Br J Ophthalmol 92:1483-1487 doi:10.1136/bjo.2008.140509
  • Original Article
    • Clinical science

23-gauge versus 20-gauge system for pars plana vitrectomy: a prospective randomised clinical trial

  1. B Wimpissinger1,2,
  2. L Kellner1,2,
  3. W Brannath3,
  4. K Krepler1,2,
  5. U Stolba1,2,
  6. C Mihalics1,2,
  7. S Binder1,2
  1. 1
    Department of Ophthalmology, Rudolph Foundation Clinic, Vienna, Austria
  2. 2
    Ludwig Boltzmann Institute for Retinology and Biomicroscopic Laser Surgery, Rudolph Foundation Clinic, Vienna, Austria
  3. 3
    Section of Medical Statistics, Medical University of Vienna, Vienna, Austria
  1. Dr S Binder, Department of Ophthalmology, Rudolph Foundation Clinic, Juchgasse 25, 1030 Vienna, Austria; susanne.binder{at}
  • Accepted 15 July 2008
  • Published Online First 14 August 2008


Aim: To compare the sutureless 23-gauge system with a standard 20-gauge system in pars plana vitrectomy.

Methods: 60 patients in two randomised groups were included in this prospective clinical trial. Pars plana vitrectomy with either 23- or 20-gauge instruments was performed. The main outcome measures were postoperative conjunctival injection and pain. Secondary outcome parameters were time of surgery, intraocular pressure, visual acuity and complications.

Results: Conjunctival injection (p = 0.0003) and postoperative pain (p = 0.01) were significantly reduced following 23-gauge vitrectomy compared with the 20-gauge procedure. Opening (p = 0.006) and closure times (p<0.00001) were significantly shorter, and vitrectomy time (p = 0.001) significantly longer in the 23-gauge system compared with 20-gauge vitrectomy. However, retinal manipulation and overall surgery times did not differ significantly between both groups. The same applies for eye pressure, distance and reading acuity. Regarding complications, two choroidal haemorrhages and one flat serous choroidal detachment occurred in the 23-gauge group.

Conclusion: The 23-gauge system for pars plana vitrectomy offers significantly higher patient comfort during the early postoperative period. Time of surgery is almost equal—a shorter time for wound closure is neutralised by a longer vitrectomy time in the 23-gauge group.

Pars plana vitrectomy (ppVE) has seen significant changes in the last several years. Self-sealing sclerotomies with smaller-gauge instruments have first been described by Chen et al in 1996, and a 25-gauge sutureless trans-conjunctival system for pars plana vitrectomy has been used since 2002.13 In 2005, Eckardt presented a sutureless 23-gauge system.4 The rationale was to create a small-incision sutureless vitrectomy system that could overcome some of the problems with the 25-gauge system,5 but still offers possible advantages, such as better patient comfort or shorter surgery time compared with conventional 20-gauge vitrectomy. However, the system has not been evaluated in a prospective randomised clinical study. This prompted us to perform the present prospective randomised controlled trial comparing 23-gauge with conventional 20-gauge vitrectomy.


Institutional Review Board approval was obtained and all interventions followed standard of care practices for vitreoretinal diseases. All interventions adhered to the tenets of the Declaration of Helsinki.

A randomised prospective clinical trial was designed, comparing 23- and 20-gauge vitrectomy systems in patients requiring vitreoretinal surgery. Inclusion criteria were a patient age of 18 years or older and signed informed consent together with one of the following diagnoses: epiretinal membrane, vitreous haemorrhage, diabetic retinopathy, asteroid hyalosis, macular oedema, macular hole or primary retinal detachment. Cases of combined vitrectomy and cataract surgery were also included. Exclusion criteria were a history of previous vitreous or retinal surgery, need for silicone oil tamponade, pregnancy and breastfeeding in women. One eye per patient was included in the study. Four surgeons were randomised to perform study procedures. Each had performed at least 10 small incision vitrectomies prior to study entry. Patients were randomised in blocks of random size using pre-prepared randomisation envelopes. The randomisation was stratified by the surgeons.

Primary outcome measures of the study were postoperative pain and conjunctival injection. Secondary outcome measures were time of surgery divided into wound-opening, vitrectomy, retinal manipulation (in eyes with epiretinal membranes, macular holes or oedema), and wound closure, as well as eye pressure, visual acuity, retinal status, and intra- or postoperative complications. In addition, the surgeon’s personal judgement of surgical tools and technical difficulties had to be documented.

Patients’ subjective pain and conjunctival injection, eye pressure and retinal status were assessed on the day before surgery, the first, second and third postoperative day, as well as 1 week, 1 month and 3 months following surgery. Visual acuity had to be measured the day before surgery, 1 week, 1 month and 3 months following surgery.

The number of subjects to be studied was determined by statistical prestudy evaluation (by a two-group t test with a 0.05 two-sided significance level). Sixty consecutive patients, who met the inclusion criteria, were divided into two randomised groups, 30 for the 23-gauge system and 30 for the 20-gauge system (fig 3).

Figure 3 Flow of participants through each stage of randomisation. ppVE, pars plana vitrectomy.

Assessment of conjunctiva

The conjunctival injection was graded on a scale from 0 to 4. 0, no conjunctival injection; 1, mild (affecting one quadrant); 2, moderate (affecting two quadrants); 3, intense (affecting three quadrants); 4, severe conjunctival injection (affecting four quadrants).

Assessment of postoperative recovery

Pain and discomfort were graded using a scale from 1 to 4 (1 = no pain, 2 = mild, 3 = moderate, 4 = severe pain).

Intraoperative time measurement

A stopwatch had to be started at the first cut into the conjunctiva. The times needed for wound opening (conjunctival and scleral preparation), for vitrectomy, for retinal manipulation and for wound closure (withdrawal of the cannulas, closing and suturing of sclerotomies and conjunctiva) were evaluated separately.

Intraocular pressure

Intraocular pressure (IOP) was measured by Goldmann applanation tonometry.

Visual acuity

Visual distance acuity using the log minimum angle of resolution (logMAR) charts according to the guidelines of the TAP study (test distance 4 m, letter-by-letter scoring, chart luminance 100 cd/m2) and visual reading acuity using standardised, German-language Radner reading charts were measured using the patient’s optimal refractive correction.6 7 In patients unable to read the largest test sentence, the Radner value was defined as 1.5.

Surgical technique

Surgery could be performed in general or local anaesthesia in all patients. Two commercially available surgical systems (DORC, Zuidland, The Netherlands; Oertli Instrumente AG, Berneck, Switzerland) were used for 23-gauge vitrectomy with cutting rates of 1200–1500/min and 1600–2500/min, respectively (20–25 vs 27–42 Hz) and a suction rate of 400 ml/min. Two 20-gauge systems (Alcon, Hünenberg, Switzerland as well as the above-mentioned Oertli system) were used with a suction rate of 150–200 ml/min.

In the 23-gauge group, the conjunctiva had to be pushed laterally using a pressure plate. Next, the two-step incisions were made tunnel-like transconjunctivally using a sharp 23 blade at a 30° angle parallel to the corneoscleral limbus facilitating insertion of the microcannulas. After vitrectomy, the microcannulas had to be removed and the sclerotomies covered by the conjunctiva.

With the 20-gauge system, the conjunctiva had to be opened in a nasal triangle and a temporal quadrangle 1 mm from the limbus followed by scleral incisions 3.5–4 mm behind the limbus. After vitrectomy, the sclerotomies and conjunctiva had to be closed with vicryl 7.0 sutures.

In both groups, a thorough vitrectomy had to be performed with the goal of removing approximately 80% or more of vitreous. An intraocular tamponade with air or gas was not required by the protocol and only performed according to the surgeon’s decision.

Statistical methods

All endpoints were analysed using exact Wilcoxon rank sum tests. To account for the multiplicity with regard to the two primary endpoints, the tests for the two primary endpoints were adjusted according to the Bonferroni correction. Hence, they were tested at the two-sided level 0.025. All secondary endpoints were tested at the unadjusted two-sided level 0.05.


Recruitment of patients started in February 2006 and was completed in March 2007. Follow-up was completed in June 2007.

No significant differences between the two treatment groups were observed for patient characteristics (table 1).

Table 1 Summary of qualitative baselines values

Conjunctival injection and postoperative pain

Postoperative conjunctival injection (p = 0.0003) and subjective postoperative pain (p = 0.01) were significantly less pronounced with the 23-gauge than with the 20-gauge system (figs 1, 2, table 2). Statistical significance is at the Bonferroni corrected level 0.025. One month after surgery, nine patients of the 20-gauge and one patient of the 23-gauge group had conjunctival injection of more than 0 on the respective scale (“>0”). Three months after surgery, one patient in the 20-gauge and no patients in the 23-gauge group had conjunctival injection >0. After the first postoperative month, four patients in the 20-gauge group and two patients in the 23-gauge had postoperative pain. After 3 months, no patients in the 20-gauge and one patient in 23-gauge group had postoperative pain.

Figure 1 Conjunctival injection before (baseline) and 1 day (1 d), 2 days (2d), 3 days (3d), 1 week (1w), 1 month (1m) and 3 months (3m) after surgery, and the p value of the Wilcoxon rank sum test for the group comparison of the conjunctival injection area under the curve. Data are presented here as means (SD) (23-gauge: n = 30, 20-gauge: n = 30).
Figure 2 Patients’ subjective pain before (baseline) and 1 day (1 d), 2 days (2d), 3 days (3d), 1 week (1w), 1 month (1m) and 3 months (3m) after surgery, and the p value of the Wilcoxon rank sum test for the group comparison of the pain area under the curve. Data are presented here as means (SD) (23-gauge: n = 30, 20-gauge: n = 30).
Table 2 Summary of the different endpoints per treatment group and p values of the Wilcoxon rank sum tests for the group comparisons

Time for surgery

Statistically significant differences were found in the duration of wound opening, vitrectomy and wound closure. The treatment groups were similar with respect to the retinal manipulation time and total surgery time (table 3).

Table 3 Means and standard deviations of the measurements for each treatment group, and p values of the Wilcoxon rank sum tests for group comparisons

Visual acuity

Distance and reading visual acuity changes from baseline to 3 months postsurgery were similar in both treatment groups (table 2). A statistically significant increase in distance visual acuity (20-gauge, p = 0.0008; 23-gauge, p = 0.0002) and a significant increase in reading acuity (20-gauge: p = 0.006, 23-gauge: p = 0.02) were observed in both treatment groups 3 months after surgery.

Eye pressure

The eye-pressure change from baseline to days 1 and 2 did not differ significantly between the two treatment groups (table 2). At days 1 and 2, eye pressure dropped below 10 mm Hg in four patients in the 20-gauge system group compared with eight patients in the 23-gauge system group (Fisher exact test: p = 0.3). The eye pressure was above 20 mm Hg in eight patients in the 20-gauge and in seven patients in the 23-gauge group (Fisher exact test: p = 1).


In the 20-gauge group, one trans-sclerotomy haemorrhage occurred during wound closure, making immediate aspiration using the vitrectome necessary. Another eye showed mild vitreous haemorrhage on the first postoperative day resolving within 3 weeks. In one case of macular hole surgery, an intraoperative retinal tear with minimal surrounding detachment occurred and was treated by endolaser.

In the 23-gauge group, one case showed a choroidal haemorrhage immediately after surgery, spontaneously resorbing after 1 week, followed by retinal detachment 5 weeks postoperatively due to a retinal tear. In this case, retinal reattachment surgery was necessary. Another eye suffered severe choroidal haemorrhage on the first postoperative day followed by tractive retinal detachment 7 months after surgery, making silicone oil surgery necessary. One case with flat serous choroidal detachment reattached after 1 week. In one case, the trocar was removed together with the vitreous cutter. In eight patients, one or more scleral incisions (13/90) as well as the conjunctiva needed suturing.

Surgeon’s comments

The stiffness of 23-gauge instruments was judged to be almost comparable with the 20-gauge system. The more peripheral cutter opening in the 23-gauge system facilitated a very precise dissection of adherent vitreous. In small eyes, placement of the trocars could be challenging. However, no switch to the 20-gauge system was necessary in the 23-gauge group.


Small-incision vitrectomy offered less pain for the patients in the present study. This mirrors the results of recent studies.2 4 5 814 Today more often working people have to undergo vitreoretinal surgery and therefore take advantage of faster recovery. Another major issue in the discussion of sutureless vitrectomy is the possible shortening of surgical time. However, so far no prospective study has compared the time of surgery between 20- and 23-gauge vitrectomy systems. In our study, the time gain in wound opening and closure was equalised by a longer vitrectomy time, when attempting a similar amount of vitreous to be removed. Regarding postoperative visual acuity, we did not see a significant difference between the two study groups. Obviously, visual rehabilitation after small-gauge vitrectomy is comparable with that after conventional 20-gauge vitrectomy. This has also been reported in other studies.5 1518 Eckardt documented no surgical complications in his series of 23-gauge ppVE.4 Theoretically, in sutureless surgical procedures there might be a higher risk of endophthalmitis. However, no case of endophthalmitis occurred in our study, but the number of patients is too small for a final judgement of this issue. Although experienced vitreoretinal surgeons participated in the study, we did see choroidal haemorrhage in two cases. One of these haemorrhages resorbed without further intervention. One eye needed revitrectomy and silicone oil tamponade after a tractive retinal detachment several months after surgery and ended up with ambulatory vision. Interestingly, neither these two cases nor the flat serous choroidal detachment were connected to severe hypotony. This complication warrants further observation although no statistically significant difference between the two study arms could be detected. A big issue discussed at the moment is the frequency of postoperative hypotony using small incision vitrectomy. In our 23-gauge group we found hypotonies in six (20%) patients, being severe (<5 mm Hg) in two (7%) on postoperative day 1. However, no statistically significant difference in postoperative eye pressure could be detected between the two groups. This stands in contrast to our previous 25- versus 20-gauge study, where significantly more hypotonies were observed in the first three postoperative days in the 25-gauge group. This is most likely related to the oblique incisions in 23-gauge vitrectomy.4 Fluid–air exchange at the end of surgery might be able to minimise this complication. In the present study, an intraocular tamponade was not required by the protocol and only given by surgeon’s decision. In a retrospective study by Fine et al examining 23-gauge vitrectomy in 77 eyes, hypotony occurred in two patients on postoperative day 1 with one patient requiring sclerotomy suture.17 In our study, one or more scleral incisions as well as the conjunctiva needed suturing in eight (26%) patients. This was done to increase safety and to prevent hypotony.

In conclusion, 23-gauge vitrectomy offered considerably less pain for patients during the early postoperative period. However, no surgical time gain could be reached in our study. The 23-gauge system seems to be a safe system suitable for a broad spectrum of vitreoretinal diseases.


  • Competing interests: None.

  • Ethics approval: Ethics approval was provided by Ethikkcommission der Stadt Wien.

  • Patient consent: Obtained.


Responses to this article