Article Text

other Versions

Download PDFPDF
Triamcinolone during pars plana vitrectomy for open globe trauma: a pilot randomised controlled clinical trial
  1. Philip J Banerjee1,2,
  2. Wen Xing1,2,
  3. Catey Bunce1,2,
  4. Malcolm Woodcock3,
  5. Aman Chandra1,2,
  6. Robert A H Scott4,
  7. David G Charteris1,2
  1. 1Moorfields Eye Hospital, London, UK
  2. 2NIHR Clinical Research Facility at Moorfields Eye Hospital, London, UK
  3. 3Royal Air force and Worcestershire Acute Hospitals NHS Trust, Worcester, UK
  4. 4Birmingham and Midland Eye Centre, Birmingham, UK
  1. Correspondence to Dr Philip J Banerjee, Vitreoretinal Service, Moorfields Eye Hospital, 162 City Road, London EC1V 2PD, UK; philip.banerjee{at}


Purpose To investigate the feasibility of conducting a randomised controlled trial in patients undergoing pars plana vitrectomy surgery following open globe trauma (OGT). Additionally, to investigate the treatment effect and toxicity of intensive anti-inflammatory agents.

Methods A 2-year, pilot, single-centre prospective, participant and surgeon-masked randomised controlled trial (RCT). Forty patients requiring vitrectomy surgery following OGT were randomised to either standard (control) or study treatment (adjuncts) in a 1:1 allocation ratio. Perioperatively, the adjunct group received intravitreal and subtenons triamcinolone acetonide, oral flurbiprofen and guttae prednisolone acetate 1%. The control group received standard care. Primary outcome was anatomical success at 6 months. Secondary outcomes included final visual acuity, occurrence of proliferative vitreoretinopathy, intraocular pressure rise, number of operations and recruitment rate.

Results 40 patients were recruited within 21 months. Primary outcome assessment showed similar results in anatomical success with 50% (10/20) in the adjunct group compared with 47% (9/19) in the standard group (OR 1.11, 95% CI 0.316 to 3.904). Visual outcomes were better in the adjunct group with a final median visual acuity of 31 Early Treatment Diabetic Retinopathy Study (ETDRS) letters compared with 25 ETDRS letters in the standard group. A higher proportion of patients gained 10, 20 and 30 ETDRS letters in the adjunct group (80%, 65% and 50%, respectively) compared with the standard group (52.6%, 52.6% and 42.1%). Fewer adjunct patients (15%, n=3) had poor visual outcomes (Zero ETDRS letters) compared with 42.1%, (n=8).

Conclusions An RCT in this population is deliverable and estimated recruitment rates are realistic. Results and patient discussions determined that the definitive study should have vision as a primary outcome. This pilot study is supportive of there being a positive treatment effect of intensive anti-inflammatory agents in OGT.

Trial registration number European Clinical Trials Database 2007-005138-35; Results.

  • Trauma
  • Eye (Globe)
  • Inflammation
  • Treatment Surgery

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Ocular trauma is an important cause of visual impairment and blindness worldwide and a leading cause of blindness in young adult males.1 Globally it has been estimated that 1.6 million people are blind as a result of ocular trauma with 2.3 million suffering bilateral low vision and up to 19 million with unilateral blindness or low vision.2 Ocular trauma is the most common cause of unilateral blindness worldwide, with significant socioeconomic repercussions in developing countries.2

Eyes sustaining penetrating or open globe trauma (OGT) are a group at high risk of severe visual impairment. Retinal detachment (RD) is common in these eyes, and multiple surgical interventions are often necessary.3 ,4 Proliferative vitreoretinopathy (PVR) is the most common cause of recurrent RD and visual loss in eyes with OGT. It is estimated to occur in 10–45% of all OGT.5–13

Corticosteroid treatment can potentially influence both the inflammatory and proliferative components of PVR. Experimental work has suggested that the corticosteroid triamcinolone acetonide can reduce the severity of PVR if administered intravitreally14 ,15 and via periocular administration.16 Laboratory work has also confirmed no demonstrable retinal toxicity of triamcinolone.17

In 2000, Jonas et al18 reported that intravitreal crystalline cortisone was well tolerated in PVR cases undergoing vitrectomy. Previous small-scale, uncontrolled clinical studies of PVR have suggested that systemic prednisolone,19 infused dexamethasone20 and intravitreal triamcinolone21 ,22 may reduce the severity of PVR, although none of these studies were of sufficient power to provide a definitive answer. A previous pilot study by our group has shown that triamcinolone is well tolerated in PVR cases undergoing vitrectomy and silicone oil exchange and that a combination of adjuncts targeting the inflammatory component of the PVR process may be a potential treatment to prevent PVR.23

Flurbiprofen is an oral non-steroidal anti-inflammatory drug most commonly used to treat musculoskeletal disorders. In ophthalmic practice, it is frequently used in the management of non-necrotising scleritis. The rationale for its use in this study stems from evidence suggesting that non-steroidal anti-inflammatory medications given perioperatively may limit the degree of blood–retinal barrier breakdown and have been shown to inhibit cellular proliferation in vitro.24–26 Furthermore, combining corticosteroids and non-steroidal anti-inflammatories has been shown to have a synergistic effect in reducing blood–ocular barrier breakdown.24

The primary aim of this pilot study (the Adjuncts in Ocular Trauma trial) was to determine the feasibility of conducting a prospective randomised controlled trial (RCT) using the aforementioned investigational medicinal products (IMPs) in this disease population. We aimed to acquire data in order to power a definitive study.

Patients and methods

This phase II pilot, single-centre, participant and surgeon-masked, prospective RCT trial was performed at Moorfields Eye Hospital NHS Foundation Trust between September 2011 and November 2013.27  An independent Data Monitoring Committee (DMC) and Trial Steering Committee (TSC) provided study oversight throughout the duration of the trial.


The study population consisted of male and female patients between 18 and 80 years of age. Eligible patients were those with an open globe injury undergoing pars plana vitrectomy (PPV) either following primary injury repair or as a primary procedure itself, that is, intraocular foreign body (IOFB) injuries. (An open globe injury is defined as a full thickness wound of the eye wall.)

The exclusion criteria were (1) history of (a) glaucoma, (b) PPV surgery to the affected eye, (c) known adverse reaction to any of the IMPs; (2) pregnant or breastfeeding females; (3) enrolment in other clinical trials; (4) inability to attend regular follow-up and (5) unable to give written informed consent.


A randomisation schedule of 40 treatment allocations against 40 study IDs was produced by a statistician using random permuted blocks of varying sizes. The randomisation schedule was provided to the pharmacy who drew up treatment packs according to the randomisation schedule. Participants were allocated to one of two groups in a 1:1 ratio. Twenty patients received standard surgical care (control group) and 20 patients received standard surgical care in addition to supplementary adjunctive anti-inflammatory agents (adjunct group).


Both groups received the standard vitreoretinal operative procedure that their ocular condition required. Consultants or senior fellows (second-year fellowship) performed the operative procedure. The following preoperative, intraoperative and postoperative treatment regimens were followed for each group.

Adjunct group

Preoperative treatment

Guttae prednisolone forte, one drop every 2 h for up to 1 week replaced any topical anti-inflammatory agents that the patient was using, and all other ocular treatments were continued.

Intraoperative treatment: (a) 4 mg/0.1 ml of intravitreal triamcinolone acetonide (Bristol Myers Squibb, UK) was injected into the vitreous cavity following closure of the scleral ports at the end of procedure, (b) 40 mg/1 mL of triamcinolone acetonide (Bristol Myers Squibb, UK) was administered as a posterior subtenons injection prior to suturing the conjunctiva and (c) standard subconjunctival antibiotic injection of cefuroxime 125 mg or gentamicin was administered.

Postoperative treatment

(a) Guttae prednisolone forte hourly for 1 week followed by a tapering regimen for 3–26 weeks thereafter depending on the degree of postoperative inflammation and cystoid macular oedema, (b) 50 mg flurbiprofen orally twice daily for 1 week and (c) routine topical antibiotics (guttae chloramphenicol 0.5% four times a day for 2 weeks) and mydriatics (guttae cyclopentolate 1%—frequency and duration at surgeon's discretion).

Control group


No additional treatment was given. Patients were instructed to continue with their current treatment, which may already include topical anti-inflammatory agents topical antibiotics and mydriatics.


Standard subconjunctival medications to include 4 mg of betamethasone and 125 mg cefuroxime or gentamicin.


Routine topical antibiotics (guttae chloramphenicol 0.5% four times a day 2 weeks) topical steroids (guttae dexamethasone 0.1%) and topical mydriatic (guttae cyclopentolate 1%—frequency and duration at operating surgeon's discretion).


Participants were masked to their treatment allocation for their entire duration of the trial, and preservation of masking status was confirmed upon exit. Additionally, the operating surgeon was masked until the end of the surgical procedure to avoid any bias in surgical management. It was not possible to mask the investigators at follow-up as the primary IMP, intravitreal triamcinolone, was sometimes visible on posterior chamber assessment.

Assessments and schedule

Postoperative study visits mirrored the routine schedule for vitreoretinal procedures at the study site and were conducted in the NIHR Clinical Research Facility at Moorfields Eye Hospital; day 1, day 10, 4–6 weeks, 3 months and 6 months. At each scheduled postoperative study visit, a full ophthalmic assessment was completed to include slitlamp biomicroscopy (±indirect binocular ophthalmoscopy when required) to assess retinal attachment status and PVR grade28 and parameters including best-corrected visual acuity (Early Treatment Diabetic Retinopathy Study (ETDRS) chart) Goldmann applanation tonometry and anterior segment assessment were recorded. Spectral-domain optical coherence tomography (OCT) was used to detect persistent submacular fluid.

Where silicone oil was used as a tamponade agent, its routine removal was not considered a reoperation and routine subsequent follow-up visits were followed until able to return to the study visit schedule. Other vitreoretinal surgical interventions over the trial period were considered reoperations and recorded as such.

The primary outcome measure was anatomic reapposition of the remaining retina to the retinal pigment epithelium in the absence of an internal tamponade agent at 6 months post-primary vitrectomy surgery.

Secondary outcomes at 6 months post-primary vitrectomy surgery were performed: (i) best-corrected visual acuity (using ETDRS chart) and proportional increase from baseline, (ii) number of procedures required to achieve retinal reattachment, (iii) presence and grade of postoperative PVR28 at any time point, (iv) persistent submacular fluid found by OCT in the presence of retinal reattachment, (v) recruitment rate, (vi) case mix of ocular trauma and (vii) study retention rate.

Sample size

A total of 40 patients was considered a feasible number over the study period and expected to provide sufficient data to estimate the SD to power a definitive study.

Statistical analysis

The time taken to recruit patients is reported together with the number of patients who failed to provide outcome data. Baseline characteristics of the two groups were compared to assess the adequacy of randomisation. The OR of patients in whom anatomical retinal attachment remained at 6 months post-primary vitrectomy between the two treatment groups was reported with 95% CIs—however, it must be acknowledged that since power is low this treatment estimate will be imprecise. Secondary end point analyses comprised summary statistics for secondary outcomes by treatment group.


Figure 1 documents the consort flow diagram. Patient recruitment opened in September 2011. Forty-five patients suffering open globe injuries were identified as proceeding to PPV surgery and screened for eligibility. One patient elected not to proceed with secondary vitrectomy surgery and four patients were ineligible due to pre-existing glaucoma (n=2) and aged >80 years old (n=2). The remaining 40 eligible patients elected to participate in the trial and were recruited within 21 months of the study commencing. The study closed at the final visit of the final patient in November 2013.

Figure 1

Consort flow chart demonstrates high eligibility to enrolment ratio. Study retention was also favourable, with one patient in the control group lost to follow-up by the primary outcome time point. All other patients were included in the analysis.

Baseline demographic and ocular characteristics are summarised in table 1 showing comparable mean age, gender and ethnicity. A relative afferent pupillary defect was present in 8 out of 20 patients (40%) in the adjunct group compared with 9 out of 20 (45%) in the control group. Ocular injury types according to the Birmingham Eye Trauma Terminology System classification29 were comparable across the two groups, although there were two perforating injuries in the adjunct group, with none in the control group and more cases of IOFB in the adjunct group. RD was present in 10 out 20 patients in the adjunct group and 11 out of 20 in the control group (table 1). Grade C PVR was present at the time of the study vitrectomy in six patients in total (two in the adjunct group and four in the control group).

Table 1

Baseline characteristics

As per study protocol, oral flurbiprofen was omitted from the postoperative regimen in 3 out of the 20 patients in the adjunct group owing to medical contraindications to its use. Table 2 outlines the intraoperative techniques used at the time of the study vitrectomy. Posterior vitreous detachment induction was performed in eight patients in both groups. Choice of intraocular tamponade was comparable between the two groups, with silicone oil used in 55% (n=11) patients in the adjunct and 57.9% (n=11) in the standard group.

Table 2

Surgical techniques used during vitrectomy

Outcome data were available for 39 out of 40 patients, with one patient in the control group lost to follow-up after the 4–6-week visit. One further patient did not attend for review after 4 months. Data collected at this time point were used to populate the final outcome assessment case report for as their clinical condition was deemed stable and therefore likely reliably representative of 6-month data.

There was no observed difference in primary outcome between groups: 50% (n=10 of 20) achieved anatomical success without internal tamponade at 6 months in the adjunct group compared with 47.4% (n=9 of 19) in the control group, OR 1.11, 95% CI 0.316 to 3.904 (table 3).

Table 3

Primary outcome variable—6 months

Considering visual acuity at 6 months following study vitrectomy: final median visual acuity was 31 ETDRS letters (IQR 12.5–47.5) in the adjunct group compared with 25 letters (IQR 0–65) in the control group. Three patients (15%) in the adjunct group could read no letters on the ETDRS chart at 6 months compared with eight patients (42.1%) in the control group. Five patients in the control group read at least 55 letters compared with three patients who had received the adjunctive regimen (table 4). The visual acuity box plot (figure 2) demonstrates that the median vision in the adjunct group initially lagged behind that in the control group at day 10 but subsequently exceeded it by month 6. The proportion of patients with visual improvement in terms of number of letters gained is reported in table 4 with a greater proportion gaining 10, 20, 30 and 40 ETDRS letters in the adjunct group compared with the control group.

Table 4

Visual acuity at 6 months/change from 6 month

Figure 2

Box plot of median visual acuity (VA) at each time point. VA in the adjunct group initially lagged behind the control group at day 10, but was noted to be higher at month 6. ETDRS, Early Treatment Diabetic Retinopathy Study.

There was no observed difference in the number of operations to achieve success (as defined in the primary outcome measure) with seven patients in each groups achieving success with a single operation, two patients in each group requiring two operations and one patient in the adjunct group undergoing three procedures to achieve stable reattachment at 6 months.

Table 5 highlights the prevalence of PVR grade C at each time point, demonstrating a higher prevalence in the control group at day 10, weeks 4–6 and month 3. At 6 months, the prevalence of PVR becomes comparable (35% adjunct groups vs 26.3% control group).

Table 5

Prevalence of grade C proliferative vitreoretinopathy at each time point

Adverse events

There were no serious adverse reactions observed in either group. The number of adverse events (AEs) was comparable across both groups with 18 out of 20 adjunct patients suffering at least one AE compared with 16 of 20 control patients. There were no cases of postoperative endophthalmitis in either group. A slightly higher number of patients in the adjunct group (n=7, 35%) suffered at least one episode of elevated intraocular pressure (IOP) (>25 mm Hg) compared with five (25%) patients in the control group. The median IOP readings at each time point is displayed in the box plot of figure 3. A higher median IOP was noted in the adjunct group at day 10, but this subsequently becomes comparable between the two groups. There were more cases of postoperative uveitis in the control group (n=5) in comparison to the adjunct group (n=2). An equal number of patients (n=5) suffered at least one episode of hypotony (IOP <6 mm Hg) during the trial.

Figure 3

Box plot of median intraocular pressure (IOP) at each time point. Day 10 median IOP was higher in the adjunct group but becomes comparable to controls thereafter.

A summary of the clinical findings at 6 months post-study vitrectomy is displayed in table 6. Complete retinal reattachment (with or without an internal tamponade) was comparable between the two groups (60% vs 63.2%). More patients in the adjunct group maintained posterior retinal attachment (90%, n=18) compared with the standard group (79%, n=15). At 6 months, on OCT assessment, only one patient (standard group) was found to have persistent submacular fluid in the presence of an attached retina.

Table 6

Clinical findings at final visit, 6 months after initial vitrectomy

Patient discussion group

We met with patients and discussed whether they felt that vision or anatomical success should be the primary outcome for the definitive study and there was overwhelming consensus that vision was the more appropriate. This also offers the additional advantage of primary outcome assessor masking, which was not possible in this trial. We conducted a power calculation based on our pilot and estimated that 302 patients in total should be recruited.


To date, there have been no prospective RCTs in this patient population. The primary aim of this pilot study was to determine the feasibility of the conducting such a trial with a view to provide data to power a definitive RCT. The study was projected to run for 24 months consisting of an 18-month recruitment period and 6-month final patient final follow-up. We were able to recruit all 40 participants within 21 months with a favourable eligibility to enrolment ratio of 97.6%. This in part may be attributable to the pragmatic design of the trial as follow-up schedules and data entry points mirrored standard care. Furthermore, the natural history of both surgical and visual outcome is poor4 in this disease group, and thus participants were eager to enrol, frequently adopting a ‘little to lose’ attitude, particularly given the well-documented and tolerable side effect profiles of the IMPs. Participant retention rate was also favourable, losing only one patient to follow-up after their 4–6-week visit. This retention rate was higher than expected, with previous surgical trials at the study site reporting a 3–5% loss to follow-up rate.30–32

We observed no difference in anatomical outcome at 6 months with an approximate 50% success rate in both groups. Similarly, we did not observe a difference in the number of operations to achieve success. This is the first study to prospectively report anatomical success rates at 6 months in this disease group. Sheng et al33 retrospectively reviewed 90 eyes in elderly patients sustaining OGI. 14 of 15 eyes undergoing secondary PPV with silicone oil for RD achieved anatomical success. However, anatomical success included eyes with an intraocular tamponade in situ. Taking our cohort as a whole, we found 19 out of 39 patients (48.7%) undergoing vitrectomy surgery after OGT maintained retinal attachment in the absence of internal tamponade at 6 months, with 14 out of 39 (35.9%) achieving primary anatomical success. If eyes with a persistent intraocular tamponade are included, then our observed overall success rate is 61.5% (n=24 of 39) for complete retinal reattachment and 84.6% (n=33 of 39) for posterior retinal attachment.

Interestingly, the prevalence of postoperative PVR was higher at each time point up to month 3 in our control group. This could simply be reflective of a higher incidence of PVR at baseline in the control group (ie, four patients in the control group compared with two patients in the adjunct group), but could also be an indication of a trend towards a treatment effect of the IMP. The prevalence of PVR at month 6 becomes comparable in both groups. Triamcinolone acetonide when administered as a 4 mg/0.1 mL intravitreal injection has been shown to have a duration of effect of between 2 and 4 months.34 It is possible that the postoperative vitreoretinal scarring response was delayed by the IMP administration, followed by a resumption in proliferative activity with falling concentrations of corticosteroid after its clearance.

As expected, the nature and severity of the ocular injuries sustained varied markedly within the groups but fortunately not between the two groups. The median Birmingham Ocular Trauma Score35 was 53 (IQR 37–70) in the control group compared with 49 (IQR 43–70) in the adjunct group. Given the sample size, we accept that the sensitivity to detecting small differences between the groups is limited, which may be further confounded by the heterogeneity of the case mix. However, the adequacy of randomisation and comparison of characteristics at baseline showed no significant difference in covariates affecting outcome.

The visual and anatomical outcome of eyes undergoing vitreoretinal surgery following OGT remains poor. The largest retrospective study to date by Andreoli et al reported outcomes of 848 eyes suffering OGT, in which 245 (29%) eyes required vitreoretinal surgery. Despite comparable median baseline visual acuities of hand motions, the reported median final visual acuity was only 20/400 in those eyes requiring vitreoretinal surgery compared with 20/40 in those that did not.4 In our cohort, the median final visual acuities at 6 months were 31 ETDRS letters (20/250) and 25 ETDRS letters (20/320) in the adjunct and control group, respectively. When considering improvement from baseline as displayed in table 4, where 80% (n=16) gained ≥10 ETDRS letters compared with only 52% (n=10) in the control group, this difference becomes more clear. We also noted fewer patients with very poor visual outcomes (Zero ETDRS letters) in the adjunct group compared with the control (15.0% vs 42.1%). As this is an exploratory pilot study, we did not test for statistical significance; nonetheless, the trend is interesting. Covariates that could affect visual outcome such as severity of PVR, number of foveal-off detachments and the use of silicone oil were comparable between the two groups.

Additional clinical factors may affect visual outcome at 6 months. Although the presence of cataract at the final assessment was comparable between both groups, visually significant corneal scarring in those patients with injuries involving zone 1 was not included in the data set at the final clinical assessment. Its presence in disproportionate numbers in either group could have either exaggerated or mitigated against any observed treatment effect. Likewise, pre-existing ocular comorbidity (eg, amblyopia or macula pathology) may also affect final visual outcome and should be recorded and adjusted for accordingly in future studies where vision is investigated as the primary outcome measure.

Triamcinolone acetonide has been used off-label via the intraocular route in clinical ophthalmic practice for >30 years. In addition to its use as an intraocular surgical adjunctive tool for visualisation of the posterior hyaloid during PPV,36 it has been used to treat a variety of posterior segment ocular inflammatory pathology.37–40 It has a well-documented safety profile with the most common significant side effect recorded as elevated IOP.41 It has been investigated specifically to determine its effect on vitreoretinal scarring (PVR) with varying success.21 ,22 ,42 Our study supports its use in future trials as it was well tolerated, with comparable rates of IOP rises noted across both treatment groups and no demonstrable evidence of toxicity. Interestingly, flurbiprofen was omitted in 15% (n=3) of adjunct patients, questioning its inclusion as an IMP in future trials in this disease group.

This is the first RCT in this patient demographic and demonstrates that a prospective RCT in this disease group is feasible. Recruitment and retention rates are realistic. The use of intravitreal triamcinolone as a pharmacological adjunct at the time of primary vitrectomy following OGT is safe and suggested a trend towards a better visual outcome in our cohort. An adequately powered definitive RCT is justified and the findings of this pilot study suggest that replacing an anatomical primary outcome with visual outcome may be a plausible vitreoretinal surgical trial design.


The authors acknowledge the significant contributions by Ms Nicola Harris (trial manager), Ms Hayley Boston (trial co-ordinator), Ms Tapiwa Margaret Zvobgo, Ms Haydee Carbonell and Mr Kerry Waller (research nurses) in the organisation and running of the trial at the study site.



  • Contributors PJB: data acquisition, data analysis, trial management and drafted manuscript. WX and CB: data analysis, contributed to drafting manuscript and approved final manuscript. MW: secured funding, contributed to drafting manuscript and approved final manuscript. AC: data acquisition, contributed to drafting manuscript and approved final manuscript. RAHS: secured funding, contributed to drafting manuscript and approved final manuscript. DGC–secured funding, data acquisition, data analysis, study supervision and approved final manuscript.

  • Funding Ministry of Defence, UK, via Qinetic Ref No. RAO 24/5/59.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Moorfields Research Management Committee approval was obtained, a favourable opinion from the Research and Ethics Committee for Wales was received (07/MRE09/60) and the study was granted a clinical trials authorisation by the MHRA. The study was conducted in accordance with the International Conference on Harmonisation for Good Clinical Practice, as set out in the European Union Clinical Trials Directive (2001) and associated UK Regulations (2004). The study complied at all times with the Declaration of Helsinki (2000)..

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional unpublished data is available to persons outside the study team.

Linked Articles

  • At a glance
    Keith Barton Jost Jonas James Chodosh