Indications and outcomes of vitrectomy surgery in a series of 1000 black African eyes

Introduction

Vitrectomy as a surgical procedure was first performed in a human eye in 1970, but since then has become very useful in combination with other procedures to treat many vitreoretinal (VR) conditions, restoring sight and preventing blinding complications in many eyes.1 Refinements in technique and instrumentation have also led to confidence in earlier intervention and increasing numbers being performed worldwide.2–4 However, the cost of viewing systems, instruments, consumables and maintenance required for vitrectomy is much higher than for other ophthalmic interventions. As a result, in Africa, there are very few ophthalmology centres providing comprehensive VR service, limiting access for patients who need the service and local training opportunities in this subspecialty. Subsequently data on indications, presentation and outcomes of vitrectomy in this population are few and far between. A key paper, in 2002, reviewing retinal detachment outcomes in East Africa by Yorston et al,5 highlighted the late presentation of patients in Africa and advocated the need for more VR services. However there have been few papers since that time. One of such studies is the paper from Ethiopia, which describes the clinical presentation of retinal detachment.6

This study aims to capture the experience of a single-centre VR team who together have performed over 1000 vitrectomies between 2005 and 2013 as a starting point for identifying further priorities for research, training and developments in this subspecialty.

Methods

A retrospective, non-comparative consecutive case series was performed involving a thousand consecutive eyes that underwent a vitrectomy surgery from May 2005 to July 2013 at the Eye Foundation Retina Institute, Lagos, Nigeria. The Eye Foundation Retina Institute is a department of the Eye Foundation Hospital, which is an indigenous private training institution that comprises multispecialty eye clinics and surgeries. The Eye Foundation Retina Institute receives VR referrals from other eye centres within the state of Lagos and the country. There are a total of five VR centres in Lagos, serving an approximate population of 20 million. The majority of the patients pay privately for their eye care, while others may be sponsored, use health insurance or request for low-cost services.

The case records of the study patients were identified from the operating room log, and information from the case records was extracted. When a case record was not available or was missing, the next in the series was selected. All non-blacks were excluded. The primary information extracted was indication for surgery, outcomes and complications. In a situation where more than one indication for vitrectomy surgery was noted, the most significant indication was chosen as the primary indication.

Other information extracted included the patient’s gender, laterality, visual acuity of the index eye and the fellow eye, symptom duration before presentation, duration of follow-up and number of revitrectomies. When retinal detachment was the primary indication, additional information including the presence of any proliferative vitreoretinopathy (PVR) and also of advanced PVR that is grade C or D and giant retinal tear (GRT) was noted. PVR grade C in this study was defined as the presence of fixed retinal folds, and grade D the presence of a funnel-shaped total retinal detachment. Also, intraoperative combination with cataract extraction, scleral buckle and use of endotamponade was also noted. Retina laser photocoagulation was the method of retinopexy in all cases. Cryotherapy was not used.

All surgeries were performed using the Accurus vitrectomy system (Alcon), and mostly 20G surgery was performed by the same VR team consisting of one senior consultant and two junior consultants. Over the study period, there were significant upgrades in technique and instrumentation (such as the change to transconjunctival surgery, the use of smaller gauge cutters and more recent wider angle viewing systems). Silicone oil was most commonly used as it was more available than expansile gases during the study period.

Results

Of the 1000 eyes that had vitrectomy, 18 case records were missing, but were replaced with the next in line in the series. The overall age of patients seen ranged from 1 to 85 years (the average age was 47 years, with 709 males and 291 females (ratio 2.4:1)).

Of the one thousand vitrectomy surgeries performed within the study period, the top three indications for surgery were rhegmatogenous retinal detachment (62%), followed by diabetic retinopathy (DR) (10%) and trauma (6%), and this is illustrated in table 1. Vitreous haemorrhage (VH) secondary to sickle cell proliferative retinopathy and branch vein occlusions were the next most common indication for vitrectomy (5% and 3%, respectively). Endophthalmitis, vitritis, haemorrhagic posterior vitreous detachment, macular hole, dislocated lens, trauma and cataract-related vitrectomy were less common indications, each accounting for about 1% of patients. Rare indications, occurring in less than 1% of cases, included macular degeneration, polypoidal choroidal vasculopathy, asteroid hyalosis, dropped nucleus, vitreous opacities and VR fibrosis. Details of indications and number of eyes are shown in table 1.

Figure 1 shows the age distribution curve of the four most common indications for surgery, which are rhegmatogenous retinal detachment, diabetic vitrectomy, trauma and proliferative sickle cell retinopathy.

Request permission

A subanalysis of the two most common indications was done, comparing demographics, presentation, outcomes and complications.

Rhegmatogenous retinal detachment

Demographics and presentation

There were 618 eyes which presented with rhegmatogenous retinal detachment (RRD), the demographic and presentation details of which are captured in table 2.

Table 2

View inline|View popup

Demographics and presentation of patients with retinal detachment

The duration of onset of symptoms to presentation was on average 13.5 months, with 43% of patients presenting within 1 month of symptoms.

One-third of RRD had some degree of PVR, and half of these cases with PVR had advanced stages of PVR. Macular ‘on’ was recorded in only 19 (3%) of 618 eyes, and the remaining 97% were macular ‘off’ at presentation. Macular hole was seen in 52 eyes (8.3%). GRT was seen in 35 eyes (5.7%). Bullous retinal detachment was present in 34 eyes (5.5%), and some amount of VH was associated with the retinal detachment in 67 (10.8%) eyes. Silicone oil was used as endotamponade in 570 eyes (92.2%), and a scleral buckle was combined with vitrectomy in 31 (5%) of eyes. There were 588 (95.1%) eyes with a follow-up of over 2 months, while only 30 eyes had less than 2 months of follow-up. The primary retinal reattachment rate among these 588 eyes after 2 months was 56%. With further surgery, the final reattachment rate at the time of last clinic visit for all patients was 88.3% in 546 eyes.

RRD visual outcomes

A quarter of the eyes (158, 25.5%) had a visual acuity of 6/60 or better postoperatively (as shown in table 3) compared with only 138 (22%) preoperatively. A majority of eyes, however, had counting fingers vision or more and could navigate. Hand motion vision or worse was seen in 272 eyes (44%) postoperatively (table 3).

Table 3

View inline|View popup

Postoperative visual outcomes from vitrectomy

The visual acuity at presentation (preoperative vision of the operated eye) versus the postoperative vision for RRD eyes is shown in a scatter plot in figure 2.

Request permission

Figure 2

Comparison of preoperative versus postoperative visual acuity outcome of rhegmatogenous retinal detachment eyes. LOGMAR, logarithm of the minimum angle of resolution.

RRD complications

As shown in table 4, 313 (50%) eyes with RRD required more than one vitrectomy to attach the retina. With further surgery, the final reattachment rate was 88.3% at the time of last clinic visit. The retina remained detached in 60 (9.5%) eyes, and there were 12 eyes (1.9%) in which the retina status could not be ascertained. Only 41.7% of the eyes had silicone oil removal. Some of the patients did not keep their postoperative scheduled clinic visits, and therefore silicone oil removal was not conducted at our facility and may have been done elsewhere. Of the 258 eyes that had the silicone oil removed, 225 (87.2%) had an attached retina as at the last follow-up. Silicone oil caused complications in 225 eyes (39%), with some patients having more than one complication.

Table 4

View inline|View popup

Complications of vitrectomy for RRD (n=618)

Vitrectomy for diabetic retinopathy (DR)

DR was the second most common indication for vitrectomy, occurring in 101 eyes (10%). The age distribution of patients with diabetes is shown in figure 1 and is similar to patients with RRD, peaking in the 50s–60s age group. In patients with DR, vitrectomy was necessary to treat VH alone in 66 eyes (74%), tractional retinal detachment (TRD) with VH in 21 eyes (19.1%) and TRD only in 14 eyes (6.7%).

Fifty-two per cent of patients with DR had poor vision in the fellow eye, with a visual acuity of less than 6/60 in the fellow eye (53 eyes), compared with patients with RRD in whom only a third had poor vision in the other eye. This may suggest that health-seeking behaviour may depend more on patients’ ability to function when the second eye is affected rather than the symptoms themselves.

PVR was noted in three eyes preoperatively, and silicone oil was used as endotamponade in 62 eyes (61.4%). Repeat vitrectomy was performed in 56.4% (57 eyes).

Visual acuity outcomes are reported in table 3 and subdivided into those who had VH alone and TRD with or without vitreous haemorrhage (±VH) as they represent different stages of the disease. Expectedly, the eyes with VH alone showed the greatest vision improvement, with 39 eyes (59%) improving compared with only 8 eyes (23%) in the TRD±VH group. The majority of eyes however, preserved counting fingers visual acuity which is sufficient for navigation.

The visual acuity at presentation (preoperative vision of the operated eye) versus the postoperative vision in DR eyes is presented as a scatter plot and is shown in figure 3.

Request permission

Figure 3

Comparison of preoperative versus postoperative visual acuity outcome of diabetic retinopathy eyes. LOGMAR, logarithm of the minimum angle of resolution.

Discussion

This study represents the largest consecutive case series of vitrectomies in Africa and offers insights into the clinical presentation and outcomes of this surgery. As a retrospective case series, it lacks controls that would enable it to be compared directly with outcomes in a more developed continent. Equally, the study was performed in a single centre in Nigeria and may not be generalisable to all of Africa, given the variation in ethnicity, disease and surgical practice. Additionally, collecting data from a team of surgeons, of whom two were more junior, may also have introduced interoperation variability and affected surgical outcomes. The study was also conducted over a period of 8 years, which coincidentally was a significant time globally for advancements in the vitrectomy technique, and so surgery was evolving over this time period.

Nevertheless, this paper adds a significant contribution to the literature by making available real-world outcomes, which may serve as a benchmark for others in Africa and drive the need for better outcomes and specialist training of VR surgeons and their teams.

Studies conducted in Africa report a variable incidence of VR diseases, but there are few population-based studies.7–10 In the national Nigerian blindness survey, Kyari et al found that the age-adjusted prevalence of diabetes in Nigeria was 3.25% and that over 10% of people with diabetes aged ≥40 years had sight-threatening DR.11 12 Trauma accounted for 1.1% of blindness, but retinal detachment was not specifically reported. With an ever-increasing population in Nigeria and the increasing prevalence of diabetes mellitus in Africa and worldwide, the need for retinal screening and VR services in Africa is essential.11 13 14

This study demonstrated that retinal detachment, while a low concern on a population basis, remains the most common indication for vitrectomy, with the possibility to prevent blindness. This supports the findings by Yorston et al, in 2002, who also highlighted the burden of retinal detachments and advocated for more resources and training.5 However, this study also highlighted the fact that patients with retinal detachment come on average 13.5 months after the initial onset of symptoms and only 43% of patients came within the first month. This translates to a majority of patients (97%) presenting with macular involvement. Longer duration between symptoms and treatment has long been associated with poorer visual outcomes.7 This study suggested that patients often presented only when their better eye deteriorated, as indicated by the high number of patients who had poor vision in the fellow eye for RRD, which was 28.9%, and for DR was 35%. Interestingly, regardless of indication for vitrectomy, a significantly higher proportion of males accessed VR services than females (80:20), which highlights the inequalities in healthcare access and supports the findings of the Nigerian national blindness survey which found that being female was a risk factor for blindness.15

The delay in presentation allows PVR to develop and is linked to poor prognosis. PVR was seen in a third of RRD eyes in this study and 16.5% of eyes had significant PVR (grades C and D). This is lower than what has been reported in a similar study from South Africa in which PVR was seen in 33% of the eyes.16 The study from east Africa reported a similar PVR rate of 18%.5 However, a study from Ethiopia having a much higher rate of significant PVR at 69% has been reported.6 It is therefore safe to say that PVR rates in Africa vary widely but are in the double digits, contrary to findings in the more developed countries which have single-digit PVR rates.17 18 This creates a significant challenge and complexity to retinal detachment presentation and treatment outcomes. Similarly, GRTs are more difficult to treat, and their frequency in this study (5.7%) and in Yorston’s study (8.3%) suggests that GRTs are more common in Africa than in the UK.19

The advanced presentation of African patients with RRD begs the question what surgical technique is best used in these complex cases. Several studies have tried to answer this question, arguing that using scleral buckle alone might have better anatomical success rates, but the complexity of cases with PVR would support the necessity for vitrectomy with silicone oil.20–22 In this series, vitrectomy rather than scleral buckle featured as the predominant technique, with only 5% of cases using additional buckle. Interestingly, the use of scleral buckle reduced and vitrectomy increased during the study period, similar to global trends as highlighted by the American Society of Retina Specialists (ASRS.org) Preferences and Trend (PAT) survey.

Final anatomical success was 88.3%, similar to the 88.2% reported in Kenya with patients having a similar amount of advanced PVR.5 The visual outcome of the RRD eyes in this study is rather disappointing compared with the East Africa study where 63.7% had a vision of 6/60 or better. However their visual acuity results were reported for patients who achieved anatomical success, whereas our study reports on all retinal detachments. Our poor results could also be explained by the significant delay in patient presentation; almost all patients were macular off at presentation and that nearly 40% of patients had complications from using silicone oil and as much as 41% of patients did not have the oil removed by the final clinic visit. Outcome of surgery using silicone oil has been shown to be poorer than with gas and has the added disadvantage of requiring a further procedure to remove the oil. Expansile gases are the way forward to improve vitrectomy outcomes for African patients, but collaborative efforts are needed to create reliable and affordable supply chains to VR units within Africa.

The visual outcomes in the DR were better than in the rhegmatogenous retinal detachment group, which is unsurprising given the indication in 60% of these vitrectomies was VH alone. However 40% of our patients had vision which was worse postoperatively compared with Guthrie et al,23 whose study showed only 25% of patients with worse vision postoperatively. However a quarter of patients in Guthrie et al’s study were non-domicile and may represent different ethnicity. Mastropasqua et al have shown that black patients have poorer visual outcomes compared with white and Asian patients, and this may explain our outcomes in an exclusively black population.24 A preoperative intravitreal injection of bevacizumab has been demonstrated by Guthrie et al in an African population to reduce intraoperative complications and reduce vision loss after vitrectomy.23 However the additional cost incurred by the patient remains a challenge and has not been adopted as routine in our practice.

This study highlights the challenges faced by VR teams performing vitrectomies in Africa. There is still much to be improved on surgically, including improvements in techniques, equipment and training. However, to improve outcomes from vitrectomy, patients with VR problems need to be identified earlier and expansile gases used as the tamponade. In the case of RRD, patient awareness of symptoms needs to be raised at the community level to trigger early consultation at the primary level, followed by assessment and evaluation of the eye by health professionals, who with fundal mobile photography and digital technology, can communicate and refer patients to tertiary care.25 For DR, a national screening programme of annual fundus examination for all patients with diabetes needs to be implemented to identify and treat proliferative disease early.