Retinal Damage from Indocyanine Green in Experimental Macular Surgery

Abstract

purpose. Previous work has shown that indocyanine green (ICG)–assisted peeling of the inner limiting membrane (ILM) may cause retinal damage. In this study, a toxic and a photodynamic effect of ICG at the vitreoretinal interface was assumed.

methods. Ten human donor eyes were hemisected, and 0.05 mL of 0.05% ICG was poured over the trephined macula in eight eyes. After 1 minute, the dye was drained by irrigation. The macula in each of two eyes was illuminated for 3 minutes with wavelengths of 380 to 760, 380 to 620, or 620 to 760 nm. Two eyes were treated with ICG only, and two eyes were illuminated only. Retinal specimens from the macula and the untreated retina were processed for light and electron microscopy. The irradiance of the light source and the absorption properties of ICG were measured.

results. Exposure of the ICG-stained ILM to wavelengths beyond 620 nm resulted in severe damage to the inner retina, including loss of ILM, cellular disorganization, and fragmentation of the cytoplasm. ICG staining alone or in combination with wavelengths of 380 to 620 nm disclosed rupture of Müller cells with detachment of the ILM, but no other cellular disorganization. Eyes subjected to illumination only showed no abnormalities.

conclusions. The spectral absorption properties of ICG may account for a possible photodynamic effect of ICG at the vitreoretinal interface. ICG alone induces ILM detachment and disruption of Müller cells even without intentional peeling of the membrane. It is assumed that accumulation of the dye at the vitreomacular interface may enhance the concentration and osmolarity of ICG at the retina beyond intravitreous values and critical limits.

Since the advent of macular hole surgery and according to Gass’s theory ¹ that tangential traction at the vitreomacular interface is involved in the formation of idiopathic macular holes, intentional removal of the inner limiting membrane (ILM) has been proposed to completely relieve traction on the foveal neuroretina. ^{1 2 3 4 5 6 7} Although initially greeted with great skepticism, it is now accepted by many that intentional removal of the ILM is a feasible and safe maneuver, leading to good anatomic and functional results in macular hole surgery. ^{6 7 8} However, removing the ILM, which is only a few micrometers thick and nearly invisible, presents a technically challenging procedure even for experienced vitreoretinal surgeons, and ILM removal remains the currently most debated issue surrounding macular hole surgery. ^{9 10}  

The introduction of staining of the ILM with indocyanine green (ICG) was widely greeted with enthusiasm and further increased the popularity of peeling of the ILM. ^{11 12 13 14} ICG selectively stains the ILM in vivo, and greatly facilitates peeling the membrane in macular surgery. ^{15 16} By providing a clear contrast between the ILM and the retina, ICG has been proposed to increase the safety of intentional peeling of the ILM, and questions regarding a potential toxicity of ICG have been toned down by enthusiasm. ^{9 17 18 19 20}  

In our institution, however, ongoing prospective evaluation of patients who had been operated with ICG-guided removal of the ILM demonstrated a tendency toward less favorable results in terms of gain in lines, in terms of postoperative visual outcome, and incidence of visual field defects (Haritoglou C, manuscript submitted), compared with peeling of the ILM without the use of chemical agents. ^{8 19 21} Ultrastructural analysis revealed plasma membranes of Müller cells and other undetermined cellular debris adhering to the retinal side of the ILM in all ICG-stained specimens (Haritoglou C, Gandorfer A, Gass CA, Kampik A, ARVO Abstract, 3516, 2002). ¹⁹ In contrast, no retinal structures were found in specimens that had been removed without the use of ICG, neither before the introduction of the dye in autumn 2000, nor after having stopped ICG staining in March 2001. Neither the surgical techniques nor the surgeon had changed, and evidence arose that ICG was responsible for the untoward outcome. 

ICG is a commonly used dye with a long history of safety after intravenous administration. ²² After its introduction in 1957, ICG soon came into general use for recording dye dilution curves, in particular for the determination of cardiac output. ^{23 24} In ophthalmology, ICG digital angiography has been a major advance in the imaging of choroidal circulation with a high level of safety. ^{22 25 26 27 28} ICG has been used as a vital dye for donor corneal endothelium before penetrating keratoplasty and for staining of the lens capsule in eyes with mature cataract. ^{29 30} After intravenous application in animal models, ICG has very little toxicity. ^{31 32 33 34} However, the tricarbocyanine dye has infrared absorption properties with a peak absorption at 800 nm, resulting in the generation of singlet oxygen and the subsequent formation of lipid peroxides after photoactivation. ^{35 36} Recently, the photodynamic properties of ICG have been used to destroy colonic cancer cells in vitro and to perform photodynamic therapy at the choriocapillary layer in an animal model. ^{36 37 38}  

The present study evaluates the possibility of a toxic and a photodynamic effect of ICG at the human vitreoretinal interface. Therefore, we simulated the situation during surgery by using ICG for staining of the ILM and the light pipe of a commonly used vitrectomy instrument and investigated the ultrastructure of the retina after exposure to ICG and light of different wavelengths. 

Methods

Ten human eyes were obtained from five donors within 16 to 30 hours after death. The donors were two women and three men with a mean age of 45 years, ranging from 17 to 62 years. After conservation of the cornea, the iris and the lens were removed and discarded. A corneal trephine of 12-mm diameter was slowly moved through the vitreous, and the posterior pole including disc, macula, and the temporal arcades of the retinal vessels was trephined. With the corneal trephine left in place, an experimental chamber was created that allowed a standardized application of ICG to the macula without treating other areas of the retina, which served as a control. One milliliter balanced salt solution (BSS; Alcon Laboratories Inc., Fort Worth, TX) containing 0.05 mL of 0.05% ICG (Pulsion, Munich, Germany) was poured into the trephine and left in contact with the macula for 1 minute. The dye was then removed by irrigation, leaving the stained ILM clearly visible. The amount of 0.05 mL of 0.05% ICG diluted in 1 mL BSS as used in this model is equivalent to the clinical application of 0.2 mL of 0.05% ICG, as administered in vitreoretinal surgery into the vitreous cavity with an assumed volume of 4 mL. 

After removal of the dye, the ICG-stained posterior pole was exposed to different wavelengths for 3 minutes, using the standard light pipe of a vitrectomy instrument (Megatron; Geuder, Heidelberg, Germany). The fiberoptic was placed 8 mm above the posterior pole in a slightly oblique angle simulating the situation during surgery. The emission spectrum of the light device was modified by filters (Andover, Salem, NH) which were interposed between the light source and the fiberoptic. For details of illumination, see Table 1 . As the control, two eyes were treated with ICG only and another two eyes were exposed to the standard light pipe without the use of ICG. 

Using a corneal trephine of 6-mm diameter, retinal specimens were obtained from the treated macula and from the untreated retina directly adjacent to the trephined posterior pole to evaluate the impact of autolysis in each eye. The specimens were placed in phosphate-buffered 4% glutaraldehyde solution for fixation. The tissue was postfixed with Dalton fixative (osmium 2%), dehydrated, and embedded in Epon. Semithin sections were stained with 2% toluidine blue for light microscopy. For electron microscopy, ultrathin sections were stained with uranyl acetate and lead citrate. Tissue sections were analyzed by electron microscope (EM 9 S-2; Carl Zeiss, Oberkochen, Germany). Two observers graded at least five ultrathin sections of each retinal specimen independently. 

Photometry and High-Resolution Spectroscopy

The spectral irradiance of the cold light source of the vitrectomy instrument was measured by two different techniques. First, the overall spectral characteristics were investigated photometrically. Narrowband (full width at half maximum [FWHM], 20 nm) calibrated interference filters were used in combination with a calibrated photodiode (model 840; Newport, Irvine, CA). For each photometric exposure, the known transmission curve of the interference filter was numerically integrated to give the total energy-throughput value, which then was used to connect the measured photocurrent to the intensity. Based on these measurements, the broadband irradiance characteristics of the light source were determined in steps of 10 nm. In a second step, the light source was investigated by high-resolution spectroscopy. A cornerstone monochromator (Oriel Swiss, Romanel-sur-Morges, Switzerland) was used in combination with a blue sensitive charge-coupled device (CCD) detector (Zimpol, Zürich, Switzerland). Slit width of the monochromator was chosen to provide a 1-nm spectral window. The spectrum between 400 and 760 nm was scanned in steps of 1 nm. The sensitivity of the instrument is known to vary monotonically with wavelengths between 400 nm and 600 nm and remains constant above 600 nm. Therefore, the measured spectral intensity curve was calibrated to spectral irradiance by comparing the narrow-band spectral curve with the photometrically obtained values for distinct 20-nm-wavelength bands. The effect of different band-pass and cutoff filters (Andover) on the irradiance of the cold light source was measured in the same way. The filters were placed between the fiber optic output of the vitrectomy instrument and the monochromator. All measurements were performed with the cold light source at maximum power. 

Spectrophotometry

The absorption properties of ICG were analyzed by a spectrophotometer (model U 2000; Hitachi, Tokyo, Japan). Between 400 and 900 nm, the optical density of the ICG solution was determined with a scan rate of 200 nm per minute. Injecting 0.2 mL of 0.05% ICG into the fluid-filled vitreous cavity with an assumed volume of 4 mL results in a solution of 0.0025% ICG or higher. Therefore, ICG solutions of 0.0025% and 0.005% were measured. 

Results

Histology

Semithin sections of the four eyes that were treated with ICG only (Fig. 1A) or with ICG and illumination with wavelengths between 380 and 620 nm (Fig. 1C) showed detachment of the ILM from the retina with otherwise well-preserved retinal cytoarchitecture at this level of magnification. In the four eyes that were treated with ICG and wavelengths of 380 to 760 nm (Fig. 1E) or wavelengths between 620 and 760 nm (Fig. 1G) , respectively, there was loss of ILM and gross disorganization of the cellular architecture of the inner retina. The nerve fiber layer was disrupted, and ganglion cells were split away from the retina. Eyes that were illuminated without application of ICG (Fig. 1I) showed a regular cytoarchitecture of the retina with mild postmortem changes. Control specimens (Fig. 1B 1D 1F 1H 1J) , which were taken from the untreated retina in all eyes had a normal and well-preserved cytoarchitecture with mild postmortem changes, such as vacuolization of the retinal layers. 

Ultrastructural Findings

ICG-Treated Eyes without and with Exposure to Wavelengths between 380 and 620 nm.

The ILM was detached in all four eyes (Fig. 2) . Ultrastructural analysis disclosed cellular swelling and disruption of Müller cell processes. In the few cells that were not ruptured, large cysts were present at the cell boundary. Despite fragmentation of the basal cell membrane in most of the cells, the intercellular tight junctions did not appear to be ruptured. The detached ILM was completely intact. On the retinal side of the ILM, fragments of Müller cell membranes and cellular debris were adherent to the ILM. There was no difference in the ultrastructure of the retina between eyes that were treated with ICG only compared with the ICG-stained eyes that were illuminated additionally with wavelengths between 380 and 620 nm. Control specimens taken from the untreated retina adjacent to the trephined posterior pole showed an intact inner retina with mild postmortem changes. 

ICG-Treated Eyes after Exposure to Wavelengths between 380 nm and 760 nm or 620 nm and 760 nm.

There was gross disorganization of the cellular architecture of the inner retina in all four eyes (Fig. 3) . Ultrastructural analysis disclosed fragmentation of the cytoplasm with disorganization of cellular organelles. Untreated retinal specimens showed an intact retina with mild postmortem changes in all four eyes. There was no difference in retinal damage between eyes that were illuminated with wavelengths between 380 and 760 nm or 620 and 760 nm. 

Eyes after Illumination without ICG Staining.

Transmission electron microscopy revealed a regular retinal cytoarchitecture after illumination of the posterior pole with wavelengths between 380 and 760 nm, when no ICG was used for staining of the ILM (Fig. 4) . 

Quantitative Spectroscopy

The light source of the vitrectomy instrument emitted between 380 and 760 nm. Below 380 nm and beyond 760 nm, no irradiance was measured. At 400 nm, we found an intensity of 13% of the maximum intensity at 550 nm (Fig. 5A) . Twenty-eight percent of the total irradiance of the light source was emitted beyond 600 nm. A cyan-subtractive filter (model 590FD24; Andover) successfully blocked the red part of the spectrum. The 50% level was reached at 612 nm and the 5% level at 628 nm (Fig. 5B) . The combination of a yellow- with a magenta-subtractive filter (models 550FD26 and 520FD22; Andover) allowed transmission of the red part of the spectrum only. Below 600 nm, there was only 3% transmission, whereas the wavelengths band beyond 615 nm was transmitted completely (Fig. 5C) . 

Absorption Properties

In 0.0025% ICG solution, the maximum absorption was reached at approximately 800 nm. Below 600 nm, there was no absorption. At 600 nm, the absorption steeply increased to a peak at approximately 700 nm (Fig. 6A) . In 0.005% ICG solution, there was a shift of the absorption maximum to 700 nm with intense absorption between 600 and 700 nm. Again, there was no absorption below 600 nm (Fig. 6B) . 

Discussion

After intravenous administration, ICG is not toxic in animals, and the median lethal dose (LD₅₀) ranges from 50 to 80 mg/kg in mice. ^{31 32 33 34} There are no equivalent data in humans. The maximum recommended dose for intravenous application in humans is 5 mg/kg. However, a low intravascular toxicity of ICG does not necessarily mean that ICG is nontoxic to intraocular structures. In the vascular system, ICG is rapidly bound to proteins and thereby excreted from the tissue. ³⁹ In the eye, there may be a different situation regarding the binding properties of ICG, the distribution of the dye, and the effect of ICG on intraocular tissue. In vitro, a 5-mg/mL (0.5%) ICG solution is toxic to cultured lens epithelial cells. ⁴⁰ Brief exposure of cultured human retinal pigment epithelial cells to 0.1% ICG solution results in decreased mitochondrial enzyme activity but does not appear to influence cellular morphology. ²⁰ There is a recent report ¹⁷ on morphologic and functional damage of ICG to the rat retina in vivo after vitrectomy and intravitreous ICG administration. Ten days after application, 0.25 mg/mL of ICG was found to cause significant morphologic damage to all layers of the retina. 

During surgery, injecting 0.2 mL 0.05% ICG into the fluid-filled globe with an assumed volume of the vitreous cavity of 4 mL, usually provides an equivalent dose of ICG of 0.25 mg/mL (0.0025%), which is half of the maximum recommended intravenous dose in humans and one tenth of the damaging dose in rat eyes. ¹⁷ However, the exposure times in the animal model and during surgery are essentially different, and conclusions regarding toxicity cannot be drawn from simple mathematics. 

Moreover, the intravitreous concentration of ICG may not reflect the real distribution of the dye within the eye, especially not at the vitreoretinal interface, where ICG may accumulate because of a possible rapid binding of the dye to proteins. In plasma, 80% of ICG is bound to globulins, probably α-1 lipoproteins. ³⁹ At the vitreoretinal interface, there are no data concerning the binding sites of ICG to the ILM. It has been reported, however, that apolipoproteins are secreted by Müller cells, effectively assembled into lipoprotein particles, and secreted into the vitreous. ⁴¹ Therefore, it can be assumed that lipoproteins and other proteins that are present at the ILM may account for binding of ICG to the ILM resulting in an accumulation of the dye at the vitreomacular interface. Persistent ICG fluorescence at the vitreoretinal interface has been reported 6 weeks after ICG administration in macular hole surgery. ⁴² The intense staining of the ILM and the clear contrast between the vitreous, the retina, and the ILM after intravitreous application of ICG further suggest an accumulation of ICG at the vitreoretinal interface, which may enhance the concentration of ICG at the inner retina beyond intravitreous values and theoretical expectations. ^{11 15}  

In the present experimental setting, application of ICG without illumination resulted in detachment of the ILM from the retina. Ultrastructural analysis disclosed plasma membranes of Müller cells and cellular debris adherent to the ILM, exactly as it has been observed after ICG guided removal of the ILM during surgery for macular holes, macular pucker, vitreomacular traction syndrome, and diffuse diabetic macular edema (Haritoglou C, manuscript submitted; and Haritoglou C, Gandorfer A, Gass CA, Kampik A, ARVO Abstract, 3516, 2002). ¹⁹ It should be noted that no attempt at peeling was made in the present study. ICG alone was sufficient to separate the ILM from the retina, reflecting the ease of membrane removal after staining with ICG in macular surgery. The cleavage plane, however, was not at the inner undulating aspect of the ILM—as in peeling of the ILM without the use of ICG—but within the innermost retinal layer (Haritoglou C, Gandorfer A, Gass CA, Kampik A, ARVO Abstract, 3516, 2002). ^{5 19}  

From the present experimental setting we cannot conclude what finally may have caused Müller cell damage. Osmolarity and pH of the ICG solution applied were 279 mOsM and 7.5, respectively (data not shown). It could be hypothesized, however, that accumulation of the dye at the ILM may have raised the osmolarity at the vitreomacular interface beyond critical limits. Marmor ⁴³ investigated the impact of hyperosmotic solutions on the retina. Within seconds to a minute, elevation and glistening of the vitreomacular interface occurred, finally resulting in retinal detachment. ⁴³ The weakest solutions that produced ophthalmoscopically visible changes to the retina were near 500 mOsM. Nonspecific shrinkage and disruption of normal cellular orientation were characteristics of osmotic cellular damage. ⁴⁴ Marmor et al. ⁴⁴ and Okinami et al. ⁴⁵ observed that severe osmotic stress may cause rupture of cells but does not split cells apart at their intercellular junctions. These findings and the observation of Marmor et al. ⁴⁴ of detachment of the vitreomacular interface after application of a hyperosmotic solution are consistent with the ultrastructural findings in the present study. Cellular swelling and formation of cysts near the cellular boundary, disruption of cells and fragmentation of the cell membrane with preserved intercellular junctions were characteristic features of osmotic damage in the series of Marmor et al. ⁴⁴ and were all found in the present study after application of ICG to the macula. Therefore, we assume an osmotic effect of ICG at the inner retina after accumulation of the dye at the vitreomacular interface, despite regular osmolarity of the ICG solution applied. This mechanism of action may account for recent reports on retinal toxicity of ICG, the ease of membrane removal during surgery, and the alteration of the cleavage plane reported previously. ^{17 19}  

However, most severe damage to the inner retina occurred after ICG staining of the ILM and illumination with wavelengths beyond 620 nm. Loss of the ILM, disruption of the nerve fiber layer, and gross cellular disorganization with fragmentation of the cytoplasm were found in all four ICG-stained eyes that were exposed to the near-infrared and infrared spectrum. Neither ICG alone, nor ICG followed by illumination of the posterior pole with wavelengths between 380 and 620 nm, nor illumination of the posterior pole without the use of ICG resulted in these ultrastructural findings, supporting evidence for a photodynamic effect of ICG at the vitreoretinal interface. 

ICG is a tricarbocyanine type of dye with infrared absorption properties and a peak absorption at approximately 800 nm in blood plasma. The light-absorption properties of ICG depend not only on the concentration but also on the solute. In water, ICG tends to aggregate at high concentrations, causing a shift of the absorption maximum from 800 to 700 nm. ³⁵ Moreover, the measurement of the absorption properties of ICG as administered in this experimental setting and during surgery disclose a shift of the absorption band starting at 600 nm and steeply increasing beyond. Regarding the irradiance of the light pipe of one randomly chosen vitrectomy instrument (Megatron; Geuder), there is still 28% of the total irradiance beyond 600 nm. These results and the presence of severe inner retinal damage in all ICG-stained eyes that have been exposed to the near infrared and infrared spectrum support experimental evidence for a photodynamic effect of ICG at the vitreoretinal interface. 

Regarding the mechanisms of action of photodynamic cytotoxicity, the molecule’s absorbed energy can be converted to heat and transferred to other molecules (photooxidation I), damaging cells by raising their intracellular temperature, as shown by the use of ICG in photocoagulation or tissue welding. ^{35 46 47 48 49} Alternatively, the photosensitizer’s energy can be transferred to molecular oxygen (photooxidation II), forming a triplet stage that interacts with oxygen and other molecules to generate reactive intermediates, such as singlet oxygen. ³⁷ Ultrastructural analysis of cultured human skin cells after ICG-mediated phototherapy revealed cytoplasmic vesiculation; dilatation of the rough endoplasmic reticulum, the Golgi complex, and the perinuclear cisternae; and chromatin condensation in the nucleus. ³⁶ A previous report demonstrated the toxic effect of ICG on rat liver mitochondria in vitro. ⁵⁰  

It is not surprising that we did not observe these ultrastructural features of intracellular damage. In living cells, uptake of ICG is a carrier-mediated and saturable transport process leading to intracytoplasmic damage. ³⁶ It is not known whether this active transport mechanism is compromised in postmortem cells and which role other mechanisms of action play, such as diffusion with subsequent accumulation of the dye at diffusion barriers. 

From the present experimental setting, we cannot conclude which type of photooxidation may have caused inner retinal damage. Given the assumption that the ILM may act as a diffusion barrier for a water-soluble molecule with a molecular weight of 775 kDa such as ICG, the ultrastructural findings in the present study are more likely to be caused by a thermal type I reaction of photoactivation rather than by a type II reaction of intracellular generation of singlet oxygen. Further evidence of this hypothesis, however, must be obtained from an animal model. 

Although an experimental setting does not reflect the situation during surgery, the present data are consistent with previous work and support further evidence that ICG staining of the ILM may cause retinal damage under certain still poorly understood circumstances. Amounts of ICG currently administered in macular surgery cause a shift in the absorption behavior of the dye toward 600 nm, which may result in a photodynamic effect of ICG at the inner retina after exposure to wavelengths beyond 600 nm. Accumulation of ICG at the ILM may exceed the concentration measured in the vitreous and may enhance the osmolarity of ICG at the retina beyond critical limits. Thus, further work is needed to define a safe surgical setting when the dye should be administered later in macular surgery. 

²  Contributed equally to the work and therefore should be regarded as equivalent senior authors.

 Supported by Friedrich-Baur-Shiftung, München, Germany.

 Submitted for publication June 5, 2002; accepted August 8, 2002.

 Disclosure: A. Gandorfer, None; C. Haritoglou, None; A. Gandorfer, None; A. Kampik, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Characteristics of Donors, Eyes, Treatment Modalities, and Ultrastructural Findings

Eyes (n)	Postmortem Time (h)	Gender	Age of Donor	ICG	Exposure to Light (nm)	Ultrastructural Findings
2	16	F	51	Yes	No	ILM detachment; rupture of Müller cells
2	26	M	49	Yes	380–620	ILM detachment; rupture of Müller cells
2	30	M	17	Yes	380–760	ILM loss; cellular disorganization; fragmentation of cytoplasm
2	16	F	47	Yes	620–760	ILM loss; cellular disorganization; fragmentation of cytoplasm
2	20	M	62	No	380–760	Intact retina

Figure 1.

 

Semithin sections of treated (left column) and untreated (right column) retinal specimens taken from the same eye each. (A) ICG only. (C) ICG and illumination with 380 to 620 nm. There is detachment of the ILM in (A) and (C). (E) ICG and illumination with 380 to 760 nm. (G) ICG and illumination with 620 to 760 nm. Note inner retinal damage in (E) and (G). (I) Illumination with 380 to 760 nm without ICG application revealed an intact retina. (B, D, F, H, J) Untreated control specimens showed a well-preserved cytoarchitecture with mild postmortem changes. Toluidine blue; magnification: (A, B, E, F, I, J) ×400; (C, D, G, H) ×250.

Figure 2.

 

Ultrastructure of ICG-treated eyes with and without exposure to wavelengths between 380 and 620 nm. (A, B, C) There was detachment of the ILM and disruption of Müller cells in all four eyes. Fragments of Müller cell membranes and cellular debris were adherent to the retinal side of the ILM. (D) Large cysts were present at the cell boundary (★). Note the intact ILM in all specimens. (E) Despite fragmentation of the basal cell membrane, the intercellular tight junctions were not ruptured (rectangle). (F) Control specimen from the untreated retina of the same eye showed an intact inner retina with mild postmortem changes. Magnification: (A, F) ×1800; (B, C, D) ×4800; (E) ×9600.

Figure 3.

 

Ultrastructure of ICG-treated eyes after illumination of the posterior pole with wavelengths beyond 620 nm. (A, B, C) There was gross disorganization of the cellular architecture of the inner retina with fragmentation of the cytoplasm and disorganization of cellular organelles. (D) Control specimen of the same eye showed an intact inner retina with mild postmortem changes. Magnification: (A, B) ×1800; (C) ×4800; (D) ×9500.

Figure 4.

 

(A) Illumination of the posterior pole with wavelengths of 380 to 760 nm did not result in ultrastructural alterations compared with the untreated control (B).

Figure 5.

 

(A) Irradiance of the light source of the vitrectomy instrument: 380 to 760 nm, with a maximum at 550 nm. (B) The near-infrared and infrared part of the spectrum beyond 620 nm was blocked. (C) Transmission of the near-infrared and infrared part of the spectrum beyond 615 nm.

Figure 6.

 

(A) Absorption properties of 0.0025% ICG solution representing an intravitreous injection of 0.2 mL of 0.05% ICG into the fluid-filled globe. There was a broad band of absorption starting at 600 nm. (B) With increasing concentration (0.005%), the absorption band between 600 and 720 nm exceeds the absorption peak of ICG in blood plasma at 800 nm.

The authors thank Monika Volkholz and Alex Feller for excellent technical assistance.