Cardiac Resynchronization Therapy: Extending Current Responses to Phrenic Nerve Stimulation

The evidence demonstrating improved outcomes with cardiac resynchronization therapy (CRT) in patients with significant left ventricular (LV) dysfunction and a prolonged QRS duration is sufficiently robust to support this therapy across a wide range of heart failure classifications.1-5 These outcomes have included 6-minute walk test, functional status, mortality, heart failure hospitalizations, and LV dimensions and ejection fraction.1-5 Clinical trials have uniformly demonstrated that these endpoints are most consistently achieved by LV epicardial pacing from a lateral branch of the coronary sinus.1-5 The phrenic nerve courses along the lateral margin of the LV on the parietal surface of the pericardium before innervating the left diaphragm. Based on the proximity of the epicardial LV lead to the phrenic nerve, and amplitude and duration of the pacing stimulus, phrenic nerve stimulation with diaphragm contractions may occur intermittently or continuously with pacing. Stimulation results in left diaphragmatic contractions that are generally intolerable to the patient. Anatomically the phrenic nerve is commonly in close proximity to the epicardial left obtuse marginal vein.6 Assessment of the relationship of LV epicardial pacing and phrenic nerve stimulation clearly indicates that stimulation is more common when LV pacing is performed from a mid lateral (14.2%), mid posterior (10%), or apical location (17.8%), when compared with other locations (2.8–8.3%).7 When a posterolateral branch was targeted, 33–40% of initial implant locations resulted in phrenic nerve stimulation.8 Multiple strategies have evolved to prevent or treat phrenic nerve stimulation during or subsequent to implantation of a CRT device. Optimally, phrenic nerve stimulation is prevented by meticulous testing once a lead has been positioned in an epicardial vein at implantation. This is commonly performed at maximal stimulus amplitude at a single pulse width during multiple deep respiratory cycles. Anesthetic agents used for patient paralysis must be withheld for an adequate period of time or reversed to assess for diaphragmatic capture. When symptomatic phrenic nerve stimulation is encountered post-procedure, it requires noninvasive programming that frequently, but not always, resolves the problem. Multiple noninvasive options may be available including programming to change the cathode and anode, so called electrical repositioning. Programming the amplitude or duration of the pacing stimulus is another noninvasive alternative. Programming the LV lead off and there by losing the potential benefits of CRT is a less desirable option. An invasive and least desirable option is repositioning the lead. This however exposes the patient to the considerable risks of a repeat procedure. Based on the limitations of all of these options, it is evident that every effort should be made to ensure absence of phrenic nerve stimulation at the time of implant. Multiple strategies have evolved to ensure lack of phrenic nerve stimulation at the time of implant. Early data suggested that LV pacing from posterlateral branches of the coronary sinus, or the areas of maximal electric or mechanical dyssynchrony resulted in maximal clinical benefit.9, 10 More recently, an analysis of data from the MADIT CRT study showed equivalent benefit of CRT from any of an anterior, posterior, or lateral location, with no benefit gained from an apical lead position.11, 12 If the more recent data are employed in the implant strategy, pacing sites less likely to produce phrenic nerve stimulation, such as anterolateral branches, could be used when phrenic nerve stimulation limits the use of more traditional sites. Any strategy to eliminate phrenic nerve stimulation other than movement of the lead to an entirely new major vein branch may not be effective after the implant procedure. Subtle post procedure lead movement back toward the main body of the coronary sinus, or out of small side braches of a large vein branch, can reposition the lead to a location where stimulation again becomes problematic. The potential impact of patient position cannot be assessed during implant. The absence of diaphragmatic stimulation at implant is not a guarantee that with alteration of the anatomic relationship of the lead to the phrenic nerve with change in body position will not result in phrenic nerve stimulation. Occasionally, very few options are available as target veins for lead placement, making the implanting physician more dependent on options other than repositioning to eliminate phrenic nerve stimulation. A major advance in avoidance of phrenic nerve stimulation in CRT is the addition of multiple pacing poles to LV leads with programmable polarity options in CRT devices. In the initial phase of CRT only unipolar leads were available with no options for reprogramming. In a study comparing unipolar LV leads to bipolar leads, all cases of phrenic nerve stimulation were resolved with reprogramming polarity in the bipolar lead group.13 By contrast, reoperation was necessary in all patients with unipolar leads.13 The introduction of a quadripolar lead has increased the likelihood of procedural success without phrenic nerve stimulation.14 The multiple options for programming pacing vectors with designation of the cathode and anode allows avoidance of phrenic nerve stimulation in the vast majority of patients. This results from the higher probability that one of the pacing vectors will be anatomically in a position to eliminate phrenic nerve stimulation.14 Quadripolar leads can be wedged distally in the vein with use of the proximal lead for pacing in a location with acceptable electrical parameters and absence of phrenic nerve stimulation. By contrast, a unipolar or bipolar lead might have to be placed more proximally in a vein to avoid phrenic nerve stimulation and be at increased risk of dislodgement.14 In a prospective study of a quadripolar lead, the postprocedure incidence of phrenic nerve stimulation was 14% when a standard bipolar configuration was used. This compared with no phrenic nerve stimulation in any patient when all pacing vectors were utilized.14 Although polarity change can be very useful to avoid phrenic nerve stimulation during or after the procedure, the appropriate choice of lead and device is needed to allow this intervention. Another technique of resolving phrenic nerve stimulation that can be attempted during implant is identifying a capture threshold margin between the phrenic nerve and the LV. If the capture threshold of the LV is lower than that of the phrenic nerve, it may be possible to successfully establish CRT with LV capture at device outputs below that needed to capture the phrenic nerve. Rarely, no other choice of vein exists, and a margin between LV and phrenic nerve capture must be accepted. Post procedure data suggest that a phrenic stimulation to LV pacing margin of >3 V results in no need for intervention in follow-up.9 This issue can be mitigated further by automatic capture algorithms that allow LV pacing far below the traditional twice threshold standard. In this issue of the Journal of Cardiac Electrophysiology, Hjortshoj and his colleagues, examine the relationship of capture of LV myocardium and the phrenic nerve, to changes in pacing voltage and pulse width.15 Eleven patients undergoing CRT implants with phrenic nerve stimulation observed were included in the study.15 During the implant procedure, LV and phrenic nerve capture thresholds were determined at varying pulse widths, from 0.3 to 2.9 milliseconds. In 9 of 11 patients, the LV threshold was reduced at a greater rate than the phrenic nerve threshold with increasing pulse width.15 The resulting strength duration curves demonstrated that the chronaxie of LV myocardium was higher than that of the phrenic nerve.14 In 3 patients, an unfavorable very small or zero margin was rendered higher to at least a factor of 2.15 The largest change between the 2 thresholds was observed up to a pulse width of 1.5 milliseconds.15 In this study of 11 patients there was no procedure detail provided beyond the testing of the LV lead nor was there clinical follow-up to determine whether changes in pulse width that avoided phrenic nerve stimulation during the procedure proved durable.15 Prior observations in 23 patients with phrenic nerve stimulation taken from a larger study of 197 patients, increasing the pulse width from 0.5 to 1.5 milliseconds did not significantly change the margin between LV and phrenic nerve threshold.13 This finding was observed in both a true bipolar and extended bipolar configurations.13 Nevertheless, these results do suggest that the LV threshold may decrease at a greater rate with pulse width increase than the phrenic nerve threshold. If proven in larger studies with longer follow-up, this approach may have clinical utility. The utility of the conclusions reached by this small study by Hjortshoj may be of value in a very specific situation.15 When no other choice of lead position or pacing polarity exists and the margin between the thresholds for LV and phrenic nerve stimulation is narrow, assessment of stimulation thresholds at a range of pulse widths may be useful. It is evident that the best approach is to prevent phrenic nerve stimulation intraoperatively. This is done with thoughtful decisions regarding the lead and its anatomic position. Outcomes are best with a multiploar lead and pacing vector that has no phrenic nerve stimulation at all. Alternatively, choice of a location with a wide phrenic nerve to LV margin might be considered. However, this approach using a wide pulse width could result in a considerable drain on the device battery. If the results of this study are further validated, increasing the pulse width may be a useful alternative to avoid phrenic nerve stimulation when no others options exist. Emerging techniques of LV endocardial pacing may eliminate the problem of phrenic nerve stimulation in the future,16 but they have not yet been demonstrated to provide a safe and durable long-term pacing site option for CRT. In the meantime, clinicians have a wide range of current options allowing them to respond to the challenge of preventing or treating phrenic nerve stimulation and thereby allow patients to benefit from CRT.