Article Text


Indocyanine green accused
  1. T L Jackson
  1. Correspondence to: Timothy L Jackson Academic Department of Ophthalmology, The Rayne Institute, St Thomas’s Hospital, Lambeth Palace Road, London SE1 7EH, UK;

Statistics from

The case for and against ICG assisted macular surgery

The concept of retinal vital staining is not new.1–3 There has however been renewed interest in the subject since indocyanine green (ICG) was introduced as a vital stain to highlight the internal limiting membrane (ILM) during macular surgery. ICG’s well known role in cardiac and choroidal angiography exploits its properties as a fluorophore, but its role as a macular vital stain relies on its properties as a chromophore or biological stain. Biological stains such as ICG have specific atomic groupings that impart colour (C = S, C = N, N = N, N = O, and NO2). ICG is therefore able to stain the thin, optically clear ILM, assisting maculorhexis in much the same way that trypan blue assists capsulorhexis. ICG has more recently been used to help remove epiretinal membranes.4

Initial reports were favourable.5–7 Subsequently, some of the most prominent exponents of ICG reported possible toxicity,8 and enthusiasm waned. More recently there have again been papers showing a favourable outcome,9,10 and the debate continues. The article by Hillenkamp et al in this issue of the BJO (p 437) supports the continued use of ICG, but it should not be read in isolation. This editorial presents the case for and against ICG macular vital staining, considering the clinical and experimental evidence in turn.

Hillenkamp’s study suggests that ICG vital staining is both safe and effective. It reports a retrospective analysis of two consecutive series of patients undergoing removal of epiretinal membranes, with and without the aid of ICG. The strength of this study lies in its detailed analysis of visual function. Outcome measures included subjective improvement, Amsler grid, corrected visual acuity (VA), slit lamp biomicroscopy, 10 and 30 degree automated perimetry, Goldmann kinetic perimetry, and optical coherence tomography. The results of surgery with and without ICG were broadly similar, except that the ICG group had significantly fewer residual or recurrent epiretinal membranes. As noted by the authors, this may be because of longer follow up in the “no ICG” group, but it is distinctly possible that ICG enhanced membrane removal. Importantly, this study suggests that visual outcome was not compromised by ICG. This finding agrees with several reports showing good outcome following macular hole surgery with ICG assisted ILM removal.5–7

The clinical argument in support of ICG rests not only on reports of good outcome, but also on the deficiencies of the patient studies showing toxicity. Many employed staining routines that are not now thought to be ideal. Some used contact times of several minutes, and solutions of up to 5 mg/ml (0.5%), yet approximately 1 mg/ml produces adequate staining,10 and many surgeons now use exposure times of 30 seconds or less. In addition, many investigators dissolved ICG in 0.5 ml of distilled water, before mixing with 4.5 ml of a balanced saline solution (BSS). This produces a 275 mOsm solution that is hypo-osmotic relative to agents such as BSS (302 mOsm).11 It has been suggested that ICG toxicity is caused,12 or at least aggravated,11 by low osmolarity.

The experimental reports of ICG toxicity are also imperfect. One of the most influential studies exposed human cadaveric eyes to ICG and illumination with a surgical endolight.13 The selected histological images were dramatic, with profound disruption of cellular architecture. This is not consistent with clinical observations, as reports of ICG toxicity generally show only mild to moderate deficits. Other authors were not able to repeat these observations in freshly enucleated pig eyes.14 Further, the validity of using enucleated eyes to model vascularised, viable retina is uncertain.

And the verdict? A good judge will want to hear more evidence

Studies of surgical ILM specimens have included retinal cellular elements and this has been marshalled as evidence that ICG alters the cleavage plane during ILM removal.15 However, this has also been observed in studies that did not use ICG.16 Cell culture experiments also need to be interpreted cautiously, as they do not simulate the complex interactions of a multicellular environment such as the retina. Many authors relied on a mitochondrial dehydrogenase enzyme assay to measure cell viability, yet there is spectral overlap of ICG absorption and the blue formazan reaction product that this assay reads. Without careful rinsing routines or appropriate assay correction, it is possible that residual ICG may produce falsely low readings of viability in a concentration dependent manner.17 This may explain why some papers showed more damage with this technique than with ultrastructural analysis.18 Potentially useful live animal studies showing histological or electophysiological damage failed to model the clinical use of ICG, with exposure times of several days.19

It should be noted that not all experimental studies show that ICG is unsafe. Retinal pigment epithelium (RPE)11,20,21 and glial11 cell culture studies, as well as macular surgery in enucleated pig eyes,14 all found that at least some clinically useful preparations did not cause damage.

As in the case supporting ICG use, the case against ICG can be considered in terms of the clinical and experimental evidence. Clinical studies have shown visual field defects,22 loss of VA,15 and RPE atrophy.23 The only randomised trial of ICG vital staining showed a small but significant reduction in VA, albeit using a hypo-osmolar preparation.24 Of particular concern are reports showing the persistence of ICG for several weeks.25 In addition, the use of ICG may be harder to justify now that 0.15% trypan blue is available as an alternative chromophore (Membrane Blue, Dorc, Netherlands).

There are experimental studies showing ICG toxicity in RPE,18 Müller cells,11,13 retinal ganglion cells,26 and both ex vivo15 and in vivo19 whole retina. The fact that so many investigators found evidence of cell damage cannot be ignored, even if exposure routines were sometimes dissimilar to those used clinically.

Advocates of ICG may suggest that infracyanine green is a safe alternative but this may not be so. Infracyanine green (Serb, Paris, France) is similar to ICG, but does not contain iodine and comes with 5% glucose as the diluent. Some investigations suggest that this alters the absorption profile,27 reducing the risk of ICG mediated phototoxicity when cells are illuminated.17 However, experiments in cadaver eyes also show cell damage,28 and Hillenkamp’s paper suggests that the effect on the absorption profile may be less than initially thought.

And the verdict? A good judge will want to hear more evidence. What is needed experimentally is a well controlled study that includes vitrectomy, short exposure times, then careful histological and electrophysiological analysis—preferably in an animal with rods, cones, and a duplex retinal circulation. Although difficult in some species, surgery would ideally include posterior vitreous detachment and removal of the ILM. Clinically, an appropriately sized, preferably multicentre, randomised trial is needed using an iso-osmotic, low concentration ICG, brief exposure times, and appropriate outcome measures. The parameters measured in Hillenkamp’s paper represent a useful template, but they did not include electrophysiology or fluorescein angiography.

Until such trials are completed, clinicians who continue to use ICG have some guidance from the literature in terms of reducing potential risk. One suggestion is to restrict use to difficult cases, where the surgical assistance offered by ICG is likely to outweigh any risk of visual loss. The literature suggests using a low concentration (0.5–1.25 mg/ml), avoiding hypo-osmotic preparations, keeping the endo-illumination distance up and power down, selecting halogen rather than xenon light sources, and rinsing fully and immediately after dye application. Another suggested technique involves the application of viscoelastic over macula holes, to prevent direct contact with the RPE.

Given that the courtroom analogies of this editorial could potentially reflect a medicolegal reality, discussing the use of ICG in the consent process may be worthwhile. Having heard the case for and against ICG, patients can then reach their own verdict.

Note in Proof

The case for and against ICG assisted macular surgery


View Abstract


  • Competing interests: none declared

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.

Linked Articles