Tracing the Right Phrenic Nerve – A Systematic Review and Meta-Analysis

Kuniewicz M12, Mazur M1, Karkowski G2, Budnicka K1, Przybycień W1, Walocha J1, Lelakowski J2

1Department of Anatomy, Jagiellonian University Medical College, Krakow, Poland.2Department of Electrocardiology, Institute of Cardiology, John Paul II Hospital, Jagiellonian University Medical College, Krakow, Poland..



The Right phrenic nerve (RPN) is vulnerable to injury during the isolation of the right pulmonary veins (RPV). The study aimed to provide a comprehensive meta-analysis of the overall prevalence of right phrenic nerve injury (RPNI), its course and its association with the superior and inferior pulmonary veins.


Through December 2017, a database search was performed on PubMed, Science Direct, EMBASE, SciELO, and Web of Science. The references were also extensively searched in the included articles.


Detection of the RPN may vary according to the identification method. It ranges from 100% in postmortem studies, 93% in intraoperative, to 57.88% in computer tomography (CT) imaging. Based on the included studies (n-507), the distance from the right superior pulmonary vein (RSPV) ostium to the RPN was 12.48mm (±6.21). In postmortem studies, the distance was 6.92mm (±3.94); in pre or intraoperative techniques, 13.32mm (±5.96) if noninvasive, 13.97mm (±7.8) if invasive. Distances ranged from 0DC342.6 mm. For the right inferior pulmonary vein (RIPV) (n-125) the mean distance was 16.53mm (±8.92) with distances from 0.4 68mm. The risk of RPNI with distance-included studies was 12.46% (47 RPNI in 377 cases). In the meta-analysis, the distance from the RSPV to the RPN that was associated with an increased risk of RPNI was 7.36mm.


RPNI is a relatively rare complication. A firm understanding of its course, relation to the PV ostium, and detection are vital for preventing future injuries and complications.

Key Words : RPN, RPNI, PNP, Pulmonary vein ablation.

AuthorMarcin Kuniewicz,Department of Anatomy, Jagiellonian University Medical College, Krakow, Poland


Catheter ablation has become a first-line treatment for the prevention of recurrences of paroxysmal atrial fibrillation after discovering myocardial sleeves activity responsible for arrhythmia, 1 and is the treatment gold standard since 2012 2. Unfortunately, the anatomy of right-sided pulmonary veins (RPV) and its close relation to the right phrenic nerve (RPN) makes it susceptible to injury during pulmonary vein isolation (PVI), which was initially reported in 2004 3. Previously, right phrenic nerve injuries (RPNI) during PVI RF ablations were a rare complication, estimated around 0.48% 4 while the first significant cryoballoon (CB) observation from 2008 shown a risk level of 7.5% 5 even reaching as high as 24.4%6 in 2014. That is why a RPNI was called “another Achilles’ heel for AF ablation” 7. Phrenic nerve injury may be transient or persistent. Transient, where the RPNI resolves at the end of the procedure, whereas a persistent injury persist beyond the procedure 8. The injury is most commonly transient, and all the clinical symptoms do not occur after postprocedural hospital discharge. In recent publications, the longest period for phrenic nerve paralysis (PNP) recovery was 314 days, while the average time to recovery was 3-6months 9,5. Through the years of practice and CB generations, various techniques were undertaken to decrease the risk of RPNI. From compound motor action potential (CMAP) in 2011 11, to immediate balloon deflation (IBD) proposed by Gosh in 2013, temperature controlled application proved by Mugnai and Kuhne 12, 13, the pull back technique proposed by Okshige 2018 33 to anatomical assessment of the left atrium and adjacent structures14. Larger RPV dimensions, early branching patterns originating from the main ostium, and shorter distances from RPV to SVC are associated with RPNI both at right superior pulmonary vein (RSPV) and right inferior pulmonary vein (RIPV) 14. For cryoballon ablation (CBA), the RSPV – RPN distance measured from the ostium was found to have the most accurate correlation in predicting RPNI. Schmidt pointed out in 2008 that the general recommendation for a minimum distance of the phrenic nerve to the RSPV ostium cannot be provided 22. Horton et al.21 suggested that a PN location within 10 mm of the RSPV poses a higher risk of RPNI when using a balloon-based ablation system. Canpolat 20 demonstrated that after their analysis estimate, the incidence of RPNI was increased when the distance of the RSPV ostium–RPA distance was 8.0 mm or less.

The RPN, associated with a bundle of pericardiacophrenic vessels, descends along the anterolateral surface of the SVC, after which it turns posteriorly approaching the superior cavoatrial junction and continues in close proximity to the RPV before reaching the diaphragm 15.

Focusing on the anatomical relation of the RPN in CBA, it is essential to obtain accurate anatomical data on the course of the nerve. The aim of the analysis was to provide a comprehensive and evidence-based assessment of the RPN distance from RSPV and RIPV. We also aimed to investigate the course-related consequences and its association for RPNI.


Search strategy

Through September 2018, searchers were performed on MEDLINE, EMBASE, Google Scholar, and Cochrane Database in order to identify eligible articles for the meta-analysis. This review was performed as per the Preferred Reporting Items for Systematic Reviews and meta-analysis guidelines. The exhaustive search strategy employed for databases is presented in [Table 1]. No date limits or language restrictions were applied. The references in the included articles were also extensively searched. The risk of bias and quality assessment scores were both performed based on previously validated tools.

Table 1. Table of included studies, CT – computer tomography, RPPA – right pericardiacophrenic artery, 3D – three dimensional, PM – postmortem, SVC – superior vena cava. *6b Horton2010 (control group of 37 patients undergoing balloon-based procedures used in Horton study 21)
Study Country Type n (detection) N (RPN) % of visualisation
1 WANG 2016 16 China CT – RPPA 62 121 51,20%
2 MATSUMOTO 2007 17 USA/Japan CT – RPPA 50 106 47%
3 NIETO-TOLOSA 2011 18 Spain CT – RPPA 10 55 20%
4 OZAWA 2018 19 Japan CT – RPPV 18 56 30%
5 CANPOLAT 2014 20 Turkey CT – RPPV 145 162 89,50%
6a HORTON 2010 21 USA CT – RPPV 71 71 100%
6b HORTON 2010 21* USA CT – RPPV 37 37 100%
6c HORTON 2010 21 USA Pacing 71 71 100%
7 SCHMIDT 2008 22 Germany 3D/pacing RSPV 7 18 39%
8 ISO 2016 23 Japan 3D/pacing RSPV 20 20 100%
9 GOFF 2016 24 USA 3D/PM 7 10 70%
11 MARTINS 2014 6 France FLUORO/P-svc 81 81 100%
12 SMITH 2017 [25] USA Cadaveric 30 30 100%
13 SANCHEZ 2005 15 Spain Cadaveric 19 19 100%
14 RANDHAVA 2014 26 India Cadaveric 30 30 100%
Summary 658 887 74%

Study Country Type n (detection) N (RPN) % of visualization
Overall 658 887 74%
1 CT imaging CT – RPPV 393 679 57,88%
2 Intraoperative 3D/pacing RSPV 186 200 93%
3 Postmortem Cadaveric 79 79 100%

Criteria for study selection

Eligible studies for meta-analysis [for RPNI] were included under certain conditions such as: reported clear, data was easy to extract for further analysis, studies were postmortem cadaveric, electrophysiological, 3D modeling or imaging studies. The exclusion criteria included: case studies, case reports, conference abstracts and letters to the editor; incomplete and unclear data, studies on animals, or not related to electrophysiology procedures. For systematic review, anatomical, clinical, and radiological reports ware taken. We used them to show all possible discrepancies in tracing and measuring the distance of the right phrenic nerve. For identification of RPA in imaging studies, RPPA was taken into consideration – [Table 1] with a possibility of vessel detection. All studies were independently evaluated for inclusion by two reviewers (MK and MM). Any disparities arising during the assessment were resolved by a consensus among all the reviewers, after consulting with the authors of the original study, if possible.

Data extraction

Data were independently extracted from the included articles by two independent reviewers (MK and MM). These data included the ability to identify the right phrenic nerve RPPA, distance from PV ostium to the RPN, or the RPPA, amount of RPNI and correlation between distance and risk of RPNI. The primary outcome was to estimate the mean distance between right-sided PV ostium and RPN. Secondary outcomes such as the possibility of visualization RPN and risk of RPNI were also noted when data were available. In the event of data inconsistencies, the reviewers did not use the study in meta-analysis.

Statistical Analysis

Statistical analysis was performed by experienced statisticians using Statistica 13 software (StatSoft).  Continuous variables are expressed as mean ±SD or median range, as appropriate. 95% Confidence intervals for individual groups, and the whole was measured. For a subset of studies with analyzable and comparable data, the results were synthesized quantitatively by performing random-effects model meta-analyses to compute absolute net changes in continuous variables (i.e., RPN – RPV ostium) and pooled OR for binary variables (i.e., RPNI versus RPN noninjured). All pooled estimates were displayed with a 95% CI. The existence of heterogeneity among effect sizes of individual studies was assessed using the Q test and the T2 index, with a value of 95% or higher indicating medium-to-high heterogeneity. To explore sources of heterogeneity, we performed subgroup meta-analyses according to RPNI – RPV ostium’s and study typesetting (postmortem, intraoperative, or imaging). Sensitivity analysis, Egger’s Test, and Rosenthal's N - the "Fail-safe" test ware used.


Study Identification

[Figure 1] presents an overview of the flow of studies in the meta-analysis. Through database searching, 12441 initial articles were identified. An additional 550 articles were identified from reference searching. After removing duplicates and primary screening, 155 articles were assessed by full text for eligibility in the meta-analysis. Of these, 14 were deemed eligible and included, while 141 were excluded for not reporting extractable RPN-RPV distance rates.

Figure 1. Flowchart of study search, eligibility, assessment, and inclusion.

Characteristics of included studies

The characteristics of the included studies in the meta-analysis are listed in [Table 1], along with the reported RPN or RPPA identification. A total of 14 selected studies from 13 research publications (n=850; identified RPN -621,) 6 CT imaging (Matsumoto 2007, Horton 2010, Nieto-Tolosa 2011, Canpolat 2014, Wang 2016, Ozawa 2018) 3 postmortem/cadaveric (Sanchez 2005, Randhava 2014, Smith 2017), 5 procedural 3D/cadaveric (Goff 2016), 1 pacing (Horton 2010), 2 3D/pacing (Schmidt 2008, Iso 2016), 1 pacing/fluoroscopy were included (Martins 2010).

RPN/RPPB Detection capability:

The detection of the RPN may vary according to the identification method used. In postmortem studies, tracing the RPN was always accurate (100%); in invasive methods/intraoperative, it was almost always detectable (93%). Noninvasive methods of tracing RPPB are not as efficient as invasive or postmortem techniques, with only 57.88% (20-100%) of detection.

Mean distance of right phrenic nerve

From all included studies (n-507), the distance from RSPV ostium to RPN was 12.48mm, (±6,21) nevertheless, the distance varies according to the method used. In postmortem studies, the mean distance is significantly shorter (6.92mm) than in pre or intraoperative techniques (13.32mm noninvasive, 13.97mm invasive) [Figure 2]. The minimal measured distance from RSPV was 0mm – meaning the RPN was directly attached to the veno-atrial junction while the longest measurement was 42.6mm [Table 2]. There were significantly fewer amounts of measurements that were identified for RIPV. From the included studies (n-125), the mean distance between the RIPV ostium and RPN was 16.53mm (±8.92) with distances from 0.4 !! 68mm [Table 3].

Figure 2. Graphical presentation of the RPN distance from the RSPV ostium in the 95% confidence interval A – among individual publications B – among different groups of publications (1 – CT; 2 – Intraoperative; 3 – postmortem)

Table 2. RPN – RSPV distance
Publication Type of study N mean SD Range form Range to (95%CI)
SANCHEZ 2005 15 PM 19 2.1 0.4 1.5 2.5 (1.92 – 2.28)
RANDHAVA 2014 26 PM 30 7.47 5.51 0 20.81 (5.50 – 9.44)
SMITH 2017 17 PM 28 9.6 3.3 4.3 18.8 (8.38 – 10.82)
NIETO - TOLOSA 2011 18 CT 10 18.9 8 3 26 (13.94 – 23.86)
CANPOLAT 2014 20 CT 145 12.5 4.9 5.9 25 (11.7 – 13.3)
WANG 2016 16 CT 62 12.39 6.18 1.72 26.44 (10.85 – 13.93)
HORTON 2010 21 CT 71 15.2 8.3 3 42.6 (13.27 – 17.13)
HORTON 2010 21 CT 37 12.99 3.02 6 21.5 (12.02 – 13.96)
HORTON 2010 21 P 71 16 8.5 3 42 (14.02 – 17.98)
ISO 2016 23 3D/P 20 8.7 3 5 11.8 (7.38 – 10.01)
SCHMIDT 2008 22 3D/P 7 8.4 6.2 1 17 (3.81 – 12.99)
GOFF 2016 24 CT/3D 7 14 12 8 20 (5.11 – 22.89)
Overall 507 507 12.48 6.21 0 42.6 (11.94 – 13.02)
Group 1 325 325 13.32 5.96 1.72 42.6 (12.67 – 13.97)
Group 2 105 105 13.97 7.8 1 42 (12.48 – 15.46)
Group 3 77 77 6.92 3.94 0 20.81 (6.04 – 7.80)

Table 3. RPN-RIPV distance
Publication Type of study N Mean SD Min max (95%CI)
SANCHEZ 2005 15 PM 19 7.8 1.2 6 10 7.68 – 7.92
RANDHAVA 2014 26 PM 30 16.86 6.88 0.4 28.55 14.77 – 18.95
SMITH 2017 17 PM 28 Not measured Not measured Not measured Not measured Not measured
NIETO - TOLOSA 2011 18 CT 10 Not measured Not measured Not measured Not measured Not measured
CANPOLAT 2014 20 CT 145 Not measured Not measured Not measured Not measured Not measured
WANG 2016 16 CT 62 17.24 7.78 7.2 33.94 13.35 – 21.13
HORTON 2010 21 CT 71 Not measured Not measured Not measured Not measured Not measured
HORTON 2010 21 CT 37 Not measured Not measured Not measured Not measured Not measured
HORTON 2010 21 P 71 Not measured Not measured Not measured Not measured
ISO 2016 23 3D/P 20 Not measured Not measured Not measured Not measured Not measured
SCHMIDT 2008 22 3D/P 7 28.2 15.3 11 68 20.04 – 36.32
GOFF 2016 24 CT/3D 7 21 6.9 10 30 17.87 – 24.13
overall 125 125 16.53 8.92 0.4 68 11.77 – 21.28
Group 1 62 62 17.24 7.78 7.2 33.94 13.35 – 21.13
Group 2 14 14 25.07 14.04 10 68 29.59 – 20.54
Group 3 49 49 13.21 6.5 0.4 28.55 11.11 – 15.31

Risk of Phrenic nerve injury with distance included studies

Combining accessible data of the number of RPNI with distance included studies, we estimated that the amount of incidence was 12.46% (47 RPNI in 377 cases). The highest risk was in Martins research – 19.7%6, the lowest in Canpolat – 2.75%20.

Risk of phrenic nerve injury – distance relation

In the meta-analysis, a decreased distance between the RSPV ostium - RPNI, was directly associated with an increased risk of RPNI. The average distance was shorter by 6.61mm from the massed average of the invasive group [Figure 3]. Measuring the distance from the RSPV ostium to RPN, the damage distance was 7.36mm. Data from Horton 2010 21, Canpolat 2014 20, and ISO 2016 23 had the most significant impact on the study while Nieto-Tolosa 2011 18 did not – because of the small sample. Additional results - The variance of the effect size (T2) and the significance level (Wi) in individual studies used to calculate the sampling error are presented in []. In the cumulative variance the Nieto - Tolosa CT study brings the most additional variability (burden) – see appendix.

Figure 3. A: Forest plot for the risk of injury RPN and the distance relationship; D = average for damage - average with no damage; B - Forest plot of cumulative effect over the years of publications; D = average for damage - average with no damage;

Appendix Table 1. Average difference data D = average for damage - average with no damage SE – standard error, RSD relative standard deviation;
Meta-analysis (average difference); Variable effect; significance p < 0,05
Study D SE Lower limit 95% PU Upper limit 95% PU P share %
HORTON CT -5,41 0,85 -7,08 -3,74 0,0000 24,32%
CANPOLAT CT -5,40 1,08 -7,52 -3,28 0,0000 22,87%
TOLOSA CT -10,40 5,62 -21,42 0,62 0,0643 4,54%
ISO 3D -4,50 1,23 -6,92 -2,08 0,0003 21,85%
SCHMIDT 3D -2,70 3,59 -9,74 4,34 0,4523 8,94%
MARTINS FLUORO -13,50 1,87 -17,16 -9,84 0,0000 17,48%
Summary -6,61 1,31 -9,18 -4,04 0,0000 100,00%

The cumulative analysis was performed relative to the year in which the study was conducted – [Figure 3]. It shows how the cumulative effect and its estimation changed over time after upcoming publications. We can see that the initial insignificant cumulative effect changed to a significant effect after attaching the Horton 2010 publication. The results are presented below as a forest plot [Figure 3], a table in an appendix [].

table 2 appendix. Cumulative analysis data; D = average for damage - average with no damage SE – standard error, RSD relative standard deviation;
Meta-analysis (average difference); Variable effect cumulative analysis (by year); significance p < 0,05
Study year D SE Lower limit 95% Cl Upper limit 95% Cl p Cumulative share N1 cum N2 cum RSD
SCHMIDT 3D 2008 -2,70 3,59 -9,74 4,34 0,4523 8,94% 2 5
HORTON CT 2010 -5,27 0,83 -6,89 -3,65 0,0000 33,26% 9 35 -76,98%
TOLOSA CT 2011 -5,38 0,82 -6,98 -3,77 0,0000 37,80% 11 45 -1,06%
CANPOLAT CT 2013 -5,38 0,65 -6,66 -4,11 0,0000 60,67% 15 186 -20,24%
MARTINS FLUORO 2014 -7,25 1,68 -10,53 -3,96 0,0000 78,15% 28 254 157,15%
ISO 3D 2016 -6,61 1,31 -9,18 -4,04 0,0000 100,00% 31 271 -21,79%

Heterogeneity and sensitivity analysis

The study determines how significant the differences between the effects obtained in individual publications have an impact on the total effect of meta-analysis. The result that was calculated was p = 0.0014. The estimated value of real variance T2, and the corresponding 95% confidence interval are 6.3592 (1.606; 17.149). Real variance accounts for around 75% (coefficient I2) of the total volatility. This means that the research is highly heterogeneous. This confirms the correctness of the decision to build a model with a variable effect. This is confirmed by the following graphs of L'Abbegi and Galbraith. – see appendix [Figure 1] appendix and [Figure 2] appendix.

We also wanted to compare data of the injured phrenic nerves between various methods – [Figure 4]. We found three main methods of tracing the right phrenic nerve – pacing (blue; Tolosa, Horton, Canpolat) 3D – (green; Schmidt, ISO) and fluoroscopic evaluation (red Martins). Each method was presented in a forest plot below. We see that pacing and 3D systems are similar in detecting the presence of the RPN, while Martins 2014 significantly differs because of a different method used.

The different method used by Martins was significantly highlighted in sensitivity analysis – [Figure 4], when individual tests were turned off. The cumulative effect and limits of the confidence interval for the overall meta-analysis are marked on this graph with vertical lines [Figure 4] and data in [].

Figure 4. A. Sensitivity analysis in various methods of tracing RPN, B. Sensitivity analysis for each publication.

Table 3 appendix. Appendix Sensitivity analysis for each publication - data; D = average for damage - average with no damage SE – standard error, RSD relative standard deviation; dSE – deviation for standar error;
Sensitivity analysis D (Odds ratio); variable effect highlited p < 0,05
Included study Group D SE Lower limit 95% Cl Upper limit 95% Cl P share % dSE
ISO 3D 2 -7,42 2,12 -11,57 -3,27 0,0005 77,02% 37,45%
SCHMIDT 3D 2 -7,28 1,68 -10,57 -3,99 0,0000 89,06% 8,90%
MARTINS FLUORO 3 -5,13 0,66 -6,42 -3,84 0,0000 80,69% -57,36%
HORTON CT 1 -7,19 2,32 -11,74 -2,65 0,0019 75,08% 50,33%
CANPOLAT CT 1 -7,04 1,81 -10,60 -3,49 0,0001 84,03% 17,71%
TOLOSA CT 1 -6,53 1,62 -9,71 -3,35 0,0001 94,12% 5,15%
Without exclusion -6,76 1,54 -9,78 -3,74 0,0000 100,00% 0,00%

Assessment of publication bias and Fail-safe number test

The funnel plot for the outcome of RPNI in the studies included in the meta-analysis was symmetric and the Egger test was not significant (P = .59), did not show the existence of susceptibility to publication bias.

The last issue examined was whether it would be necessary to further analyze the distance of the RPN over the right pulmonary veins. For which we used Rosenthal's N (the "fail-safe" number) described by Rosenthal (1979). This value determined the number of works that did not indicate an effect (e.g., a mean difference of 0) that was needed to reduce the summary effect from statistically significant to statistically insignificant. In our analysis, the number of "fail-safe" publications determined by the Rosenthal method amounted to 72. Therefore, the obtained effect could be considered stable because it will not (with a small number of works) undermine the obtained effect. Besides, the Rosenthal factor fs = 4.35 was greater than 1, which means that the probability of publishing significance is minimal. As we know (obtained), the common difference was D = -6.61. The Orwin method checked how many papers with a relative risk of -3 needed to be included to reduce the relative risk to 1.12. The obtained result was 16 research papers.


Present results and their variability showed that methods of distance measurement of the RPN may have a significant impact on the final result. postmortem study results have a different approach - where the distance assessment is taken externally from the epicardium, differently from the non-invasive visualization where the RPP bundle is traced, and finally, the intraoperative method with endocardial stimulation. The same intraoperative methods also have high variability relative to each other, as demonstrated by statistical analysis.


From all included studies (n-507), the distance from the right superior pulmonary vein (RSPV) ostium to the RPN was 12.48mm (±6.21). In postmortem studies, the distance was significantly shorter, ranged 6.92mm (±3.94). From imaging studies, the distance to RPPB was 13.32mm (±5.96). A similar distance was found collecting intraoperative techniques - 13.97mm (±7,8). Distances ranged from 0 to 42.6 mm.

Postmortem study

The most cited publication (Sanchez 2005) 15 of RPN relation to the right side of the heart has a very interesting and unique methodology. The distance was measured from the vein margins to the presence of the RPN, with minimal and maximal distances on the right side of the heart, and not from the vein ostium. The differences of the minimal and maximal values for the same RPV are because of the shape of the cross-section of RPV, but most notably, the veins in postmortem studies were without blood pressure and their shape differs from a working heart. The minimal value from this publication was taken into meta-analysis because its distance relationships were often taken into discussion in many publications 22. Randhawa 2014 24 used the same methodology and had different measurements, more similar to the endocardial distances. Finally, Smith 2017 25 combined both of these techniques to create a reliable landmark for operators. The values were comparable with those from pacing or CT imaging

CT imaging/Preoperative visualization

In CT imaging, the RPN is not directly visualized, but the course of RPPB is. Among these structures, the exact course of the RPN may be variable. Delineation of the PN is highly dependent on image quality, which may raise the question of whether this technique can be commonly used before the procedure 23. The best example was illustrated by Nieto-Tolosa18 in which RPN detection was on the 20% level. The authors suggested that patients in whom RPPB could be visualized tended to have a higher BMI. Canpolat 20 three years later proved that this technique might be used both in average and higher BMI patients.

Nevertheless, distance measurement before the procedure will only keep the operator more alert during RSPV freeze. This information should be correlated with RPV ostium size and depth setting of the CB during the procedure. Thus, the preoperative assessment of RPPB will only have informative value for an operator.

Pacing/3D mapping

Intraoperative measurements are most precise in detecting RPN. Stimulation from the place of cryoapplication may be the most accurate. However, many factors which play a role in proper RPN localization, such as the strength of stimulation impulse, usage of the 3D system, and force of balloon setting. Another issue is comparing stimulation from the SVC to the balloon tip in AP view 6, which showed no significant compatibility to other studies. Stimulation strength plays a role in depicting RPN, and the bias margin rises with the power output 6, 21, 27 or wall thickness 15. Usage of the 3D system in CBA is relatively rare, however, Iso study23 was a milestone into understanding how the CB changes the shape of PV ostium23. Iso uses both methods of measuring the RPN distance: from RSPV orifice to pacing the RPN capture, and the postmortem type method, from RSPV to the PN location corresponding with the RA in a 3D system.

Interestingly the distances in the second method were significantly longer than in the real postmortem studies (2,1mm Sánchez-Quintana; 7,74 Randhava; vs. 11,4 Iso p<0,05). The reason is that the RPN location is somewhere between RSPV and RA, including wall thickness, pericardial sack, and epicardial adipose tissue presence. The last thing which plays a role in RPN distance is the deep setting of the CB. The distances before and after the balloon setting to the RPN significantly differ on average, by about 3.4mm 23. The same results were presented in Goff research 24. He proved that small variations in balloon placement could lead to substantial changes in phrenic – to – balloon distances. In summary, RPN pacing is clinically useful for delineating the precise course of the RPN. The RPN capture is dependent on the relative distance between the pacing site and the nerve position and course. Relative distance is dependent from force balloon setting in the RPV ostium, possibly decreasing the distance to the danger zone.


We harvest possible data from publications used for this study. The only parameter to estimate was the mean distance of the RPN to the ostium from the RIPV. The mean distance (n-125) was 16.53mm (±8.92) with distances from 0.4 !! 68mm and was significantly longer than from RSPV (p<0,05). From that reason the risk of RPNI is significantly lower than from RSPV. Because of a small amount of RPNI in RIPV, insufficient data makes meta-analysis unattainable. The interesting fact presented in the publications was the presence of a right common ostium (RCO) and vigorous wedging against a large PV ostium will provoke the balloon's distal displacement close to the RPN 30,31.

Danger zone of RPNI

In our analysis, the mean risk of RPNI was surprisingly high: 12.46% from RSPV, and 3.5% from RIPV. This high value occurred because the publications were concentrated on RPNI and were from the early period of cryoballoon ablation using the first-generation CB. There was no research in the database of CB studies without the risk of RPNI. This value is decreasing over the years, due to new balloon generations, novel techniques in avoiding RPNI risk and volume of procedures. The lowest risk was in Franceschi's research with feasible usage of CMAP with a risk of incidence of 0.7% 32. The danger zone was estimated at 7.36mm from the RSPV ostium, i.e., it was close to the diaphragmatic nerve distance average in the postmortem examination. It is slightly closer than proposed by Canpolat 20: -8mm. Again, the distance is not the only risk predictor. Many research papers noticed that the balloon tip was often significantly further from the projected distance relative to a given phrenic nerve without RPNI 6,20,21,24. Preparing this meta-analysis, we harvest the PNP data from over 10K CBA finding 783 RPNI in 54 significant publications from 2008 estimating the mean risk level to be 7.7% - but that was not the problematic aspect in this research. These studies suggest that the location of the nerve near the PV ostium is associated with a higher risk of RPNI.

Questions for further investigation

After obtaining the distance information on relatively large groups of measurements (n-507), the first question is, will further research may contribute to a change in average distances? From our analysis, we know that a minimal amount of publications – with different values – need to exceed at least 16 in the Orwin method while in Rosenthal’s, 72 research publications.

Another question concerns the idea that phrenic nerve injury may not only be dependent on course proximity to the PV ostia, but of other issues such as flow size through vessels accompanying RPN in RPPB or thickness of fat surrounding the bundle. Interestingly, not every RPN in the range of balloon freeze was paralyzed.

Finally – how to prevent or decrease the risk of RPNI? Still, the five main strategies play a role in preventing RPNI during cryoablation. From most commonly used fluoroscopy and palpation, through electromyography – CMAP, to intracardiac echocardiography (ICE) and auditory cardiotocography least often28. The most crucial direct protection of the RPN during CBA, in our perspective, is CMAP 29.


RPNI is a relatively rare complication. A firm understanding of its course, relation to the PV ostium, and detection are vital for preventing future injuries and complications.


  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]
  • [ PubMed ]