Thermal Latency in Radiofrequency Ablation
Background Progression of unintentionally induced atrioventricular delay is occasionally observed directly after termination of radiofrequency delivery in the vicinity of the atrioventricular node. We postulated that the application of a radiofrequency pulse may result in a tissue temperature rise that continues after the pulse.
Methods and Results Using the thigh muscle preparation, 5-, 10-, 20-, and 30-second pulses were applied at 30 to 40 W via a standard 4-mm tip electrode with 10-g contact pressure. Forty-one undisturbed pulses were delivered while recording intramural temperatures at 2-, 4-, and 7-mm depth. Maximal “thermal latency” was observed with the shortest pulse duration and at greatest depth. With 5-second applications, tissue temperature at 7-mm depth peaked 11.6 seconds after termination of radiofrequency delivery and stayed above end-of-pulse value as long as 34.5 seconds after the pulse. The additional rise in tissue temperature was 2.9°C. If only recordings within the lesion border zone were considered, the duration of latency was maximal with 10-second pulses: an additional gain in tissue temperature of 3.4°C was observed 6.4 seconds after the pulse while tissue temperature stayed above end-of-pulse value during 18.3 seconds.
Conclusions With relatively short applications, tissue temperature continues to rise after termination of radiofrequency delivery. This “thermal latency” may result in lesion growth after the pulse and may so explain the incidentally observed progression of conduction block after short pulses in the vicinity of the atrioventricular node. It also may explain the apparent discrepancy between lesion growth rate and intramural temperature rise studies.
During ablation in the vicinity of the atrioventricular node, unintentional lengthening of atrioventricular nodal delay is occasionally observed. Despite immediate termination of radiofrequency (RF) delivery, however, atrioventricular delay often progresses after the pulse, sometimes up to complete, usually transient, block (Fig 1⇓). This could be explained by a delayed effect of heat on atrioventricular nodal conduction properties. However, we hypothesized that under certain conditions, tissue temperature may continue to rise after termination of RF delivery.
During RF application, a very thin rim of tissue in contact with the ablation electrode is heated resistively while surrounding tissue is heated by radial conduction of heat predominantly.1 2 3 4 5 6 Termination of RF delivery will immediately eliminate resistive heating of tissue in contact with the ablation electrode. The temperature of that tissue, however, may temporarily remain higher than surrounding tissue, and radial heat flow may continue.
In the present study, we investigated the hypothesis that RF pulses can result in an intramural temperature rise that continues after termination of RF delivery.
The experimental protocol was approved by the University of Oklahoma Institutional Animal Care and Use Committee. Three mongrel dogs weighing between 20 and 24 kg were used in this study. The animals were anesthetized with sodium pentobarbital (25 mg/kg) and mechanically ventilated with room air. General anesthesia was maintained with supplementary doses of sodium pentobarbital. The thigh muscles were exposed and skin cradles created as previously described.7 A standard 7F deflectable quadripolar electrode catheter (Cordis Webster, Inc) with a 4-mm-long tip was used in this study. The electrode also contained a thermocouple to allow measurement of tip temperature. The electrode was positioned perpendicular to the tissue, with 10-g contact pressure. RF current was delivered between the catheter tip and an adhesive electrosurgical dispersive pad applied to the shaved skin of the opposite thigh. Tissue temperatures were measured with fluoroptic probes (Luxtron Inc, model 3000-4): four fibers were bundled together with one touching the tissue surface and the other three positioned at approximately 2-, 4-, and 7-mm, depth, directly adjacent to the ablation electrode.7 During RF application, root mean square voltage, current, and impedance were monitored and recorded on optical disk (Bard LabSystem), along with the temperatures measured from the fluoroptic tissue probes and the ablation electrode.
After the ablation electrode was positioned and the fluoroptic probes were inserted, the cradle was filled with blood and circulation in the cradle was started as previously described.7 RF energy was randomly applied for 5, 10, 20, or 30 seconds at various sites on the thigh muscle. RF energy was supplied by a 500-kHz voltage source generator (ACAC). Depending on the observed ablation impedance, RF voltage was set at values between 45 and 60 V to obtain an RF power level between 30 and 40 W. Distances between successive ablation sites were chosen to avoid lesion overlap. RF ablation was interrupted in the event of a sudden rise in impedance >10Ω or audible “pops.”7 8 After each RF pulse, the cradle was depleted of blood and the ablation site was examined for catheter tip dislocation and coagulum formation. After approximately 10 pulses, the skin incision was closed, and the dog was turned onto its other side. The procedure was then repeated on the other thigh muscle.
RF voltage, average impedance, pulse duration, tip electrode, and intramural temperatures were analyzed by means of the Bard LabSystem. From each intramural temperature recording, the end-of-pulse temperature was measured together with the following three characteristics of local thermal latency (Fig 2⇓): (1) timing of maximum temperature, (2) time between termination of the pulse and point at which temperature returned to end-of-pulse value, and (3) additional rise in temperature after the pulse.
A hypothetical local tissue temperature rise after termination of RF delivery is only clinically relevant if it can create additional transient or permanent damage.9 10 We therefore separately analyzed those temperature curves with an end-of-pulse and/or maximum tissue temperature between 45°C and 55°C.
The effects of pulse duration, depth, and end-of-pulse temperature on the three above-mentioned characteristics of thermal latency were analyzed by two-way ANOVA and multiple regression analysis with pulse duration, depth, and end-of-pulse temperature as covariates. A value of P<.05 was considered significant.
In the three dogs, a total of 59 RF pulses were delivered to the thigh muscle preparations. During 18 of these, sudden discontinuities in impedance and temperature curves were observed, usually in combination with an audible “pop” or coagulum formation. This mainly occurred at longer pulse durations. Visual inspection of ablation sites with “pops” after depletion of the cradle frequently revealed some electrode displacement, probably the result of the small vapor explosion responsible for the “pop” sound.7 8 An intramural temperature above 100°C was only observed in 3 of these 18 pulses but also in 3 of the 41 undisturbed pulses. The shape of the surface (0 mm) and 3 intramural temperature curves of these 41 undisturbed pulses was quantified by measuring the above-mentioned characteristics. Temperature at the tissue surface stayed below 45°C during 20 of the 41 pulses.
As illustrated in Fig 2⇑, tissue temperature gradually rises during RF delivery, as expected. This rise, however, continues after termination of RF delivery. The shape of these curves was affected by pulse duration and recording depth. Both the timing of maximum temperature and duration until return to end-of-pulse temperature were reached significantly later (P<.001) with shorter pulse durations and at greater depth, with a significant interaction (P<.01) between the latter two parameters. The magnitude of the additional rise in temperature after the pulse decreased with increasing pulse duration (P<.001) but was unaffected by depth (P=.8).
Multiple regression analysis revealed that of the three characteristics of thermal latency, only the duration until return to end-of-pulse temperature was significantly affected by the end-of-pulse temperature with a shorter duration at higher temperatures. Thermal latency was absent in all tip temperature recordings; tip temperature invariably fell instantly at termination of RF delivery.
In 31 of the 123 recordings, the end-of-pulse or maximum temperature reached a value between 45°C and 55°C (Table 2⇓). The average depth at which these curves were recorded increased with increasing pulse duration (P<.001) as the result of growth of the lesion during RF application. The magnitude of the three characteristics of thermal latency, listed at various pulse durations in Table 2⇓, therefore reflects the direct effect of pulse duration but also the simultaneously affected depth of the lesion border zone. In these tracings, the intervals from end-of-pulse to maximum temperature and the return to end-of-pulse temperature are maximal with a 10-second pulse duration, 6.4 and 18.3 seconds, respectively, with an average additional rise of 3.4°C.
This study demonstrates an important thermal latency between RF delivery and tissue temperature. While tip temperature drops instantly at termination of the pulse, tissue temperatures may continue to rise, especially after short pulses. This phenomenon can be explained by the coexistence of two different heating mechanisms: rapid resistive heating of the thin rim of tissue surrounding the ablation electrode and slower conductive heating of tissue at greater distance.1 2 3 4 5 6 The different response times of these two processes may result in a large temperature gradient near the ablation electrode at the beginning of the pulse and a surplus of heat that will continue to flow to surrounding tissue after the pulse. While thermal latency is still present after 30-second pulses, the clinical significance of this additional temperature rise may be reduced by its decreasing magnitude with longer pulse durations.
Because of the growth of the lesion with increasing pulse duration, the thermal latency in the border zone of the lesion (Table 2⇓) reflects the combined effect of pulse duration and depth. With a duration of latency that decreases with increasing pulse duration and increases with increasing depth, the decrease in latency with increasing pulse duration in the border zone is less dramatic than when all recordings (Table 1⇓) are considered. This explains why the duration of latency in the lesion border zone after 5- and 20-second pulses is similar (Table 2⇓). The additional rise in temperature, however, still decreases with increasing pulse duration (P<.001) because this latter parameter only depends on pulse duration.
The observed thermal latency may explain a progression of conduction block directly after short RF pulses in the vicinity of the atrioventricular node but will also play a role in all other radiofrequency catheter ablation procedures. Premature interruption of RF application in the vicinity of the atrioventricular node caused by unintentional lengthening of atrioventricular conduction delay may be one of the few occasions where thermal latency becomes clinically manifest.
A rise in tissue temperature after short RF exposures may explain the apparent discrepancy between tissue temperature rise and lesion growth studies.6 11 While intramural temperatures continue to rise during RF delivery even after 60 seconds,1 3 4 7 12 lesion size appears to mature within 20 to 30 seconds.1 13 14 15 The main difference between these two types of study is that lesion size is measured some time after delivery of RF energy, whereas tissue temperatures are recorded during RF delivery. Potentially ablative effects of a tissue temperature rise after termination of RF delivery are thus not accounted for in the latter studies. If we arbitrarily define the duration of an RF “pulse” in the border zone as start-of-pulse until return of tissue temperature to end-of-pulse value, a 5-, 10-, 20-, and 30-second pulse is lengthened to approximately 16, 28, 30, and 35 seconds (Table 2⇑). The difference in lesion size between a 10- and 20-second pulse is small, probably not because the lesion does not grow in that interval but because the difference in “thermal” pulse duration is much less than 10 seconds: it is only a 2-second lengthening of a 28-second pulse if we use the above-mentioned definition of “pulse” duration. In other words, the time scale of lesion growth during RF application measured at necropsy1 13 14 15 is transformed by the tissue temperature rise directly after RF delivery. Lesion growth curves should therefore be interpreted such that lesions listed at, for example, 10 seconds, do not reach their size at 10 seconds but only after a full 10-second pulse, including the thermal latency effect after the pulse.
Eighteen of the 59 pulses were disturbed by a sudden temperature discontinuity, usually in combination with an audible “pop,” suggesting intramural boiling.7 8 These discontinuities disabled a sensible analysis of the curves, which were therefore excluded from analysis. In our series, these discontinuities only occurred during RF application and never thereafter. This is not in disagreement with the observed thermal latency effect. The presence of thermal latency was postulated within the conductively heated zone around the hottest, resistively heated zone in contact with the ablation electrode. Within the latter area, elimination of resistive heating at termination of RF delivery will immediately result in a drop in temperature. Moreover, a transition of intramural temperature from below 100°C during the pulse to above 100°C after the pulse was not observed in our series. This may explain the absence of “pops” after RF applications.
An intramural temperature above 100°C was only observed in 3 of these 18 pulses. This may be due to a practical limitation of our tissue temperature recording technique. Tissue surface temperatures were measured as close as possible to but not at the electrode-tissue contact site while the first intramural temperature was recorded at 2-mm depth. Higher intramural temperatures closer to the tissue surface would thus have been missed.
Limitations of the Study
This study was performed in the thigh muscle preparation with standardized tip orientation and pressure. Differences between this experimental setup and the clinical situation may affect the precise characteristics of thermal latency.7
The repetition rate of the Luxtron model 3000-4 resulted in a temperature read-out every 0.4 seconds. This resolution may have affected the accuracy of our time measurements. During analysis, however, we have attempted to compensate for this limited resolution by interpolation of the stepped tissue temperature tracings.
The observation that the temperatures recorded with the fiber at the tissue surface stayed relatively low (<45°C) during 20 of the 41 pulses suggests that on average, the fluoroptic probes were positioned at the correct depth within the tissue. Nevertheless, the depth of the probes may have varied slightly between different ablation sites, which may have affected the data listed in Table 1⇑. The precise position of the probes, however, is not relevant for the clinically more important duration of latency as listed in Table 2⇑. With the above-mentioned temperature range criterion, we investigated the thermal latency effect in the border zone of the lesion irrespective of the actual depth of that border.
The RF generator used in this study was voltage controlled, whereas most clinically used generators are power regulated. When initial power is set to the same value, the only difference between both types of generator is their response to a change in impedance during RF application. With a power-controlled generator, a 10% drop in impedance will result in an approximately 5% decrease in RF voltage and a 5% increase in RF current. With a voltage controlled generator, this will lead to a 10% increase in RF current and total power. It is unlikely that this difference will lead to a different thermal latency effect after the pulse because it is the total amount of heat generated during the pulse and its delayed conduction into surrounding tissue that is responsible for the thermal latency effect.
With short (5 to 30 seconds) RF applications, tissue temperature continues to rise after termination of RF delivery. This “thermal latency” may result in lesion growth after termination of RF delivery and may so explain the apparent discrepancy between lesion growth and tissue temperature rise studies. Thermal latency may explain incidentally observed phenomena occurring shortly after RF applications such as progression of conduction block after short pulses in the vicinity of the atrioventricular node.
This study was supported by a grant (R01-HL39670) from the National Institutes of Health and a grant (HRI-104) from the Oklahoma Center for the Advancement of Science and Technology.
- Received October 25, 1995.
- Revision received December 27, 1995.
- Accepted January 2, 1996.
- Copyright © 1996 by American Heart Association
Kalbfleisch SJ, Langberg JJ. Catheter ablation with radiofrequency energy: biophysical aspects and clinical applications. J Cardiovasc Electrophysiol. 1992;3:173-186.
Wittkampf FHM. Temperature response in radiofrequency catheter ablation. Circulation. 1992;86:1648-1650.
Nakagawa H, Yamanashi WS, Pitha JV, Arruda M, Wang X, Ohtomo K, Beckman KJ, McClelland JH, Lazzara R, Jackman WM. Comparison of in vivo tissue temperature profile and lesion geometry for radiofrequency ablation with saline irrigated electrode versus temperature control in a canine thigh muscle preparation. Circulation. 1995;91:2264-2273.
Avitall B, Morgan M, Hare J, Kahn M, Lessila C. Intracardiac explosions during radiofrequency ablations: histopathology in the acute and chronic dog model. Circulation. 1992;86(suppl I):I-191. Abstract.
Nath S, Lynch C, Whayne JG, Haines DE. Cellular electrophysiological effects of hyperthermia on isolated guinea pig papillary muscle. Circulation. 1993;88:1826-1831.
Wittkampf FHM, Simmers TA, Velema E, Hauer RN. Discrepancy between lesion growth and temperature rise in radiofrequency catheter ablation. Circulation. 1992;86(suppl I):I-192. Abstract.
Wittkampf FHM, Hauer RNW, Robles de Medina EO. Control of radiofrequency lesion size by power regulation. Circulation. 1989;80:962-968.
Haines DE. Determinants of lesion size during radiofrequency catheter ablation: the role of electrode-tissue contact pressure and duration of energy delivery. J Cardiovasc Electrophysiol. 1991;2:509-515.