The Royal College of Ophthalmologists' National Ophthalmology Database study of cataract surgery: report 1, visual outcomes and complications

Aims To describe the outcomes of cataract surgery in the United Kingdom. Methods Anonymised data on 180 114 eyes from 127 685 patients undergoing cataract surgery between August 2006 and November 2010 were collected prospectively from 28 sites. Outcome measures included intraoperative and postoperative complication rates, and preoperative and postoperative visual acuities. Results Median age at first eye surgery was 77.1 years, 36.9% cases had ocular co-pathology and 41.0% patients underwent cataract surgery on both eyes. Preoperative visual acuity was 0.30 logMAR or better in 32.0% first eyes and 47.7% second eyes. Postoperative best-measured visual acuity was 0.00 and 0.30 logMAR or better in 50.8 and 94.6% eyes without ocular co-pathology, and 32.5 and 79.9% in eyes with co-pathology. For eyes without co-pathology, postoperative uncorrected distance visual acuity was 0.00 and 0.30 logMAR or better in 27.3 and 80.9% eyes. Posterior capsule rupture or vitreous loss or both occurred in 1.95% cases, and was associated with a 42 times higher risk of retinal detachment surgery within 3 months and an eight times higher risk of endophthalmitis. Conclusion These results provide updated data for the benchmarking of cataract surgery. Visual outcomes, and the rate of posterior capsule rupture or vitreous loss or both appear stable over the past decade.


Introduction
Cataract is the leading cause of blindness in the world 1 and cataract surgery is the most commonly performed operation by the UK National Health Service (NHS) with around 330 000 cataract operations performed per year in England. Robust data are critical for the benchmarking of surgical outcomes, as a quality improvement tool, and for commissioners and patients to make informed choices.
The Royal College of Ophthalmologists' National Ophthalmology Database (RCOphth NOD) was established to provide national audit and research data, and to provide an evidence base for revalidation standards allowing Ophthalmologists to compare their surgical outcomes with those of their anonymised peers. The RCOphth NOD is the formalised successor to the ad hoc collaboration that resulted in the Cataract National Dataset publications. [2][3][4][5][6] The RCOphth NOD covers a range of conditions and operations 7-10 and collates pseudoanonymized data collected during routine clinical care using electronic medical record systems (EMRs).

Materials and methods
Data were extracted from 31 UK NHS Trusts of which 28 had recorded data for cataract surgery, as detailed in the acknowledgements section.
All cataract operations were performed between August 2006 and November 2010 and this time frame was specifically chosen to exclude data used in previous analyses of cataract surgery outcomes in the Cataract National Dataset publications. [2][3][4][5][6] All data were recorded using a single EMR system (Medisoft Ophthalmology, Medisoft Limited, Leeds, UK). The lead clinician and Caldicott Guardian (responsible nominee for data protection) at each NHS Trust gave written approval for anonymised data extraction. Anonymized database analyses of this type do not require ethical permission as they are viewed as audit or service evaluation (see http://www.hra.nhs.uk/research-community/beforeyou-apply/determine-whether-your-study-is-research/) and this was confirmed by a Research Ethics Committee. This study was conducted in accordance with the declaration of Helsinki, and the UK's Data Protection Act.

Case definitions
Eligible cataract operations were those performed on patients aged 18 years or older using phacoemulsification and where the primary intention was cataract surgery, not combined 'cataract þ other' operations where the cataract component may not have been the primary reason for surgery. Surgeon grades were categorised as: Consultant surgeons, independent non-consultant surgeons (staff grades, associate specialists and trust doctors) and trainee surgeons (Foundation Year doctors, Specialist Trainees, Senior House Officers, Specialist Registrars and Fellows.) Owing to progression through training an individual surgeon can have data recorded at more than one grade. Ocular co-pathology was a compulsory question with responses of none, a list of common categories and a free text field for rare options.

Intraoperative complications
In all centres the EMR software mandated the collection of the presence or absence of surgical complications. If the surgeon indicated that a complication occurred, then they had to select from a pre-populated list of complications specific to that operation, or select 'other' and record the complication using free text. 'Posterior Capsular Rupture or vitreous loss or both' (abbreviated as PCR) was defined as unintentional communication with the posterior segment from the occurrence of any of the following intraoperative complications: PCR with or without vitreous loss, zonule rupture with vitreous loss, vitreous loss, vitreous to the section at the end of surgery, IOL into the vitreous or nuclear/epinuclear fragment into the vitreous or the performing of any of an automated anterior vitrectomy, 'sponge and scissors vitrectomy,' secondary IOL or scleral fixed IOL during surgery; or if either of vitreous to the section or anterior chamber were recorded within 2 months of cataract surgery, or an operation for a dropped nucleus was recorded within 3 months of cataract surgery.

Postoperative complications: retinal detachment and endophthalmitis
Owing to varying availability of postoperative complication data, the results for post cataract retinal detachment surgery and endophthalmitis treatment were confined to centres where this could be cross-checked with other RCOphth NOD treatment data. Data on retinal detachment surgery were available from 15 centres providing 139 537 (77.5%) of all cataract operations and results are reported for within 3 months, 6 months and 1 year of cataract surgery. For endophthalmitis data from 19 centres providing 145 868 (81.0%) cataract operations were available and results are reported within 3 months of cataract surgery.

Visual acuity
Preoperative visual acuity data was defined as the better value of uncorrected distance visual acuity (UDVA) or corrected distance visual acuity (CDVA). Eligible VA measurements were recorded within 3 months prior to cataract surgery where the VA measurement closest to the date of cataract surgery was used in the analysis.
Postoperative 'best-measured' visual acuity was defined as the best CDVA measurement within 2 weeks and 4 months of cataract surgery when present and the best measurement of UDVA or pinhole VA within the time period when CDVA was absent.
For postoperative UDVA only analysis, data was used for eyes where the predicted postoperative refraction was ±0.5 dioptres spherical equivalent.
Visual acuity values were recorded as Snellen or logMAR, and Snellen values were converted to logMAR for analyses. LogMAR values corresponding to count fingers (CF), hand movements (HM), perception of light (PL) and no PL (NPL) were substituted with 2.10, 2.40, 2.70 and 3.00 logMAR, respectively, in keeping with previous publications from this group.
Significant visual loss was defined as visual acuity deteriorating by Z0.30 logMAR units (doubling or worse of the visual angle) 4  In 67.3% cases the operation was on the patient's first eye to undergo cataract surgery. For those who had surgery on both eyes (41.0% of all patients), the median time between operations was 3.7 months (IQR: 2.4-6.6, range: 0-114.6 months). 436 (0.3%) patients underwent immediate sequential bilateral surgery performed by 120 surgeons.

Grade of surgeon
Of the 995 surgeons, 285 were Consultant surgeons, 451 independent non-consultant surgeons and 325 trainee surgeons. Sixty-six surgeons performed surgery while at more than one grade of surgeon.

Surgical details
Local anaesthesia was used for 96.2% operations, general anaesthesia for 3.5% operations and not known for 0.3% operations. Pupil size at the time of surgery was considered to be 'small,' 'medium,' or 'large' in 3.0, 10.7, and 86.2% eyes, respectively, and was not recorded for 0.1% of eyes. The intraocular lens was placed in the capsular bag in 98.6% eyes, in the sulcus or partly in the bag for 1.0% eyes, and in the anterior chamber for 0.1% eyes. In 0.2% eyes, no intraocular lens was placed (aphakic) and in 0.2% of the intraocular lens position was not recorded.

Intraoperative complications
Overall 4.2% (95% CI: 4.1-4.3%) cases had an intraoperative complication, the most common being PCR (1.95%, 95% CI: 1.89-2.02%; Table 2). The rate of PCR was 1.63% in eyes without co-pathology (1847/113 610) and 2.51% (1667/66 504) in those eyes with a co-pathology. The presence of at least one ocular co-pathology was associated with any intraoperative Intraoperative complication rates were lower for consultant surgeons than other surgeon grades (P ¼ 0.000), see Table 2. Over the study period, the rate of PCR by surgical grade and the overall rate appears stable, see Figure 1a.

Discussion
This study provides pragmatic data on current cataract surgery outcomes using data collated from a large  2 Over the past 20 years there has been a dramatic change in practice to operating on eyes with cataract at lower levels of visual impairment. The reasons for this are multifactorial and include better surgical techniques and outcomes, patients' visual expectations and a need for higher visual function to maintain independence, including that for driving. 12 In 1990, 8.3% cases in England had a preoperative visual acuity of 0.30 logMAR or better, 13 in 1997 this had increased to 31%, 14 and in Cataract National Dataset analysis for 2001-2006 was 43%. 2 In this series, 36% had a preoperative visual acuity of 0.30 logMAR or better. We found 32% first eyes and 48% second eyes had a preoperative visual acuity of 0.30 logMAR or better compared with 35 and 55%, respectively in the Cataract National Dataset papers. 2 Thus suggesting there is no longer a trend towards performing cataract surgery at lower levels of visual impairment in the UK, and if anything this may have reversed slightly. This may be owing to cataract surgery rationing in recent years. 15 The analyses of postoperative visual acuity showed 44 and 89% all eyes achieved 0.00 and 0.30 logMAR or better. These results are similar to those reported in the Cataract National Dataset paper where 46 and 91% were 0.00 logMAR and 0.30 logMAR or better, however immediate comparison is complicated by differing proportions of those with ocular co-pathology (37% RCOphth NOD vs 29% in the Cataract National Dataset). Part of this may be explained by a change in the EMR definition of co-pathology from conditions likely to be a cause of a guarded visual prognosis in the years contributing to Cataract National Dataset publications, to presence or absence of co-pathology in the years covered by this analysis. For those without ocular co-pathology, we found 50 and 94% were 0.00 and 0.30 logMAR or better, respectively compared with 51 and 95% in the Cataract National Dataset papers. We used slightly different definitions for pre and postoperative visual acuity measures than those in the Cataract National Dataset (we did not consider pinhole visual acuity as a proxy for missing preoperative visual acuity data, and used the better of UDVA and pinhole for absent postoperative CDVA data, rather than the best of UDVA, CDVA, or pinhole). 2 Repeat analyses using the earlier definitions made no clinically meaningful difference to our results. In comparison with the outcomes from the recent European Registry of Quality Outcomes for Cataract and Refractive Surgery (EUREQUO) sub-analysis 16 on 274 000 eyes without ocular co-pathology, our results appear slightly less favourable with EUREQUO reporting 70 and 98% had a postoperative CDVA of 0.00 and 0.30 logMAR or better. However, these are CDVAs only, so each case has undergone a formal refraction.
Cataract surgery does not always result in an improvement in visual acuity or patient satisfaction with visual function. The EUREQUO analysis reported that for those with a preoperative visual acuity of 0.00, 0.10, and 0.20 logMAR or better, 11.9, 2.3, and 1.4% cases had worse CDVA following surgery. 16 For those with 0.00 A strength of this study is that the data were nonselective, pooled, and anonymized, so they may be more generalisable than data obtained from randomized controlled trials, and less subject to publication bias than single-centre case series. 18 The 28 sites providing data for these analyses represent B15% of the annual number of cataract surgeries in England and Scotland as although there is encouragement to collect data to nationally agreed standards, each NHS Trust is independent resulting in variable adoption of EMR systems. All data used in these analyses were from a single EMR platform (Medisoft Ophthalmology) and were from Trusts that have agreed to contribute data to the NOD and were able to provide this. The recording of intraoperative complications is mandatory on the EMR and accuracy depends on surgeons recording their complications faithfully. It is not possible to confirm how reliably a surgeon records these, and surgeons may have a natural reluctance to record complications even though their results were anonymized. Data on possible underreporting rates may be partly inferred by the o0.1% operations that had an unplanned anterior vitrectomy without PCR having been reported to have occurred. Another limitation of our study is incompleteness of particularly postoperative data, with for example, only 53.1% of eyes being eligible for the postoperative visual acuity analyses.
The finding of an eight times higher risk of endophthalmitis if PCR occurred is consistent with data from a recent meta-analysis that found PCR was associated with a six times higher risk of endophthalmitis, 19 and was the intraoperative factor with most contribution to endophthalmitis risk. 19 Although the association of PCR with a higher risk of retinal detachment is well known, our findings of 42, 24, and 18 times higher likelihoods of retinal detachment surgery within 3, 6, and 12 months, respectively, of cataract surgery if PCR occurred are higher than previously reported, but do suggest the difference in risk decreases with time. Our results are comparable with the Swedish NCR where a 15 times higher odds of retinal detachment surgery within 3 years was found if a capsule complication occurred. 20 The rates for retinal detachment and endophthalmitis after cataract surgery (both 0.03%), must be interpreted with caution as they are likely to represent minimum values owing to the possibility that cases with these complications may have been treated at other NHS Trusts not contributing to this data set.
Overall, this analysis reports pragmatic data from a national database of cataract surgery outcomes allowing benchmarking of surgical units and individual surgeons, and gives an overview of the results that may be anticipated in the UK NHS. Our findings suggest that the trend of operating at lower levels of visual impairment has stopped or even reversed, and that visual outcomes and surgical complication rates have been stable over the past decade.

Summary
What was known before K Posterior capsule rupture or vitreous loss or both occurs in B2% of cataract operations. K For eyes without co-pathology, approximately half achieve 0.00 logMAR or better best-measured visual acuity after cataract surgery, and 95% are 0.30 logMAR or better.
What this study adds K Visual outcomes, and the rate of posterior capsule rupture or vitreous loss or both appear stable over the past decade. K For eyes without co-pathology, postoperative UDVA was 0.00 and 0.30 logMAR or better in 27 and 81% eyes. K Posterior capsule rupture or vitreous loss or both was associated with a 42 times higher risk of retinal detachment surgery within 3 months and an 8 times higher risk of endophthalmitis.

Conflict of interest
PHJD's employer received unrestricted funding from The Special Trustees of Moorfield's Eye Hospital to analyse these data. RLJ is the Medical Director of Medisoft Limited, which developed the electronic medical record from which the data were extracted for the first iteration of the National Ophthalmology Database.