During the 141-month study period, there were 40 cases of surgically induced scleritis: 23 cases after PPV and 17 cases after SB. During the same period, there were 38,935 PPV in 28,439 patients and 1893 SB from 1595 patients conducted at all sites of Moorfields Eye Hospital NHS Foundation Trust. Seven cases of scleritis post-SB were excluded for incidence calculation as the inciting SB surgery was performed elsewhere (n = 4) or outside the study period (n = 3). Hence, the estimated annual incidence of surgically induced scleritis in our institution was 5.0 cases (95% CI: 3.4–7.5) per 100,000 PPV and 45.0 cases (95% CI: 24.4–82.5) per 100,000 SB. When considering only necrotising scleritis cases, the estimated annual incidence was 1.3 (95% CI: 0.60–2.86) cases per 100,000 PPV and 9.0 (1.6–32.7) cases per 100,000 SB.

Of the 23 patients (23 eyes) who developed scleritis after PPV, 16 (70%) were male, 12 (52%) were of white Caucasian ethnicity and their median age was 54 years (IQR 39–66) (Table 1). A history of significant ocular trauma was noted in two (8.7%) eyes. Scleritis developed after a single PPV operation in 11 (48%) eyes including three (13%) with prior phacoemulsification cataract surgeries. The remaining 12 (52%) eyes had multiple retinal procedures with a median of three PPV (IQR 2.25–4.75), including two eyes with previous SB, ten combined phacovitrectomy and one Baerveldt glaucoma implant. None had extracapsular cataract surgery through limbal incision or pterygium surgery, which are two leading causes of SINS in previous literature [1, 12].

Indications for PPV were retinal detachment (RD) (9 cases, 39%), removal of silicone oil post-RD repair (6 cases, 26%), scleral-fixated intraocular lens (IOL) implantation (3 cases, 13%), full-thickness macula hole (2 cases, 8.7%), macula proliferative vitreoretinopathy (1 case, 4.3%), iris-fixated IOL implantation (1 case, 4.3%), and intraocular foreign body removal (1 case, 4.3%). The vitrectomy gauge used were 23 G in 19 (83%) cases, and 25 G in 4 (17%) cases. Sclerostomy ports were sutured in 15 (65%) cases (Table 2).

Scleritis presented 3.6 weeks (IQR 2.1–51.7) after PPV, with BCVA of 0.69 logMAR (IQR 0.30–1.35) (Table 3). Scleritis was frequently associated with mild anterior uveitis of at least grade 0.5+ anterior chamber cells (10 cases, 43%), occasionally with ocular hypertension exceeding 21 mmHg (2 cases, 8.7%) and rarely with vitritis, cystoid macular oedema (CMO) or chorioretinal folds (1 case, 4.3%). None had documented interstitial or peripheral ulcerative keratitis.

Scleritis following PPV was anterior non-necrotising in 17 (74%) eyes and necrotising in 6 (26%) eyes. Anterior non-necrotising scleritis were of diffuse subtype in 16 eyes and nodular subtype in one eye. Two (7.8%) eyes with anterior necrotising scleritis also exhibited posterior scleritis – one at the time of anterior scleritis (with findings of serous retinal detachment, posterior coat thickening and T-sign on B-sign), and one at recurrence.

Oral NSAIDs in combination with topical steroids and/or NSAIDs were sufficient for post-PPV anterior non-necrotising scleritis cases. Flurbiprofen was the most commonly used oral NSAIDs with a median daily dose of 150 mg (IQR 150–300) and treatment duration of 2 weeks (IQR 2–4) in the post-PPV scleritis group.

Of the six eyes with necrotising scleritis following PPV, two required systemic steroids or immunomodulatory therapy (IMT); these two cases were microbiology-proven infective scleritis. One case required surgical removal of scleral-fixated IOL and the associated expanded polytetrafluoroethylene (GORE-TEX) suture. The remaining three cases initially presented as diffuse anterior scleritis, with scleromalacia being subsequently recognised during follow-up examinations. In these cases, scleritis responded adequately to oral NSAIDs and was deemed not to require further escalation of treatment by our uveitis team. Delayed recognition of necrotising scleritis was also noted in the landmark study by O’Donoghue et al., in which less than 60% of the SINS were identified at the initial clinical evaluation, with the remainder recognised during subsequent follow-up examinations or with anterior segment fluorescein angiography [1].

The aetiology of surgically induced scleritis post PPV was non-infective in 21 (91%) cases. There were two (8.7%) cases of microbiology-proven infective scleritis. Subconjunctival abscess was a feature in both cases, albeit not an initial presenting feature in one of the two cases. The causative pathogens were Haemophilus Influenza in one case, and mixed organisms in the other (Pseudomonas aeruginosa and Aspergillus fumigatus). Oral NSAIDs were insufficient to control inflammation in both cases. The former required oral prednisolone, moxifloxacin and surgical intervention for debridement and application of scleral patch graft. Oral prednisolone was tapered over 19 months. The latter initially responded to oral prednisolone, before deteriorating when dose was reduced. Treatment was escalated to intravenous methylprednisolone pulse therapy and ciclosporin which continued for another month before the diagnosis of co-existing fungal infection.

At 15 months’ follow-up (IQR 5–69), the final BCVA in the post-PPV scleritis cohort was 0.54 (IQR 0.08–1.61). The median change in BCVA at final follow-up from diagnosis was −0.18 (−0.40–0.13), indicating most eyes maintain or gain vision with treatment. Vision loss of at least one line was seen in five (22%) eyes. Four (17%) cases had one episode of scleritis recurrence. All recurrences had the same morphology as the initial presentation except in one eye (suspected posterior scleritis based on mildly reduced vision and T-sign on ultrasound B-scan). These responded to oral NSAIDs and did not require escalation of treatment to immunomodulatory treatment.

An underlying connective tissue or autoimmune disease was uncommon (13%) in the post-PPV scleritis cases; two patients had Marfan syndrome, and one had ulcerative colitis. Systemic vasculitis or rheumatoid disease was not detected in the five (22%) patients who underwent laboratory and radiographic work-up.

Of the 16 patients (17 eyes) who developed scleritis after SB, seven (44%) were male, five (31%) were of white Caucasians ethnicity and their median age was 38 years (IQR 31–57) (Table 1). A history of significant ocular trauma was noted in three (18%) eyes. Scleritis developed after a single SB procedure in 11 (65%) eyes. The remaining six (35%) eyes had multiple retinal procedures including previous SB (2 cases), combined PPV-SB procedures (2 cases), and subsequent PPVs after the initial SB (2 cases). Removal of suture triggered development of scleritis in one eye.

The scleral explants used were encircling bands (47%) in eight eyes (including four with an additional silicone tire segment, and one with an additional sponge), segmental silicone tire buckle in seven (41%) cases (including one with additional sponge), and sponge in two (12%) cases (Table 2).

Scleritis presented 33.4 weeks (IQR 5.7–756.7) after SB, with BCVA of 0.54 logMAR (IQR 0.08–1.36) (Table 3). Buckle extrusion was a presenting feature in seven (41%) cases. The latency period between surgery and scleritis onset is significantly longer for post-SB than post-PPV (p = 0.013).

Scleritis following SB was anterior non-necrotising (diffuse subtype) in eight (47%) eyes, necrotising in eight (47%) eyes and posterior in one (6%) eye. Scleritis after SB was frequently associated with mild anterior uveitis (4 cases, 24%) and rarely with ocular hypertension or vitritis (1 case, 6%). No keratitis, CMO or chorioretinal folds were reported.

The aetiology of surgically induced scleritis post SB was non-infective in 16 (94%) cases. There was only one (6%) microbiology-proven infective scleritis caused by Staphylococcus aureus.

Treatment included oral NSAIDs in the majority of cases; Flurbiprofen was most commonly used with a median daily dose of 300 mg (IQR 275–300) and treatment duration of 3 weeks (IQR 2–4) in the post-SB scleritis group. Five (29%) cases of necrotising scleritis required systemic steroids and two (12%) required immunomodulatory therapy (IMT). Oral prednisolone was successfully tapered over 6–9 weeks in three cases. One patient received long-term IMT due to subsequent diagnosis of a systemic vasculitis and the other received mycophenolate mofetil for 2.5 years before being successfully weaned off following the removal of scleral explant.

Oral moxifloxacin was used in three cases in which infection was suspected due to finding of eyelid erythema, swelling and mucopurulent discharge. Wound swabs and scleral explants were sent for microscopy, culture and sensitivity but only one (6%) showed positive culture with staphylococcus aureus from both conjunctival swab and scleral explant. One other case showed gram negative rods on microscopy but culture was negative.

The scleral explant was removed in 10 (59%) cases, including one (6%) requiring a scleral patch graft. One patient declined removal of the scleral explant despite recurrence of scleritis due to concerns of re-detachment. Amongst those who underwent removal of scleral explant, there was no recurrent retinal detachment.

At 8 months’ follow-up (IQR 5–21), the final BCVA was 0.48 (IQR 0.10–1.00). Scleritis recurred in 5 (29%) cases. The median change in BCVA from diagnosis to final follow-up was 0.00 (−0.20–0.08), with majority of patients maintain or gaining vision with treatment. Three (18%) patients lost at least one line of vision.

Prior to surgery, one patient was known to have multiple sclerosis. Of the six (35%) patients who underwent laboratory and radiographic work-up, one was found to have a systemic vasculitis (granulomatosis with polyangiitis). Taken together, an underlying autoimmune disease was uncommon (12.5%).