Tips on Improving Your Use of Endoillumination

Tips on Improving Your Use of Endoillumination

David R. Chow, MD

There has been a dramatic evolution in light sources for vitreoretinal surgery over the past decade. The preceding generation of light sources—namely, the halogen bulbs on the Alcon Accurus system and the metal halide bulbs on the Bausch + Lomb Millennium—though safe, lacked the power capabilities to drive the smaller fibers necessary for smaller-gauge vitrectomy, chandeliers and lighted instrumentation.

The newer xenon and mercury vapor light sources have significantly increased power output, allowing adequate illumination for these indications. This increased power capability also comes with attention to improved safety through the use of standard lower-wavelength filters and optional stronger filters. This article will review some practical aspects of endoillumination for the vitreoretinal surgeon.


For most surgical subspecialties, the key to illumination is brightness—the brighter the better. The problem for retinal surgeons is that, although we would like our surgical environment to be brighter, we face the risk of retinal photo-toxicity, a risk unique to ophthalmologists and highest for retina surgeons. So we must balance our desire for a brighter surgical field against the risk of creating retinal damage.

Many models of phototoxicity damage have been established and published to date. The work of Ham et al.1 ultimately led to the creation of what is referred to as an aphakic hazard curve. In that study, the authors determined the risk of creating phototoxicity in monkeys exposed to progressively lower wavelengths of light. They noted, as have many others, an increasing risk of phototoxicity with exposure to shorter and shorter wavelengths of blue and UV light.

To assess the safety of a given light source, its output is determined with a spectrophotometer, and any output that intersects with the aphakic hazard curve is summed together and multiplied by the risk factor associated with each wavelength (Figure 1).

Figure 1. Calculating the aphakic hazard sum: the aphakic hazard curve and the spectral output of a tested light source.

In 2002 and then again in 2010, I performed studies on all of the clinically available light sources. Calculation of the aphakic hazard sums revealed that the currently available mercury vapor light sources provided the highest safety calculations, followed by xenon light sources with a 435-nm filter and the halogen light sources, and then the xenon light sources with a 420-nm filter and the metal halide light sources.

What is important to note about comparing light sources based on their aphakic hazard sums, however, is that the safety calculations are valid only for comparing light sources at an equal brightness, working time and working distance. In other words, if all other settings are equal, then this calculation is a valid indication of the safest light source. In reality, however, the more clinically relevant calculation of safety is the retinal threshold time. This calculation incorporates not only the aphakic hazard sum of the light source but also the brightness of the light pipe, its working distance, and its numerical aperture (cone of illumination). It then determines the amount of time you could theoretically work under these conditions safely, assuming a cut off for damage of exposure to 25 J/cm2.

One should note that these threshold times are “theoretical,” based on the somewhat arbitrary usage of a damage threshold determined by the work of Ham et al. The times that are calculated sometimes seem surprisingly short and do not seem to correlate well with clinical experience. What is relevant about these threshold times is that they can be used comparatively to determine the relative changes in safety that occur as you vary the light source, working distance, brightness, etc. Here are some examples to show you the interplay of these variables in the retinal threshold time calculation.

• Effect of a light source with a lower aphakic hazard sum. The Alcon AHBI xenon light source with a 420-nm filter used at 8 lm and a working distance of 4 mm has a threshold time of 7:43 min. By switching to a halogen light source on the Accurus with the same settings, the threshold time increases to 10:07 min. The xenon light source on the Alcon Constellation with a 435-nm filter and the same settings has a threshold time of 12:25 min, while the mercury vapor light source on the Bausch+Lomb Stellaris PC and the Synergetics Photon 2 has a threshold time of 15:16 min. These calculations show the increased working time available by improving the safety of the light source.

Further proof of the increased safety from switching to a light source with a better aphakic hazard calculation: At a working distance of 4 mm, if you keep the threshold time constant at 7:43 min, you can use the Alcon AHBI xenon light at a brightness of 8 lm but can increase the brightness on the Alcon halogen light source to 12 lm and the Synergetics Photon 2 or Bausch+Lomb mercury vapor (Stellaris PC) light source to 16 lm.

• Effect of turning down the brightness. If you use the Constellation xenon light source at an output of 8 lm, threshold time is 12:25 min; if you increase the brightness to 10 lm the threshold time drops to 9:56 min, while at 12 lm it drops to 8:17 min. On the other hand, if you drop the power to 6 lm the threshold time increases to 16:34 min.
• Effect of doubling your working distance. If you increase your working distance from 4 mm to 8 mm with an 8 lm light source, the retinal threshold time increases on the Alcon Accurus halogen from 10:07 min to 33:48 min, on the Alcon Constellation xenon from 12:25 min to 41:29 min, and on the Bausch+Lomb Stellaris PC and Synergetics Photon 2 from 15:16 min to 50:59 min. This adjustment is by far the greatest increase in safety that can be obtained in the calculations. The nice thing about this variable is that it is totally surgeon-dependent—we as surgeons can exponentially increase our safety in any situation by just holding the light pipe further away from the retina!
• Effect of adding stronger filters to the light source. As outlined in the description of the aphakic hazard sum, the safety of a light source is a function of its output in the range of the aphakic hazard curve. Filters can be used to eliminate the output in the lower wavelengths, thus increasing the safety of the light source. The newest generation of light sources all feature stronger blue-light filters (at least 435 nm), which has resulted in improved safety calculations with these light sources.

The question then becomes: why not filter out even more blue light? There are some concepts to be explored in this direction. First, by filtering out more blue light you can increase the safety of any light source. This increased safety is reassuring, particularly if a surgeon desires to turn up the brightness at any point during a surgical procedure.

In 2004, I did a study looking at the effect of various filters on the Synergetics Photon 1, and I showed that yellow filters with a cutoff around 485 nm created a working environment that was visually pleasing and a safety calculation that was 10 times higher than the Photon 1 with its baseline filter. I showed that this increased safety could be translated into brightness by allowing you to turn up the power on the Photon 1 to 30 lm from 10 lm, while still maintaining a safety calculation that was 4.5 times greater than the baseline (Figure 2).

Figure 2. The value of a yellow filter for safety and brightness.

In the last few years, many of the newest light sources have incorporated filter options to give surgeons the opportunity to take advantage of this concept. Presently, the DORC Xenon White Star, Stellaris Xenon PC and the Synergetics Photon 1 and 2 all have filter options.

In a study I completed earlier this year, I showed that the filters on the DORC Xenon Bright Star resulted in a 20% increase in safety when going from the 420-nm filter setting to the 435-nm filter setting, with a fivefold further increase in safety when switching to 475 nm and another sevenfold increase when switching to 515 nm (a 33x increase from baseline). The Synergetics Photon 1 and 2 have an adapter with a 485-nm filter that improves the safety calculation by 26 and 30 times, respectively. The Stellaris PC features a yellow tint filter that increases safety by 16% and an amber filter that improves it by 118 times.

A second concept that can be explored with the use of filters is the effect they have on the color of the viewing environment. One of the trade-offs that occurs as you incorporate stronger and stronger baseline filters into a light source is that the color of the light changes, becoming increasingly yellow as you remove blue light. Customized filters can be created to allow you to have a viewing environment of virtually any color. This concept of colored light can be explored to see if there are certain colors that enhance our ability to see tissues in vitreoretinal surgery.

One of the concepts that has received attention in our field is the concept of “true white light.” Various companies have marketed this idea as being an advantage in microscope illumination and in the design of retinal light sources. The advantage of “true white light” to us as retina surgeons is the concept of color consistency, meaning that tissues will have their native, “true” color.

The problem with “true white light” is that to create a white light you need to include blue (and all the other colors of the spectrum), and by doing so, you reduce the safety of a given light source. The other problem with this notion is that, even if you shine a true white light onto the retina, the light we will see up at our eyepieces will not be white in the presence of any element of a nuclear sclerotic cataract or a UV blocker intraocular lens. These media opacities will, in fact, act as filters, turning the “true white” light yellow.

• Effect of using a chandelier. The retinal threshold times created with use of a chandelier are dramatically increased. This is the result of maximizing the working distance inside the eye. For example, in comparison to the Alcon Xenon Constellation light source, which has threshold time of 12:25 min (working distance = 4 mm, brightness = 8 lm), a Synergetics 25-g Awh Chandelier used at maximum illumination has a threshold time of 4 hr 17 min 39 sec.
• Is phototoxicity really a clinical issue or are these all just theoretical calculations? For years, surgeons using the halogen (Alcon Accurus) or metal halide (Bausch + Lomb) light sources at the maximum of 10 lm reported no cases of clinical phototoxicity. Phototoxicity seemed to become an irrelevant clinical topic to most retina surgeons, and attitudes toward safety became relaxed and cavalier.

Since the release of the newer xenon and mercury vapor light sources with significantly more power, there has been some misuse of the additional power available, and there are now many reports of phototoxicity over the last few years. The FDA has accumulated a series of over 16 photo-toxicity cases occurring with one of the xenon light sources in which higher powers were used in conjunction with a filter system that only blocked blue wavelengths up to 420 nm.2

This combination of a brighter light being used with a light source that did not more aggressively filter blue light created a potentially dangerous working environment. This spring, at the Retina Fellows' Forum in Chicago, we surveyed the graduating retina fellows in North America and found that 30% have seen at least one case of phototoxicity, and, alarmingly, 3% had seen more than four cases. The bottom line is, yes, we need to be careful to use the extra power that is available in the newer light sources for the right reasons and in the right situations (Table 1).

Table 1. Tips to Reduce Your Risk of Phototoxicity
• Increase your working distance
• Turn down your brightness
• Use a light source with a good (>435 nm) base filter
• Add a stronger filter to light source
• Use a chandelier


The brightness of a light source is measured by determining its output with a power meter and then modulating this output by our own photopic response curve (Figure 3). The corresponding output is measured in lumens. The important implications of this are that a light source can be made very powerful, but if its output does not intersect with our photopic response curve, then we will not “see” the light.

Figure 3. Perceived brightness of a light source: photopic response curve and its intersection with the spectral output of a tested light source.

The ideal light source with perfect “luminous efficacy” would have an output that exactly matched our photopic response curve. In this situation, all light and energy created would be translated into light we would “see.” This concept is also important to understand as customized filters are designed so that energy put into creating a particular color for viewing is not done at the expense of inadequate brightness, which ultimately would be more important.

In the studies I completed in 2010, data on the brightness of the new light sources showed that all of the new light sources have a tremendous increase in their power capabilities. This increase has been particularly important in 25-g vitrectomy, in which the previous halogen and metal halide light sources only allowed 4 lm of output. Many clinicians complained about inadequate illumination in this context, and fortunately this is no longer the case with any of the newer xenon or mercury vapor light sources, which have outputs in excess of 8 lm.

The other clinical context in which increased output has proved valuable is the creation of lighted instrumentation and chandeliers, both of which suffered with previous light sources from inadequate brightness delivered down the smaller fibers necessary for these clinical concepts.

One of the recurring themes that I found during the testing of the newer light sources were inconsistencies in output. There were many sources for this inconsistency. First, you needed to be aware that xenon lamps have a finite life span (around 400 hrs) and that the output tends to drop off through that life span. The greatest drop in performance occurs during the “burn-in” period of the bulb. To alleviate this difference in clinical performance, many manufacturers have completed the burn-in process on these bulbs before they are shipped. Despite this change, there will be a drop-in performance of your xenon lamp as a function of its duration of usage.

Second, you should be aware that when you turn on a xenon lamp, it takes about 10 min for the lamp to “warm up” and produce its full power output. Third, there are large inconsistencies in output on the same light source between different ports. During the testing, I found that the output varied by as much as 30% between ports on the same machine fed by the same lamp.

Fourth, the output on the same port varied by as much as 30% in some light sources and by at least 15% in others, depending on the way you attach the connecter. The clinical pearl from this, of course, is that if you ever feel you are not getting as much light out of your light pipe as you are used to, have your staff wiggle the connector around, and you will see a difference.

How much brightness do we need? The current generation of retina surgeons all trained on light sources that had a maximum capability of around 10 lm of light. This fact was not based on any science but on a basic clinical observation that this seemed to be enough. Because this output seemed adequate for surgeons' needs and was certainly safe given the large clinical experience, there was largely no drive toward providing more brightness.

It was only once we moved into the 25-g domain that the output of our previous light sources was found to be inadequate. The question still remains then: What is the optimal amount of illumination for retinal surgery? Presently we do not know the answer to this question, but in my opinion, the answer varies depending on the situation. For instance, an older surgeon operating with an older microscope, which has a 50/50 beam splitter and active laser filter, on a patient with 3+ early nuclear sclerosis and a dense vitreous hemorrhage will need much more light than a younger surgeon operating with a newer microscope with no beam splitter or laser filter on a pseudophake with an epiretinal membrane. Not only will the ideal amount of illumination depend on the clinical issues, but it will depend on what level of filtering is being used on the light source. As previously discussed, if a strong yellow filter is used, brightness levels will be able to be turned up significantly without any safety concerns.


In my experience, one of the greatest advantages of the newer light sources has been the creation of clinically useful lighted instruments and chandeliers. These have essentially freed up a hand for me during surgery, allowing true bimanual surgery. I routinely use a chandelier for retinal detachment repairs and diabetic dissections. In both of these scenarios, I am able to depress the periphery for myself, performing the vitreous base dissection with more control and confidence that I am not missing any small peripheral breaks.

For diabetic dissections, the second hand allows me to dissect membranes freely, with a forceps in one hand and the cutter in the other. This capability has made even the most complicated cases “easier” in my hands. When inserting the chandelier, you should pay attention to be sure that the angle of entry matches the angle at which you would like the chandelier fiber to be oriented. Also, secure the fiber to the drape in a manner that maintains optimal illumination. When using a single-fiber chandelier, do not worry if the illumination does not fill the whole eye. Just position the chandelier to allow you to work on 180º, and then reposition the fiber when necessary to complete the other side of the eye.

The potential advantages of two fiber chandeliers are that they eliminate the need for repositioning of the fiber and will eliminate the shadowing sometimes seen with single-fiber chandeliers. This is done, of course, at the expense of an extra sclerotomy.

Shadowing with single-fiber chandeliers is somewhat irrelevant to the surgical procedure, except at the 12 o'clock position, at which the shadow would actually be in the desired working zone. I have found that this effect can still be managed clinically by adjusting the angle of the fiber.

Another concern with chandeliers is the glare encountered when going to air. This concern is genuine and can often be quite difficult. Tips to manage this include changing the angle of the fiber, retracting the fiber back into the metal sleeve (certain designs), or using the standard pack-supplied light pipe for this part of the procedure.


Despite the significant progress that has been made in the last decade in endoillumination, there are many frontiers that we only now are beginning to explore. Newer LED light sources with unique capabilities are entering our field, and the clinical utility of filters for safety, increased brightness, and tissue enhancement are evolving concepts. The “ideal” viewing environments and brightness levels need to be better identified through scientific work. RP


1. Ham WT, Ruffolo JJ, Mueller HA, et al. Histologic analysis of photochemical lesions produced in rhesus retina by short-wavelength light. Invest Opthalnol Vis Sci. 1978;17:1029-1035.
2. MAUDE—Manufacturer and user facility device experience. US Food and Drug Administration Web site.,Data_Date_Year:2009,Data_Date_Year:2008,Data_Date_Year:2007,Data_Date_Year:2006,Data_Date_Year:2005&pn=10&sc. Accessed April 1, 2011.

David R. Chow, MD, practices at the Toronto Retina Institute and is an assistant professor at the University of Toronto. He is a consultant to Synergetics, Bausch+Lomb and Arctic Dx. He can be reached at