Ultrasound-Induced Thermal Incision Contracture

36 Ultrasound-Induced Thermal Incision Contracture

Randall J. Olson

An incision burn is a complication that is infrequent but can be devastating. I have seen video examples of severe wound burns that happened within three seconds of the introduction of ultrasound, and yet resulted in such profound scarring that, even with a partial transplant in the area of the burn, vision did not return to 20/20 even a year after the surgery. Because we generally look at the phacoemulsification tip during surgery, our first indication of an incision burn are the corneal folds from the thermal contraction that have made their way into the central cornea, obscure our view, and make us realize we have a major problem on our hands (Fig. 36.1).

Studying the Problem

Our in vitro work over the years points to a temperature of 60°C at the incision site as the level at which a second or two is enough for a wound burn to start to form. It is important to understand that this is not really a burn or an oxidative change as such, but rather is what happens to collagen once it reaches ∼ 60°C; there is a contracture of the collagen, which has been used in the past to contract the cone in keratoconus prior to keratoplasty. This contracture starts as a focal change and extends, if the heat source is maintained, to the point of substantial blurring of the central cornea as well as induction of huge amounts of astigmatism. One case referred to me had over 20 diopters of induced irregular astigmatism! So clearly it is important to understand this process and strive to do all we can to avoid it.

I first became very interested in this problem when claims were being made about different thermal properties of phaco machines that simply did not make sense. What was claimed did not follow the laws of thermodynamics, and so I became curious about understanding the thermal properties of phacoemulsification instruments. Everyone agreed that wound burns were due to ultrasound-induced heat change in the incision. With an ultrasound needle having a peak velocity of about 50 miles per hour, even though the length it travels is very short, friction is generated very rapidly. To demonstrate, lightly squeeze a vibrating phaco needle between your fingertips to realize how fast it can get hot.

Our first major publication on this subject showed that the percentage of power was not a constant that you could use to compare one machine to another¹ (Fig. 36.1). Looking at the three common machines available at the time, and performing the experiment in saline solution (some earlier work had even been done in air, which made no sense at all), we found that Legacy (Alcon, Fort Worth, TX) at 100% power produced the same heat that Sovereign (Abbott Medical Optics [AMO], Abbott Park, IL)did at 43% power and Millennium (Bausch and Lomb, Rochester, NY) did at 36% power. Therefore, any comparisons without understanding this difference made very little sense.

Additional work demonstrated that the machines also behaved quite differently when there was a load in place² (Table 36.1). Just going from air to water dramatically changes the load; however, we further used weight on the sleeve, thereby creating increased friction on the tip, just as a tight incision would do, or in trying to phacoemulsify a hard nuclear fragment, and here the clear outlier was Legacy, which had an eightfold increase in its overall heat production over that of Sovereign and Millennium. In fact, with a load, the total heat generated by Legacy exceeded at the same power setting what Sovereign and Millennium created. Thus, it became clear that a stroke-protected machine such as Legacy acts like a cruise control in a car such that if you set it for 50 miles per hour, you will have very little energy needed to go downhill and a lot of energy to go uphill. In contrast, Millennium and Sovereign and most machines today are more like a gas pedal—you push the pedal a certain distance down and then the energy created is the same so you would race downhill and then go more slowly uphill. We surmised that this potentially was a wound burn risk for Legacy, because the surge in power would occur without any movement of the foot peddle and would occur at the riskiest time—when there was a lot of friction on the phaco needle or a hard nuclear fragment was being removed, often associated with occlusion of flow through the needle tip.

All of this early work also showed that the riskiest step we take is when we block fluid flow through the tip and engage long runs of ultrasound, such as when we have impaled and are removing a nuclear fragment. Heat buildup to near 60°C was not possible without blocking flow, but we could rapidly meet and exceed this temperature with the loss of the cooling effect of fluid flow. So we surmised that long runs of ultrasound with a hard nuclear fragment that also blocks the flow of fluid is the main cause of incision burns.

More recently, we have seen entirely new ultrasound modalities in the marketplace where the needle tip subtends an arc (OZil [Alcon] or torsional phaco) or an ellipse (Ellips [AMO] or transversal phaco). Claims were made that these types of tip action essentially prevented the risk of wound burn, and all studies to date show a dramatic decrease in frictional heat generation. So is that the end of the story and the end of wound burns?

Looking at this in more detail, we measured the heat generated, especially looking at metal stress heat generated from the wagging motion of the tip³ (Table 36.2). We were able to document that there was heat generated in the direction of the tip motion, and then compared it to the same amount of power with the same machine with continuous as well as micropulsed (6 ms on and 12 ms off) longitudinal ultrasound. We found, as did others, that continuous longitudinal ultrasound created three times or more heat than torsional, but micropulsed ultrasound created less heat, being on only a third of the time when compared with continuous longitudinal ultrasound. Repeating this experiment with fluid flow blocked showed that torsional heat could rise to levels where wound burns were possible. Clearly, therefore, blockage is a key concern, and even though the horizontal motion associated with torsional creates about one third the heat of continuous longitudinal ultrasound, if it were used consistently at a very high power, and longitudinal ultrasound were used at a much lower power setting, which we felt was the more likely clinical scenario, then there may not be any wound burn protection with torsional ultrasound as it was currently being used.

Fig. 36.1 A wound burn a year after the procedure. (Courtesy of Jorge Alio, Professor and Chairman of Ophthalmology, Universidad Miguel Hernandez, Alicante, Spain.)

Another key element had to be the efficiency with which we are using ultrasound energy. Some surgical approaches use mainly mechanical energy in which no heat is produced and very little ultrasound energy is used. We suggested that such approaches should be protective of wound burn. Furthermore, we have had a chance to study in some detail just what makes a phaco machine more efficient in the cataract removal process.⁴,⁵ From this particular combination of work, we have been able to discover many variables in association with efficiency—all of which suggests that how the machines operate, the parameters chosen, and the cataract surgical approach are all potentially important factors in wound burn creation.

Another important clinical consideration gleaned from studies using thermography of the incision site during surgery showed that temperatures close to the “cliff” of wound burn (60°C) are not uncommon, and there is no clinical clue that we are nearing the cliff. So we are blindfolded and wandering near a cliff that we can only appreciate once we are actually falling off (a wound burn occurs). Using this cliff analogy and without any simple means of removing this blindfold, it only makes sense to do all we can to stay as far from this cliff as possible. We can’t fall off a cliff if we never get close!

An additional important finding had to do with the role of viscoelastics in the creation of wound burns. Videos have documented that such burns can happen in a matter of seconds after applying ultrasound! A possible explanation is that the viscoelastic could block fluid flow through the needle, but the rapidity with which these wound burns have occurred simply cannot be explained by just fluid blockage. This was a mystery that had to be unraveled. So we studied different viscoelastics in an artificial chamber and compared them to the heat generated with saline solution with flow through the needle blocked.⁶ Incredibly, we found that all viscoelastics were variably exothermic when ultrasound is applied when compared with saline solution. These heat generation ratios when compared with that of saline were on the low side, with Healon GV (AMO) 2.4 to 7.1 times that of Viscoat (Alcon) (Table 36.3). This would mean that when you combine both this exothermic factor with the loss of fluid flow, you could reach wound burn temperatures in a matter of a couple of seconds, which is exactly what we have seen clinically. So clearly, viscoelastics would also appear to be a potential major risk for incision burn creation.

All of this work was theoretical and did not pinpoint the actual clinical factors related to, or the incidence of, wound burn. We had too many potential variables to ever discern the clinical basis of wound burns in prospective clinical studies, and with no clue as to the incidence, it would be impossible to power such a study. So we felt our best next step was a clinical survey. Surveys looking at absolute numbers of cases is always problematic, as they are based on a guess as to the number of cases; however, we were confident that surgeons would not forget a wound burn that they had in their clinical practice, so we felt that comparative ratios would be valid. Our first survey (Table 36.4) reported on 75 wound burns from a projected total of 75,000 surgeries, for an incidence of 1 in 1,000.⁷ We were able to determine that the vast majority of them (53) occurred during fragment removal, which supports exactly what we felt was the riskiest moment, and that is when the port of the needle is occluded and there is no cooling going on inside of the needle.