Sir,

We thank Mr Ahfat1 for the comments offered in response to our report of a series of patients with transient retinal artery occlusion (TRAO) occurring during phacoemulsification cataract surgery.2

Ocular compression devices were not used in any of the patients described in our series. However, we agree that consideration of potential mechanisms that may logically underlie TRAO is important to identify preventative strategies. Conceptually, any peri-operative manoeuvre that elevates intraocular pressure (IOP) above ocular perfusion pressure (OPP) may produce retinal arterial hypoperfusion. This may be ‘absolute’ if the induced IOP elevation exceeds systolic OPP, or ‘relative’ if IOP exceeds diastolic OPP—for the duration throughout which the stimulus is applied.

Excessive ocular compression, high infusion pressures, or over-hydration of corneal wounds may all contribute to the risk of TRAO. IOP may exceed 85 mm Hg during nuclear fragment removal depending on flow rates and infusion parameters.3, 4 Newer-generation phacoemulsification technologies can maintain a target IOP5 and protect OPP.4 This may reduce the risk of TRAO when compared with phacoemulsification platforms that utilise gravity fluidics or fixed pressure irrigation systems.

Patient factors must be considered when evaluating candidate risk factors for TRAO. High diastolic systemic arterial pressures may be protective, maintaining OPP throughout the cardiac cycle despite peri-operative elevations in IOP. Conversely, carotid artery stenosis is associated with reduced ipsilateral OPP6 and may risk TRAO in the context of relatively minor fluctuations in IOP.

TRAO may have a permanent impact on post-operative visual function despite technically uncomplicated phacoemulsification cataract surgery.2 We encourage the clinicians to exclude TRAO in all patients presenting with unexplained post-operative visual loss. Improved detection and reporting of TRAO may enable a prospective case-finding study in which candidate patient risk factors and intra-operative manoeuvres may be identified. In the absence of such a study, associations may be suggested, but cannot be proven.