required intraocular air or gas. One was injected with air, the second with 18% SF6, and the third with 14% C3F8. Visual recovery was delayed only in the third patient. The effect of the gas bubble on smoothing retinal folds and flattening fish-mouth tears is only required for a short period, hence SF6 or air should be used in SB when required.

Cataract formation has been recorded as a complication of vitrectomy in many studies (Machemer et al. 1971; Stern 1992; Hakin et al. 1993; Heimann et al. 1996; Miki et al. 2000; Oshima et al. 2000; Tewari et al. 2003). The process is more often pronounced in cases where intraocular gas or silicone oil has been used as a tamponade agent. Postoperative cataract formation in this series occurred in five patients in the PPV group, although all patients had a clear lens preoperatively. Four of these patients were aged > 50 years. The occurrence of cataract following vitrectomy is caused by the removal of the retrolental vitreous with direct gas lens contact (Heimann et al. 1996).

Raised IOP was present in two cases in each group. As a non-expansile concentration of C3F8 gas from a standard source was used in the study group, the cause of raised IOP can only be speculated upon. The raised IOP was transient in both groups and controlled by topical antiglaucoma medications or systemic carbonic anhydrase inhibitors, with no serious complications. Elevated IOP is a known complication following frequent fluid−gas exchange in vitrectomy (Bartz-Schmidt et al. 1996).

Our study had some limitations. The intraoperative durations of the two surgical procedures and levels of patient discomfort during surgery and postoperatively were not compared. Although SB is expected to induce greater levels of intra- and postoperative discomfort (pain, redness, lid oedema), vitrectomy is accompanied by the discomfort of postoperative positioning. A discomfort scale designed to include all these factors and applied to test both procedures may help elucidate these questions. The costs involved in the two procedures were also not compared.

The strong points in favour of this study are its prospective nature and the randomization of consecutive patients, and the fact that possible confounding factors were either excluded (giant retinal tears, posterior retinal tears, trauma, unseen breaks) or found not to differ significantly between the two groups (age, duration of the retinal detachment, configuration of retinal detachment, macular status, presence of risk factors, PVR status, preoperative VA).

To conclude, in this small, randomized series, we found that with primary vitrectomy, high final anatomical and functional success rates can be achieved in uncomplicated forms of phakic RRD. However, this would not justify the use of primary vitrectomy over SB techniques if the latter achieved similar results, particularly as cataract formation is a significant disadvantage of vitrectomy in phakic eyes. Visual rehabilitation also occurred significantly earlier in cases managed by conventional SB surgery in our study. A similar multicentre trial is being conducted to compare SB and primary vitrectomy in RRD, although it also includes pseudophakic and aphakic patients (Heimann et al. 2001, 2005), the results of which are awaited. We have conducted a similar study in pseudophakic retinal detachments at our centre (Sharma et al. 2005).

In view of the above, we recommend conventional scleral buckling as the primary surgical modality in the treatment of primary, uncomplicated, rhegmatogenous retinal detachments where the media are sufficiently clear to visualize the fundus using indirect ophthalmoscopy and fundus biomicroscopy. Pars plana vitrectomy may be used as a secondary surgical procedure in cases of redetachment that result after scleral buckling. However, a larger multicentre trial may be required to consolidate the results.