Paradoxical increase in peripapillary retinal nerve fibre layer thickness during the acute phase of steroid‐induced increase in intraocular pressure in a patient with vernal keratoconjunctivitis

A 9yearold boy was referred to our outpatient clinic with severe vernal keratoconjunctivitis (VKC). He was treated with topical dexamethasone and ketotifen but required bilateral injections of supratarsal triamcinolone acetonide to obtain sufficient disease control. Shortly after the second injection with triamcinolone acetonide, the intraocular pressure (IOP) increased in both eyes, but this could be managed by pressure lowering topical medication. During this time the peripapillary retinal nerve fibre layer (RNFL) and ganglion cell layer (GCL) thickness were normal on optical coherence tomography (OCT), see Figure 1 and Figure 2a,e. Approximately 1 year later, the IOP in the right eye increased uncontrollably despite maximum medical therapy. Fundoscopy revealed severe cupping of the right optic nerve head (ONH) even though the RNFL thickness was within normal range, see Figure 2b. We therefore decided to perform subacute 360 degrees trabeculotomy and the IOP quickly normalized. Four months later, the RNFL and GCL were reduced, and corresponding glaucomatous damage was present on automated perimetry (AP), Figure 2c– d and Figure 3. One year later, the IOP increased in the left eye after a third supratarsal injection with triamcinolone acetonide, see Figure 1. As in the right eye, cupping of the left ONH was noted during increased IOP even though the RNFL and GCL thickness were within normal range, see Figure 2f. Trabeculotomy was therefore performed in the left eye also. Six months after IOP normalization we found glaucomatous visual field defects and reduction of the RNFL and GCL thickness in the left eye, se Figure 2g and Figure 3. 1 | DISCUSSION

A 9-year-old boy was referred to our outpatient clinic with severe vernal keratoconjunctivitis (VKC). He was treated with topical dexamethasone and ketotifen but required bilateral injections of supratarsal triamcinolone acetonide to obtain sufficient disease control. Shortly after the second injection with triamcinolone acetonide, the intra-ocular pressure (IOP) increased in both eyes, but this could be managed by pressure lowering topical medication. During this time the peripapillary retinal nerve fibre layer (RNFL) and ganglion cell layer (GCL) thickness were normal on optical coherence tomography (OCT), see Figure 1 and Figure 2a,e. Approximately 1 year later, the IOP in the right eye increased uncontrollably despite maximum medical therapy. Fundoscopy revealed severe cupping of the right optic nerve head (ONH) even though the RNFL thickness was within normal range, see Figure 2b. We therefore decided to perform subacute 360 degrees trabeculotomy and the IOP quickly normalized. Four months later, the RNFL and GCL were reduced, and corresponding glaucomatous damage was present on automated perimetry (AP), Figure 2c-d and Figure 3.
One year later, the IOP increased in the left eye after a third supratarsal injection with triamcinolone acetonide, see Figure 1. As in the right eye, cupping of the left ONH was noted during increased IOP even though the RNFL and GCL thickness were within normal range, see Figure 2f. Trabeculotomy was therefore performed in the left eye also. Six months after IOP normalization we found glaucomatous visual field defects and reduction of the RNFL and GCL thickness in the left eye, se Figure 2g and

| DI SC US SION
Topical glucocorticoid therapy is a cornerstone in the management of several vision-threatening eye diseases such as VKC and uveitis. Unfortunately, glucocorticoids increase the IOP in up to 30% of children (Ang et al., 2012). It is therefore paramount to monitor the IOP and to detect glaucomatous changes during treatment. Fundoscopic evaluation of the ONH and automated perimetry can be difficult to perform in children due to difficulties in cooperation. Screening for glaucomatous changes therefore often rely on peripapillary OCT.
In the present case peripapillary OCT did not demonstrate thinning of the RNFL or GCL-hallmarks of glaucomatous damage-as one would have expected with the observed cupping of the ONH. On the contrary, the RNFL thickness was increased while the IOP was acutely raised. Shortly after the IOP was surgically normalized severe thinning of the RNFL and GCL revealed that glaucomatous damage had in fact occurred during the periods of high IOP.
This paradoxical phenomenon is sparsely described in the literature, and the mechanism is not fully understood. In one study of acute angle closure glaucoma the RNFL thickness was significantly increased during episodes of very high IOP (>40 mmHg). When the IOP was later normalized progressive RNFL thinning became evident over the following 6 months (Liu et al., 2010). In contrast to this, the RNFL thickness remained stable in the comparison group with chronic angle closure glaucoma without episodes of very high IOP. Some studies suggest that, during episodes with very high IOP, the  blockage of retrograde axonal transport causes RNFL oedema (Lampert et al., 1968). This may cause falsely elevated measurements of RNFL thickness on OCT.
The present case is, to our knowledge, the first to describe such a phenomenon in children. Interestingly, 'swelling' of the RFNL and subsequent glaucomatous damage occurred in the left eye despite only a moderately increased IOP of around 30 mmHg for a short period of time.
This case demonstrates that peripapillary OCT can be misleading and should not be used as a stand-alone measurement in the monitoring of glaucomatous changes in children with high IOP. Inspection of the ONH and automated perimetry should also be used whenever possible. A paradoxical increase in the RNFL thickness during high IOP may be a sign of impending glaucomatous damage, but the value of this phenomenon should be demonstrated in larger studies.