British Journal of Ophthalmology 2002;86:827-829
© 2002 British Journal of Ophthalmology

Retinal detachment surgery in district general hospitals: an audit of changing practice

R Ling1, C James1, P Simcock2, R Gray3 and S Shaw4

1 Department of Ophthalmology, Torbay Hospital, Torquay, UK
2 West of England Eye Unit, Royal Devon & Exeter Hospital, Exeter, UK
3 Department of Ophthalmology, Musgrove Park Hospital, Taunton, UK
4 Department of Mathematics and Statistics, University of Plymouth, Plymouth, UK

Accepted for publication 11 February 2002

Keywords: retinal detachment surgery; district general hospitals; audit

Despite advances in the techniques of vitreoretinal surgery, rhegmatogenous retinal detachment (RRD) continues to pose a serious threat to vision. In the past, general ophthalmologists in the United Kingdom undertook most RRD surgery, with only the more "complex" cases being referred to retinal specialists. The recent trend towards subspecialisation has led to increasing numbers of primary RRD being managed by vitreoretinal surgeons in tertiary referral centres.1

We conducted an audit to determine the current success rate of primary RRD surgery in our subregion (south west subregion, SWSR). SWSR consists of the catchments of four district general hospitals (DGH) in Devon and Somerset (population 1.2 million). A previous subregional audit in 1991 showed that all consultant ophthalmologists undertook primary RRD repair, achieving a success rate of 71% with one operation.2 Since then, retinal subspecialisationhas meant that all RRDs are managed by three retinal specialists(CJ, PS, RG) in their individual DGHs, providing a subregionalsurgical retinal service. An informal cross cover arrangementexists between the three DGHs, with the result that patientsare very rarely referred out of the subregion.

Case notes of all patients who underwent surgery for primary RRD from January to December 1999 were retrospectively reviewed. Case ascertainment was achieved by reviewing theatre logbooks and identifying all vitreoretinal procedures that took place during 1999. A single investigator (RL) collected details of all the patients; 1999 was chosen so that there was a minimum of 1 year follow up. Retinal reattachment was assumed to be stable in the absence of any history of further retinal surgery in the notes.

The main outcome measures were (1) primary success: retinal reattachment with one operation, with no re-intervention during follow up, (2) final success: retinal reattachment with or without further intervention during follow up, and (3) visual acuity outcome.

The retrieval rate for case notes was 99%. A total of 127 cases(DGH A 36%, DGH B 41%, DGH C 23%) were treated in 1999, witha mean age of 59.6 years (range 1495).

Table 1 shows the characteristics of the retinal detachments at presentation. There was no significant difference in any of the characteristics listed in Table 1 between the three DGHs ({chi}2 test, p>0.05). A total of 102 (80%) patients were operated on within 48 hours of presentation. Detachment subtypes in the 20% of operations delayed for more than 48 hours included chronic macula-off detachments (13), chronic inferior detachments with atrophic holes (eight), chronic dialyses (three), and inferior combined schisis detachment (one). No patients with macula-on detachment developed macular detachment before surgery.

Table 1 Characteristics of primary RRD in the south west subregion during 1999
Macula attached
>6D myopia
 Single retinal break
Inferior retinal breaks*
PVR (grade C1 or worse)3
Total retinal detachment
Types of retinal breaks:
       Horseshoe tears
       Round holes
       Combined schisis detachment
       Giant tear
No (%) of retinal detachments 

60 (47)

97 (76)

33 (26)

72 (57)

29 (23)

6 (5)

11 (9)

83 (65)

22 (17)

8 (6)

3 (2) 

3 (2)

PVR = proliferative vitreoretinopathy.
   *Breaks in the 4 clock hours from 4 to 8 o'clock.

Overall, 74 (58%) cases had primary vitrectomy. The three DGHsdiffered significantly in their surgical approach (primary vitrectomyrate: DGH A 65%, DGH B 40%, DGH C 79%). Consultants performed85% of all operations, while trainees performed the remainderunder supervision.

The primary success rate of retinal reattachment was 85% overall,with a mean of actual documented follow up of 8.2 months (225months). This was a significant improvement from the 71% primarysuccess in 107 cases of primary RRDs in the previous subregionalaudit (p<0.05, {chi}2 test). There was no significant difference between the three DGHs in their individual primary success rate (DGH A 87%, DGH B 85%, DGH C 83%). Included in the primary success were three vitrectomy/oil cases that had not undergone oil removal at final follow up.

There were 19 (15%) primary failures. 11 were "early failures"(mean interval to redetachment 8 days) whose retinas did notreattach, or immediately redetached after absorption of gastamponade. They were due to inadequate retinopexy/adhesion (five),new/missed breaks (four), inadequate buckle (one), and proliferativevitreoretinopathy (one). In contrast, the eight "late failures"(mean interval to redetachment 69 days) all had successful initialreattachment, but subsequently redetached due to proliferativevitreoretinopathy. Logistic regression with success or "latefailures" as the dependent variable found retinal break in asuperior position (from 10 to 2 o'clock), myopia >-6 dioptres,and "aphakia"/pseudophakia without an intact posterior capsule,to be significant predictors of "late failures" (p<0.05).

A total of 15 primary failures underwent further retinal surgery;14 patients were reattached with one further operation (oneneeded two further procedures). The final success rate was 97%.This was not significantly different from the 93% of the previousaudit (p>0.05, {chi}2 test).

Table 2 summarises the visual outcome of our patients, represented by changes in logMAR visual acuity. Visual acuity was significantly improved in the primary success, macula attached and macula detached subgroups (p<0.05, the Wilcoxon test). Seventy per cent of primary success patients achieved a Snellen acuity of 6/18 or better at discharge.

Table 2 Preoperative and postoperative logMAR visual acuity
Overall (n=127)                                                                     
Macula attached (n=60)                                                           
Macula detached (n=67)                                                                       
Primary success (n=108)                                                                  
Primary failures following successful reintervention (n=15)
Preop Median (SD)

0.78 (0.88)

0.21 (0.62)

1.83 (0.90)

0.60 (0.88)

0.78 (0.90)

Postop Median (SD)

0.18 (0.65)*

0.16 (0.35)*

0.85 (0.65)*

0.18 (0.54)*

0.60 (0.99)

*p<0.05, Wilcoxon test

In 1991, all consultant ophthalmologists in our subregion undertookprimary RRD surgery, each managing on average 8.2 cases peryear. In the current audit cycle, the caseload had increasedto 42 cases per consultant, with three retinal specialists managingall the RRD. Assuming there was no significant change in thedetachment case mix between 1991 and 1999, the improvement inthe primary success rate is most likely to be associated withthe advent of vitreoretinal subspecialisation.3

Results of RRD surgery from three vitreoretinal units (VRU) in the United Kingdom1,4,5 provided indicators for comparison.With the exception of pseudophakic rate (Cambridge 12%, SWSR24%), our case mix was most comparable to the Cambridge subregion,1 with both series describing RRD from geographically well definedpopulations. Initially, there would appear to be significantdifferences in the primary success rate between the four studies(Cambridge VRU 90%, SWSR 85%, St Thomas's 84%, Moorfields 80%,p<0.05, {chi}2 test). However, it is debatable whether retinal reattachment achieved with silicon oil in situ can be considered "stable." After excluding eyes with silicon oil still in situ, the "oil-out primary success rate" was comparable (Cambridge VRU 80%, SWSR 83%, p>0.05, {chi}2 test).

The primary success rate, although significantly improved, wasnot 100%. Improvement on our "early" failures, mostly due to"technical" errors, can be achieved by appropriate choice ofsurgical technique. For the "late" failures, all consequenceof PVR, improvement may depend on the identification of "at-risk"cases,6 and the selective use of an "anti-PVR cocktail."7

To conclude, we have shown that the primary success rate of RRD surgery has improved following a change in practice towards retinal subspecialisation in our subregion. The anatomical success and functional outcome were comparable to results from other VRUs in the United Kingdom.1,4,5 The benefits of providing alocal vitreoretinal service, in contrast to referring patientsto a VRU in a tertiary centre, include prompt on-site surgicalaccess, and the availability of support from family and friendslocally to aid visual rehabilitation in the postoperative period.


  1. Comer MB, Newman DK, George ND, et al. Who should manage primary retinal detachments? Eye 2000;14:5728.
  2. James C. Results of primary retinal detachment surgery in the south west subregion, September 1990 to August 1991. Presented at the British & Eire Association of Vitreo-Retinal Surgeons meeting, 1992.
  3. Retinal Society Terminology Committee. The classification of retinal detachment with proliferative vitreoretinopathy. Ophthalmology 1983;90:1215.
  4. Sullivan PM, Luff AJ, Aylward GW. Results of primary retinal reattachment surgery: a prospective audit. Eye 1997;11:86971.
  5. Minihan M, Tanner V, Williamson TH. Primary rhegmatogenous retinal detachment: 20 years of change. Br J Ophthalmol 2001;85:5468.
  6. Asaria RHY, Kon CH, Bunce C, et al. How to predict proliferative vitreoretinopathy. A prospective study. Ophthalmology 2001;108:11846.
  7. Asaria RHY, Kon CH, Bunce C, et al. Adjuvant 5-fluorouracil and heparin prevents proliferative vitreoretinopathy. Results from a randomised, double-blind, controlled clinical trial. Ophthalmology 2001;108:117983.