A Database Study of Vitreomacular Traction Timothy L Jackson PhD, FRCOphth1, Paul HJ Donachie MSc2, and Robert L Johnston FRCOphth2 for the Vitreomacular Traction Study Group* King’s College London1 and Gloucestershire Hospitals NHS Foundation Trust, Cheltenham2, United Kingdom.
*Collaborators (Vitreomacular Traction Study Group): Fiona Bishop, Ted Burton, Ahmed El-Amir, Paul H.J. Donachie, Alan Fitt, Steven Harsum, Richard Haynes, Julian Hickman-Casey, Timothy L. Jackson (Chair), Robert L. Johnston, Dominic McHugh, Richard Newsom, Rubina Rahman, Martin Snead, Steve Winder, Yit Fung Yang, Rahila Zakir, and Hadi Zambarakji.
Word count: 3673.
Proprietary Interest and Funding: T. Jackson and P. Donachie’s employer received unrestricted funding from Alcon-ThromboGenics to analyse these data. Alcon-ThromboGenics had no data access, nor any input in the study design, data analysis, or manuscript preparation, with no opportunity to review the manuscript prior to first submission.
Database study; vitreomacular traction; vitreomacular adhesion; natural history; pars plana vitrectomy; complications; cataract; United Kingdom; surgery; audit.
A database study of 1,399 eyes with vitreomacular traction found that vitrectomy was associated with intraoperative complications in 11.5% of cases and retinal detachment in 2.7%. Only 36% gained two Snellen lines or more.
Purpose: To audit the natural history of vitreomacular traction (VMT), and the risks and benefits of surgery.
Methods: Database study of 1,254 patients (1,399 eyes) with VMT from 16 United Kingdom vitreoretinal units.
Results: The median age was 75 years, with 36.2% of patients diagnosed as diabetic. In 986 (70.5%) eyes managed conservatively, the median presenting visual acuity (VA) of 10/20 was unchanged over 10.8 months. Of 413 eyes (29.5%) that required pars plana vitrectomy, the median preoperative logarithm of the minimum angle of resolution VA improved from 0.60 to 0.50 (5/20 to 6/20), with 36% gaining at least 0.3 units (approximately 2 Snellen lines). Vitrectomy was combined with internal limiting membrane peel in 43.8%, epiretinal membrane peel in 42.4%, gas/air tamponade in 46.2%, and cataract surgery in 27.9%. One or more intraoperative complication occurred in 11.5%; most commonly retinal breaks (8.0%), retinal trauma (1.9%), and retinal hemorrhage (1.0%). Postoperative retinal detachment occurred in 2.7% and macular hole in 2.2%. The 1, 2 and 3 year rates of post-vitrectomy cataract surgery were 28.2%, 38.2% and 42.7% respectively.
Conclusion: Many eyes with VMT have stable VA. Those progressing to vitrectomy have relatively modest VA gains, and complications are not infrequent.
Vitreomacular traction (VMT) syndrome occurs in the context of an incomplete posterior vitreous detachment (PVD), where persisting macular adhesion leads to distortion of the macular anatomy. It can be associated with variable loss of visual function manifest as distortion, metamorphopsia, and reduced visual acuity (VA).
Traditionally there are two main management options, pars plana vitrectomy (PPV) or observation, the latter in the expectation that some eyes will resolve spontaneously.1-9 A recent meta-analysis of PPV for VMT revealed that only one third of eyes gained two or more Snellen lines.10 Equally, observation may have disadvantages, with a natural history study reporting that only 11% of eyes showed spontaneous resolution over a mean follow up of 5 years, whereas 64% of eyes lost at least two Snellen lines over this timeframe.11 Spontaneous resolution may be more frequent in eyes without coexisting epiretinal membranes and with small areas of adhesion, but conversely VMT can progress to macular hole.12, 13
More recently, ocriplasmin has been licensed for the treatment of symptomatic vitreomacular adhesion. Ocriplasmin is given as an intravitreal injection, and is designed to dissolve the vitreous and loosen its adhesion to the internal limiting membrane (ILM).14-17 A recent, large, double-masked, randomized controlled trial reported that vitreomacular adhesion was fully resolved in 26.5% of cases following a single intravitreal injection of 125 µg of ocriplasmin.16 Other authors have reported pilot data using an intravitreal gas injection to mechanically release VMT, with 6 of 15 cases (40%) responding.18
A systematic review identified several important gaps in the literature on VMT.10 In particular, few peer reviewed reports detailed the natural history of VMT, making it hard for clinicians and patients to determine the risks and benefits of the treatment options. Also, there are no large randomised controlled trials of PPV for VMT, and most of the literature arises from uncontrolled, single centre studies.
This database study aimed to provide further insight into the routine management and clinical course of VMT syndrome, and the risks and benefits of surgery.
MATERIALS and METHODS
Data were extracted on eyes diagnosed with VMT between June 2006 and December 2012 from 16 participating vitreoretinal units. All data were captured using one electronic medical record (EMR) system (Medisoft Ophthalmology, Medisoft Limited, Leeds, UK), and then pooled for analysis. The lead clinician and Caldicott Guardian (who oversees data protection) at each hospital gave written approval for the data extraction. Anonymized database analyses of this type do not require ethical permission as they are viewed as audit or service evaluation.19 This study was conducted in accordance with the declaration of Helsinki, and the UK’s Data Protection Act. It was routine practice at all sites for nurses to record a pre-operative assessment on the EMR, during which a standardised set of questions were asked in relation to systemic comorbidities and the use of topical and systemic medications.
The EMR required surgeons to record whether or not there were any surgical complications, before they could save the operation note. If the surgeon indicated that a complication occurred, then he or she had to select from a pre-populated list of well-recognised complications specific to that operation, or select ‘other’ and record the complication using free text.
All hospitals using the EMR for vitreoretinal procedures also recorded cataract surgery using the same system. It was therefore possible to estimate the incidence of post-vitrectomy cataract surgery (PVCS).
Eligibility and Statistical Analysis
All eyes with a diagnosis of primary VMT were eligible. Eyes were excluded if they had undergone PPV prior to VMT diagnosis, or if they had non-VMT eye surgery after diagnosis and no record of VMT treatment.
The time to PVCS was modelled using the Kaplan-Meier method,20 with PVCS modelled as failure. Eyes were censored at the last date on which follow up data of any type was recorded on the EMR, if they had not had cataract surgery.
Visual acuity data were expressed as the logarithm of the minimum angle of resolution (LogMAR), where the baseline measurements were taken at diagnosis for the non-surgical eyes and day of surgery for the surgical eyes. We chose to primarily report the median VA as the most appropriate VA outcome, but to allow comparison with similar studies21-23 we also reported the mean VA by substituting CF, HM and PL with 2.1, 2.4 and 2.7 respectively, where each step represents a doubling of the visual angle.24 For each substitution we provided the number of cases with CF, HM and PL, so that other researchers could calculate mean VA using different values.
Visual success was defined as eyes improving by ≥0.30 LogMAR units (approximately 2 Snellen lines) from baseline and is reported at the last recorded VA measurement (final review). All VA values less than 6 weeks after diagnosis or primary VMT surgery (for the non-surgical and surgical groups respectively) were excluded and VA was categorised as ≤0.30, >0.30 - 0.60, >0.60 - 0.90, >0.90 - 1.20, >1.20 at the time of surgery and the last VA record. All analysis was conducted using STATA version 11 (StataCorp, College Station, TX) except for the confidence interval (CI) which were calculate using Confidence Interval Analysis version 18.104.22.168
Vitreomacular traction syndrome was diagnosed in 1,451 eyes from 1,301 patients over the study period. Of these, 52 eyes from 47 patients were excluded from all analyses: 17 as they had previously undergone PPV surgery at a median of 5 months prior to VMT diagnosis, and 35 as they had non-VMT surgery after diagnosis and no record of VMT treatment. Therefore 1,399 eyes (741 left eyes and 658 right eyes) from 1,254 patients were eligible for analysis, of which 145 patients had bilateral VMT.
Of the 1,254 patients, 507 (40.4%) were male, 746 (59.5%) were female and the gender was not specified for 1 patient. The patient’s ethnicity was recorded for 919 (73.3%) patients, where 830 were Caucasian, 52 Asian, 28 black, 6 mixed race and 3 other races. At the time of primary VMT diagnosis the median age was 75 years (range; 11 – 98 years) and 624 (49.8%) patients were recorded as taking prescribed medications; 134 eye drops only, 274 systemic medications only, and 216 both eye drops and systemic medication.
Previous diseases and systemic comorbidities
Prior to primary VMT diagnosis 78 (6.2%) patients had experienced a myocardial infarction, 64 (5.1%) a cerebrovascular accident and 59 (4.7%) a transient ischemic attack. At the time of primary VMT diagnosis 611 (48.7%) patients were recorded as having no systemic comorbidities, 350 (27.9%) one comorbidity, 155 (12.4%) two, 90 (7.2%) three, 37 (3.0%) four and 11 (0.9%) five comorbidities.
At the time of primary diagnosis 454 (36.2%) patients were record as having diabetes mellitus, 296 (23.6%) hypertension, 66 (5.3%) angina, 58 (4.6%) suffering from anxiety, panic attacks or claustrophobia, 51 (4.1%) hyperlipidemia, 46 (3.7%) asthma, 27 (2.2%) bronchitis, 27 (2.2%) hypercholesterolemia, 20 (1.6%) bleeding diathesis, 14 (1.1%) anemia, 7 (0.6%) cancer, 4 (0.3%) tuberculosis and 63 (5.0%) other non-specified conditions.
Of the 1,399 eyes diagnosed with VMT, 986 (70.5%) eyes did not undergo surgery to treat the disease (512 left eyes and 474 right eyes). The 986 non-surgical eyes were from 862 patients (median age 75 years; range 30 – 96 years), 137 of whom had bilateral VMT. Data were recorded after diagnosis for 587 (59.5%) of the non-surgical eyes for a median follow up time of 10.0 months (range: 1 day – 6.0 years). Of these, 54 had no VA measurement at the time of diagnosis, 20 had no VA measurements recorded after diagnosis and 36 had less than 6 weeks follow up data recorded, leaving 477 non-surgical eyes eligible for VA analysis. For these eyes the median VA at diagnosis was 0.30 LogMAR units (Snellen 20/40) (mean 0.44; range: -0.20 – NPL) including 10 eyes with CF, 3 with HM, 1 with PL and 1 with NPL. At the time of final review after a median follow up of 10.8 months (range: 22 days – 5.4 years) the median VA remained 0.30 LogMAR units (mean 0.47; range: -0.20 – HM) including 15 eyes with CF and 6 eyes with HM; 29 eyes (6.1%; 95% CI: 4.3% TO 8.6%) achieved visual success (a gain of ≥0.30 LogMAR units, ≈2 Snellen lines).
There were no differences observed between patients progressing to surgery or not for either gender or the age at diagnosis, but there were differences observed for diabetic status and ethnicity, where a lower proportion of patients with diabetes had surgery than did not (28.6% vs. 41.1%) and a lower proportion of patients from ethnic minorities had surgery than did not (4.7% vs 8.3%). Both the diabetic status and ethnicity were ‘not specified’ for many patients, although the observed differences hold when these estimates are restricted to singleton eye patients with the respective data.
Surgical technique, anaesthesia and setting
Surgery was conducted on 413 of 1,399 (29.5%) eyes, of which 381 (92.2%) were undertaken as day cases and 32 (7.8%) as inpatients. The 413 primary VMT operations were performed on 229 left eyes and 184 right eyes from 405 patients (median age 75 years; range 11 – 98 years) by 72 surgeons. Consultant surgeons performed 272 (65.9%) of the operations, independent non-consultant surgeons, including Fellows, 131 (31.7%) operations, and more junior trainee surgeons 10 (2.4%) operations.
All operations included PPV, which was combined with an ILM peel in 181 (43.8%), epiretinal membrane (ERM) peel in 175 (42.4%), retinopexy in 98 (23.7%) and cataract surgery in 115 (27.9%) operations. A tamponade was used in 194 (47.0%) operations: 3 (0.7%) used silicone oil; 64 (15.5%) used sulfahexafluoride (SF6); 30 (7.3%) used hexafluoroethane (C2F6); 14 (3.4%) used perfluoropropane (C3F8); and 83 (20.1%) used air. For 78 (18.9%) operations the type of anaesthesia used was not recorded on the EMR. For the remaining 335 operations general anaesthesia was used for 43 operations (12.8%) and local anesthesia for 292 (87.2%), 7 with sedation.
Of 413 operations, 361 (87.4%) were recorded as having no intraoperative complication. Of the 52 (12.6%; 95% CI: 9.7% to 16.1%) eyes recorded as having a complication, the most common was an iatrogenic tear, which occurred in 33 operations (Table 1). Two intraoperative complications occurred in 4 eyes. If the complications associated with cataract surgery were excluded, then the complication rate attributable to VMT surgery alone was 11.5% (47/408; 95% CI: 8.8% to 15.0%).
Of the 413 eyes undergoing primary VMT surgery, 235 were excluded from the PVCS analysis: 45 because they were known to have had previous cataract surgery; 115 as they had combined VMT and cataract surgery; and 75 as they had less than 3 weeks follow up. Of the remaining 178 eyes eligible for analysis, the median follow up was 8.2 months (range 22 days – 6.3 years), and 39 (21.9%) were subsequently recorded as having cataract surgery. The 1, 2 and 3 year PVCS rates were 28.2%, 38.2% and 42.7% respectively (Figure 1).
Other subsequent operations, not related to cataract, occurred in 35 eyes, 8 of which had more than 1 further surgery. A repeat VMT operation was performed in 7 (1.7%) eyes at a median of 3 months (range 0.6 - 7.4 months) after primary VMT surgery; 11 (2.7%) eyes had retinal detachment surgery at a median of 1.9 months (range 1 day - 12.6 months); 9 (2.2%) had macular hole surgery at a median of 5.9 months (range 2.2 - 10.0 months); and 11 (2.7%) had other types of vitreoretinal surgery at a median of 2.8 months (range 11 days – 15.7 months).
Visual acuity change following surgery
Of the 413 eyes undergoing VMT surgery, 182 were excluded from the VA analysis as they had no VA measurement at the time of surgery, 52 as they had no further VA measurement and 19 as there was less than 6 weeks follow up VA data. From the 160 surgical eyes eligible for VA analysis the median preoperative LogMAR VA was 0.60 (Snellen 20/80) (mean 0.69; range -0.10 – HM) which included 3 eyes with CF and 1 with HM.
The last recorded VA measurement ranged from 6 weeks to 6.3 years (median 10.2 months) after primary VMT surgery. At this time the median VA was 0.50 (Snellen 20/63), the mean VA was 0.62 (Snellen 20/83), 58 eyes (36.3%; 95% CI: 29.2% to 43.9%) had achieved visual success, and the worst visual outcomes were CF in 5 eyes and HMs in 2. Twenty-two eyes (13.8%) lost >0.30 LogMAR units at final review. Of these, 7 had repeat surgery within this timeframe (1 for VMT, 1 for macular hole and 5 for retinal detachment) at a median of 2.0 months (range 0.6 – 12.6 months). The worst visual outcomes were CF in 5 and HMs in 2 eyes. The general trend for greatest VA improvement was for the eyes with preoperative VA in the middle of the LogMAR scale (Figure 2).
Only 79 surgical eyes had VA data covering the period from diagnosis to surgery (the others first appeared on the EMR at or near the time of surgery). These cases lost a median of 0.04 logMAR units (mean 0.11, range -0.7 -to +0.3) over a median of 2.8 months prior to surgery (range 5 days – 2.5 years).
Of the 160 surgical eyes eligible for VA analysis 29 were known to be psuedophakic. These eyes had a median VA at surgery of 0.60 LogMAR (Snellen 20/80), a mean VA of 0.68 LogMAR (Snellen 20/96) and the range was 0.20 – CF which included 1 eye with CF. At final review the median VA was 0.50 LogMAR (Snellen 20/63), the mean VA was 0.65 LogMAR (Snellen 20/89) and the range was 0.00 – HM which included 2 eyes with HM. Visual success was achieved by 8 (27.6%) eyes and 3 (10.3%) eyes lost >0.30 LogMAR units.
Diabetic subgroup analysis
Because 36.2% of patients had diabetes we undertook a subgroup analysis of this group. The 454 patients with diabetes contributed 500 eyes to the analysis, and of these 115 eyes progressed to surgery and 385 did not.
Of 385 non-surgical eyes from 342 patients with diabetes (188 males, 156 females and median age at diagnosis 73 years), 179 had VA measurements at diagnosis and more than 6 weeks follow up. Of these the median VA at diagnosis was 0.40 LogMAR (Snellen 8/20) (mean 0.56; range: -0.10 – NPL) including 8 eyes with CF, 2 with HM and 1 with NPL. After a median follow up of 10.5 months (range: 6 weeks – 5.4 years) the median VA was 0.46 LogMAR (Snellen 7/20) (mean 1.20) which included 8 eyes with CF and 4 eyes with HM. Visual success was achieved by 11 eyes (6.2%; 95% CI: 3.5% to 10.7%).
Of 115 surgical eyes from 115 patients with diabetes (58 males, 57 females and median age at diagnosis 73 years), a PPV was combined with ILM peel in 43 (37.4%), epiretinal membrane peel in 45 (39.1%), retinopexy in 22 (19.1%), cataract surgery in 28 (24.4%) and a tamponading agent in 47 (40.9%) eyes.
Fourteen eyes experienced an intraoperative complication (10 iatrogenic tears and 4 iatrogenic retinal traumas’). Repeat VMT surgery was performed in 2 (1.7%) eyes, a post VMT surgery macular hole operation was performed in 4 (3.5%), retinal detachment in 1 (0.9%), other types of vitreoretinal surgery in 6 (5.2%) and cataract surgery in 15 (13.0%) eyes.
Fifty-five of the 155 surgical eyes from patients with diabetes had preoperative and postoperative VA measurements as well as more than 6 weeks follow up. Of these the median VA at surgery was 0.70 LogMAR (Snellen 4/20) (mean 0.76; range: 0.20 – CF) including 2 eyes with CF. After a median follow up of 11.8 months (range: 6 weeks – 6.3 years) the median VA was 0.60 LogMAR (Snellen 5/20) (mean 0.70) which included 1 eye with CF and 2 eyes with HM. Visual success was achieved by 16 eyes (29.1%; 95% CI: 18.8% to 42.1%).
DISCUSSION This database study aimed to collect pragmatic clinical data on the management of VMT. It found that many patients are managed conservatively, with relatively stable VA over the short term. Those who undergo surgery obtain modest VA improvements, with 12.6% having an intraoperative complication, most commonly retinal breaks. The most frequent postoperative complications were cataract in 42.7% of eyes, retinal detachment in 2.7%, and macular hole in 2.2%.
A systematic review and meta-analysis of VMT surgery found that small, petechial, self-resolving retinal hemorrhages were the most common intraoperative complication (5.6% of eyes), followed by retinal breaks (1.6%).10 The present study also found retinal breaks and retinal hemorrhage or trauma to be the most common intraoperative complications, although the rates were higher and the order reversed (8.0% and 2.9% respectively). The studies contributing to the systematic review may be more subject to publication bias than our study, as surgeons may be reluctant to publish poor outcomes and journals may be more likely to publish novel interventions that improve clinical outcomes.26 By contrast the present database study selected routine cases pooled anonymously from a range of hospitals. In addition, the EMR mandated data collection at the time of surgery, such that the surgeon had to confirm whether or not a complication occurred. This, combined with anonymised data collection, should encourage accurate reporting of surgical complications and may explain the higher intraoperative complication rates in this analysis, although it may not fully overcome an innate tendency to underreport complications and the true complication rate could therefore be higher.
Whereas the surgical procedures and complication rates in this study are likely to be reliable, the VA data were less complete and need to be interpreted with caution. Nonetheless, the data suggest that VA was relatively stable in those patients managed conservatively, at least over the median 10.8 months follow up period. There are very few natural history studies, but Hikichi et al found that VA tended to decline over the longer term, with almost two thirds of patients losing at least 2 Snellen lines over 5 years.11 It is possible our patients would show a similar decline if follow up was longer. Also, patients whose VA did decline may have been more likely to undergo surgery, and selectively exit our observation cohort. The limited VA data support this assumption, as eyes that underwent surgery lost 0.04 logMAR units over a median of 2.8 months prior to surgery, which extrapolates to a change from 20/54 to 20/80 in the year prior to surgery. Therefore, our non-surgical group may not fully reflect the natural history of VMT.
In those patients who did undergo surgery, median VA improved only modestly from 0.6 to 0.5 logMAR (20/80 to 20/63) over median follow up of 10.2 months, with only 36% gaining at least 0.3 LogMAR units (≈2 Snellen lines). This is similar to a value of 33% reported in a systematic review.10 Whilst VA is an accepted measure of visual function, patients with VMT may benefit if surgery reduces their metamorphopsia, and VA does not directly measure any such gain.
There was a greater than expected number of patients receiving intravitreal gas injection. Gas may be selected for a variety of reasons including intraoperative retinal breaks, pre-existing macular hole or concern that peeling of the ILM or ERM might have caused a macular hole, to help seal sutureless sclerotomies, or perhaps to help tamponade and flatten elevated macular tissue. It is likely that retinal breaks were responsible for many cases as they were relatively common, whereas we expect pre-existing macular holes to be a less frequent explanation, as these cases are more likely to be coded as a macular hole than VMT, and air was the most commonly used tamponade (air is seldom selected for the treatment of macular holes).21
The rate of post-vitrectomy cataract surgery is less than might have been expected. For example, a UK database study reported that 92% of phakic eyes required cataract surgery within three years of vitrectomy undertaken for macular hole,21 with an equivalent figure of 74% for RD.22 Whilst both macular hole and retinal detachment require intravitreal tamponade, which is thought to increase lens opacity,27 a database study of vitrectomy undertaken for a range of conditions also reported a three year cataract surgery rate of 74%,23 suggesting that gas tamponade cannot fully explain the difference between the present analysis and others. Also, many patients undergoing VMT surgery also received gas tamponade, albeit often with short-acting gases. Although speculative, one explanation for the low cataract rate in our study might be the relatively simple nature of VMT surgery. Another is that those patients expected to require cataract surgery following vitrectomy, due to preoperative subclinical lens opacity, were treated with combined phacovitrectomy, which occurred in more than a quarter of eyes. In addition, our figures may include some pseudophakic patients in the denominator, as pseudophakic patients who had their cataract surgery recorded before the EMR was implemented may not have been recognised as pseudophakic in the present analysis; hence the incidence of cataract in phakic eyes may be somewhat higher than we report. We also had a relatively high proportion of diabetic patients, and diabetes is known to reduce post-vitrectomy cataract.28
Strengths of this study include pooled anonymized collection of pragmatic clinical data from a range of hospitals, mandated intraoperative recording of complications, and its large size relative to all prior publications on VMT. It is also the first database study of VMT. Notwithstanding the advantages of randomised controlled trials for testing new treatments, a database study may better reflect routine care, delivered without the rigid eligibility criteria and mandated follow up of a clinical trial. As such it may be more representative of usual care. The main weakness of this study relates to incomplete data collection on VA change over time, and hence VA outcomes should be considered exploratory. However, database studies seldom have full levels of data completeness and their interpretation necessarily involves this consideration.
Future database studies might consider requesting sites to provide mandated outpatient VA data collection, but until that occurs this study may help clinicians to benchmark their surgical complication rate, and help patients consider the risks and benefits of surgery.
Collaborators: Vitreomacular Traction Study Group: Fiona Bishop, Ted Burton, Ahmed El-Amir, Paul H.J. Donachie, Alan Fitt, Steven Harsum, Richard Haynes, Julian Hickman-Casey, Timothy L. Jackson (Chair), Robert L. Johnston, Dominic McHugh, Richard Newsom, Rubina Rahman, Martin Snead, Steve Winder, Yit Fung Yang, Rahila Zakir, and Hadi Zambarakji.
The authors thank the following sites for contributing clinical data: Leeds Teaching Hospitals NHS Trust; Norfolk and Norwich University Hospitals NHS Foundation Trust; Royal Berkshire NHS Foundation Trust; Peterborough and Stamford Hospitals NHS Foundation Trust; Epsom and St Helier University Hospitals NHS Trust; University Hospitals Bristol NHS Foundation Trust; East Sussex Hospitals NHS Trust; King’s College Hospital NHS Foundation Trust; Gloucestershire Hospitals NHS Foundation Trust; University Hospital Southampton NHS Foundation Trust; Calderdale and Huddersfield NHS Foundation Trust; Cambridge University Hospitals NHS Foundation Trust; Sheffield Teaching Hospitals NHS Foundation Trust; Portsmouth Hospitals NHS Trust; Imperial College Healthcare NHS Trust; Whipps Cross University Hospital NHS Trust.
1. Levy J, Klemperer I, Belfair N, et al. Rapid spontaneous resolution of vitreomacular traction syndrome documented by optical coherence tomography. Int Ophthalmol 2004;25:247-51.
2. Cheng CL, Hoh ST, Chuan JC, et al. Acute vitreomacular traction with early spontaneous resolution. Clin Experiment Ophthalmol 2003;31:161-3.
3. Sulkes DJ, Ip MS, Baumal CR, et al. Spontaneous resolution of vitreomacular traction documented by optical coherence tomography. Arch Ophthalmol 2000;118:286-7.
4. Selver OB, Parlak M, Soylemezoglu ZO, Saatci AO. Spontaneous resolution of vitreomacular traction: a case series. Clin Exp Optom 2013;96:424-7.
5. Hung KH, Yang CS, Lin TC, et al. Optical coherence tomography in spontaneous resolution of vitreomacular traction syndrome. J Chin Med Assoc 2010;73:334-7.
6. Sarup V, Wald KJ, Gross NE, et al. Spontaneous resolution of unilateral vitreomacular traction. Ann Ophthalmol (Skokie) 2006;38:347-9.
7. Rodriguez A, Infante R, Rodriguez FJ, Valencia M. Spontaneous separation in idiopathic vitreomacular traction syndrome associated with contralateral full-thickness macular hole. Eur J Ophthalmol 2006;16:733-40.
8. Carpineto P, Ciancaglini M, Aharrh-Gnama A, et al. Optical coherence tomography and retinal thickness analyzer features of spontaneous resolution of vitreomacular traction syndrome: a case report. Eur J Ophthalmol 2004;14:67-70.
9. Kusaka S, Saito Y, Okada AA, et al. Optical coherence tomography in spontaneously resolving vitreomacular traction syndrome. Ophthalmologica 2001;215:139-41.
10. Jackson TL, Nicod E, Grimaccia F, et al. Pars plana vitrectomy for vitreomacular traction syndrome: a systematic review of safety and efficacy. epub ahead of print.
11. Hikichi T, Yoshida A, Trempe CL. Course of vitreomacular traction syndrome. Am J Ophthalmol 1995;119:55-61.
12. Odrobina D, Michalewska Z, Michalewski J, et al. Long-term evaluation of vitreomacular traction disorder in spectral-domain optical coherence tomography. Retina 2011;31:324-31.
13. Lee MW, Cheng Bobby CL, Ang CL. Optical coherence tomography evidence of rapid progression from vitreomacular traction to full thickness macular hole. Clin Experiment Ophthalmol 2007;35:379-80.
14. Gandorfer A, Rohleder M, Sethi C, et al. Posterior vitreous detachment induced by microplasmin. Invest Ophthalmol Vis Sci 2004;45:641-7.
15. Gandorfer A. Enzymatic vitreous disruption. Eye (Lond) 2008;22:1273-7.
16. Stalmans P, Benz MS, Gandorfer A, et al. Enzymatic vitreolysis with ocriplasmin for vitreomacular traction and macular holes. N Engl J Med 2012;367:606-15.
17. Gandorfer A. Pharmacologic vitreolysis: rationale, potential indications, and promising agents. Retina 2012;32 Suppl 2:S221-4.
18. Rodrigues IA, Stangos AN, McHugh DA, Jackson TL. Intravitreal injection of expansile perfluoropropane (c(3)f(8)) for the treatment of vitreomacular traction. Am J Ophthalmol 2013;155:270-6 e2.
19. National Health Service Health Research Authority. Defining research. 2010. Available at http://www.nres.nhs.uk/applications/is-your-project-research/ (accessed 5th October 2013)
20. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Statist Assoc 1958;53:457-81.
21. Jackson TL, Donachie PH, Sparrow JM, Johnston RL. United kingdom national ophthalmology database study of vitreoretinal surgery: report 2, macular hole. Ophthalmology 2013;120:629-34.
22. Jackson TL, Donachie PH, Sallam A, Johnston RL. United Kingdom national ophthalmology database study of vitreoretinal surgery: report 3, retinal detachment. Ophthalmology. epub ahead of print.
23. Jackson TL, Donachie PH, Sparrow JM, Johnston RL. United Kingdom National Ophthalmology Database Study of Vitreoretinal Surgery: Report 1; Case mix, complications, and cataract. Eye (Lond) 2013.
24. Sparrow JM, Taylor H, Qureshi K, et al. The Cataract National Dataset electronic multi-centre audit of 55,567 operations: risk indicators for monocular visual acuity outcomes. Eye (Lond) 2012;26:821-6.
25. Bryant T. Computer software for calculating confidence intervals. In: Altman DG, Machin D, Bryant TN, Gardner MJ, eds. Statistics with confidence, 2 ed. London: BMJ Books, 2000.
26. Newcombe RG. Towards a reduction in publication bias. Br Med J (Clin Res Ed) 1987;295(6599):656-9.
27. Thompson JT. The role of patient age and intraocular gases in cataract progression following vitrectomy for macular holes and epiretinal membranes. Trans Am Ophthalmol Soc 2003;101:485-98.
28. Smiddy WE, Feuer W. Incidence of cataract extraction after diabetic vitrectomy. Retina 2004;24:574-81.
Figure 1: Figure shows a Kaplan-Meier curve of the rate of post-vitrectomy cataract surgery (PVCS), over time.
Figure 2: Figure shows the change in logarithm of the minimum angle of resolution (LogMAR) visual acuity from baseline to final follow up, for those eyes undergoing vitrectomy for vitreomacular traction (VMT). The size of each circle is proportional to the number of patients in that category, with those above the dashed line of identity demonstrating an improvement by at least 0.3 LogMAR units (approximately 2 Snellen lines). Of 160 cases, 71 improved by ≥ 0.3 units, 58 remained in the same category, and 31 deteriorated by ≥0.3 units.