|Home | About | Journals | Submit | Contact Us | Français|
Claims in the medical literature suggest that neck fullness and witnessed neck pulsations are useful in the diagnosis of typical AV nodal reentrant tachycardia (AVNRT).
Neck fullness and witnessed neck pulsations have a high positive predictive value in the diagnosis of typical AVNRT.
We performed a cross sectional study of consecutive patients with palpitations presenting to a single electrophysiology (EP) laboratory over a 1 year period. Each patient underwent a standard questionnaire regarding neck fullness and/or witnessed neck pulsations during their palpitations. The reference standard for diagnosis was determined by electrocardiogram and invasive EP studies.
Comparing typical AVNRT to atrial fibrillation (AF) or atrial flutter (AFL) patients, the proportions with neck fullness and witnessed neck pulsations did not significantly differ: in the best case scenario (using the upper end of the 95% confidence interval [CI]), none of the positive or negative predictive values exceeded 79%. After restricting the population to those with supraventricular tachycardia other than AF or AFL (SVT), neck fullness again exhibited poor test characteristics; however, witnessed neck pulsations exhibited a specificity of 97% (95% CI 90–100%) and a positive predictive value of 83% (95% CI 52–98%). After adjustment for potential confounders, SVT patients with witnessed neck pulsations had a 7 fold greater odds of having typical AVNRT, p=0.029.
Although neither neck fullness nor witnessed neck pulsations are useful in distinguishing typical AVNRT from AF or AFL, witnessed neck pulsations are specific for the presence of typical AVNRT among those with SVT.
Palpitations are one of the most commonly reported symptoms in patients presenting to internists and cardiologists,1 and almost half are attributable to cardiac arrhythmias.2 The majority of these cardiac arrhythmias will have a supraventricular origin, but the appropriate management will vary widely depending on the nature of the arrhythmia.3 For example, patients with atrial fibrillation (AF) and atrial flutter (AFL) may benefit from early consideration of anticoagulation, and patients with other paroxysmal supraventricular tachycardias (SVT) may be readily cured with catheter ablation. Specifically, the most common cause of SVT, atrioventricular node reentry tachycardia (AVNRT), can be successfully treated in 98–100% of cases with particularly low risk.4
Unfortunately, the exact diagnosis of palpitations often requires event monitoring over several weeks and, in the case of SVT, often ultimately requires an invasive electrophysiology study. Conventional wisdom suggests that the presence of typical AVNRT in a patient with palpitations can be immediately determined from the history and physical exam, potentially guiding accurate management and appropriate patient counseling.5 While perceived fullness in the neck has been associated with typical AVNRT, the predictive value of this symptom has not been previously reported.6 Moreover, although important and influential review articles in the literature as well as prominent text books state that witnessed neck pulsations during tachycardia indicate the presence of typical AVNRT, this association has never been studied.1, 7–10 Importantly, neither of these historical points have been evaluated in patients with AF or AFL, arguably the most clinically relevant group of patients to identify or exclude given the pertinent anticoagulation issues. With the hypothesis that typical AVNRT could be distinguished from AF, AFL, and other mechanisms of SVT by patient history of neck fullness and/or witnessed neck pulsations, we analyzed the test characteristics of these historical points in making an accurate diagnosis.
We performed a prospective observational study of consecutive adult patients with a history of palpitations presenting to a single electrophysiology (EP) laboratory over a 1 year period. Patients with a documented history of both SVT and AF or AFL, known or suspected ventricular arrhythmias, or patients undergoing a procedure that was not potentially immediately diagnostic (such as implantation of a loop recorder) were excluded; therefore, patients presenting for cardioversion for AF or AFL or EP study and/or ablation for AF, AFL, or SVT were included.
In addition to obtaining a medical history from chart review, each patient underwent a structured interview with a standardized questionnaire. Subjects were asked if they had any sensation of fullness in their neck during episodes of palpitations and/or if anyone had ever told them that their neck appeared to be pulsating during their palpitation episodes. While “neck fullness” was intentionally broad (with a “yes” designation for any neck sort of fullness, including “pounding” for example), patients that were not certain were designated as “no.” For patients with known or suspected SVT (as opposed to AF or AFL), the interviewer was blinded to the ultimate diagnosis. In those with AF or AFL, the rhythm was often apparent to the interviewer due to telemetry monitoring and therefore blinding could not be guaranteed. For those undergoing procedures for AF or AFL, the diagnosis was determined by either chart review (with histories confirmed by documented arrhythmia by telemetry monitoring or 12 lead electrocardiogram) or documented evidence of the sustained arrhythmia on the day of the interview. For those undergoing ablation for SVT (no patients underwent cardioversion alone for SVT), results of the EP study were used to determine the exact mechanism: all patients underwent placement of standard right atrial, His, right ventricular, and coronary sinus multielectrode catheters under fluoroscopic guidance; determination of the arrhythmia mechanism was performed with standard intracardiac recordings and pacing maneuvers.11 When more than one arrhythmia was observed in the electrophysiology laboratory, the primary diagnosis was determined by the treating electrophysiologist (who was blinded to the results of the interview) based on the presence of a sustained rhythm, and, when available, correlation with the clinical 12-lead electrocardiogram during tachycardia.
All patients provided witnessed and informed consent. The study was approved by the University of California, San Francisco Committee on Human Research.
Normally distributed continuous variables are expressed as means ± SD. Analyses of normally distributed continuous variables were assessed using t-tests or analysis of variance as appropriate, and categorical variables were compared using the χ2 test. Test characteristics, including sensitivity, specificity, positive and negative predictive values with 95% confidence intervals (CI) for the historical points were calculated using documented histories in conjunction with electrocardiograms (in the case of those undergoing cardioversion for AF or AFL) and invasive EP study in the remainder as reference standards. Multivariable analysis was performed with logistic regression analysis, and covariates/potential confounders were selected for the regression analysis based on both important demographics (eg, age, race, and gender) and those covariates significantly associated with both the predictors and outcomes of interest with p values < 0.10. Two-tailed p values < 0.05 were considered statistically significant.
A total of 239 patients were enrolled. Of these, 48 patients (20%) experienced neck fullness during tachycardia (Table 1). Compared to other races, Latino patients were more likely to report neck fullness (p=0.003). In contrast, white patients were less likely than all other races to report neck fullness, but this was not statistically significant (p=0.053). There were no other subject characteristics that were significantly associated with neck fullness.
Twenty three patients (10%) reported witnessed neck pulsations (Table 1). While women and those with a history of coronary artery disease more often reported witnessed neck pulsations, none of these findings reached statistical significance.
Patients characteristics of those with typical AVNRT, AF/AFL, and SVTs other than typical AVNRT are described in Table 2. Of the 124 subjects in the AF/AFL group, 82 had AF and 42 had AFL. Of the 115 SVT subjects, the diagnosis made during EP study was typical AVNRT in 51 (43%); those with “Other SVT” (SVT other than typical AVNRT) included atrioventricular reciprocating tachycardia (AVRT) in 29 (24%), focal atrial tachycardia in 15 (6%), atypical AVNRT in 4 (2%), reentrant atrial tachycardia in 1 (1%), and junctional tachycardia in 1 (1%). Two patients with typical AVNRT exhibited other arrhythmias during electrophysiology study (1 with atrial tachycardia and 1 with AVRT). Patients with typical AVNRT were significantly younger than patients with AF/AFL, but significantly older than patients with other types of SVT. Compared to patients with AF/AFL and other SVTs, patients with typical AVNRT were more likely to be female. Finally, patients with AF/AFL were significantly more likely to have congestive heart failure.
The proportion of subjects reporting neck fullness and witnessed neck pulsations in each group are shown in Figure 1.
There was no difference in the proportion of subjects reporting neck fullness between those with AF/AFL and those with typical AVNRT (18% in each group). Of note, AF patients more often reported neck fullness than AFL patients (17 [24%] versus 4 [9%], p=0.04). As shown in Table 3, while neck fullness was not at all sensitive for distinguishing typical AVNRT from AF/AFL (18%), it was moderately specific (82%). However, neither of the predictive values was high.
While more patients with typical AVNRT reported witnessed neck pulsations than those with AF/AFL (19% versus 9%), the difference was not statistically significant. While witnessed neck pulsations were found to be highly specific (91%) for distinguishing typical AVNRT from AF/AFL (Table 3), the positive predictive value at best did not exceed 70% (per the 95% CI).
Bivariate analysis revealed no difference between patients reporting neck fullness in either group (18% in the typical AVNRT group versus 22% in the Other SVT group, p=0.55). As was the case in comparison with AF/AFL, among subjects with SVT, neck fullness was not sensitive (18%) and was moderately specific (78%) for the presence of typical AVNRT (Table 3). Neither predictive value exceeded 60% (or 67% if the upper end of the 95% CI is used).
Table 4 depicts the percentage of patients reporting neck fullness by specific type of SVT.
Patients with typical AVNRT were significantly more likely to report witnessed neck pulsations compared to patients with other SVTs (20% versus 3%, p=0.003). After adjusting for age, gender, and race, typical AVNRT patients remained significantly more likely than other SVT patients to report witnessed neck pulsations: typical AVNRT patients had a 7 fold greater odds of reporting witnessed neck pulsations (95% CI 1.2–45 greater odds, p=0.029). Among SVT patients, witnessed neck pulsations were highly specific for typical AVNRT (97%), and exhibited a moderately high positive predictive value (83%, Table 3).
Table 4 depicts the percentage of patients reporting witnessed neck pulsations by specific type of SVT. No statistically significant differences were observed.
Contrary to classic teaching regarding the utility of neck fullness and witnessed neck pulsations in making an accurate diagnosis of typical AVNRT,1, 12 neither historical point was useful in distinguishing typical AVNRT from AF or AFL. However, amongst those with SVT, while a sensation of neck fullness demonstrated little utility, witnessed neck pulsations were highly associated with typical AVNRT in both univariate and multivariate analyses, demonstrating a high specificity and a moderately high positive predictive value.
The theoretical basis of both neck fullness and neck pulsations in typical AVNRT arises from the fact that the arrhythmia is, essentially by definition, an “A on V” tachycardia, or one where the atria and ventricles are simultaneously (or nearly simultaneously) contracting.13 Hemodynamic assessments during AVNRT support this theory, revealing a single AV wave and a higher degree of flow reversal on right atrial angiography.6 In contrast, in short RP tachycardias (such as in AVRT) or long RP tachycardias, such as in sinus tachycardia or the majority of atrial tachycardias, the sequential contraction of the ventricles followed by the atria (or visa versa) should not result in the atria contracting against closed atrioventricular valves. Finally, neither AF nor AFL are thought to generate enough of an atrial contraction to cause either symptoms of neck fullness or cannon A waves.
Gursoy et al. queried 244 consecutive patients with a history of palpitations regarding any symptoms of rapid and/or regular pounding in the neck (i.e., neck fullness) during palpitations.6 These symptoms were then correlated with the results of the electrophysiology study when an arrhythmia was observed or induced or when the patient felt palpitations. In all but 4 of 54 patients with AVNRT, patients reported “neck pounding,” whereas none of the other 190 patients with other arrhythmias reported this symptom. This study has become a widely cited paper associating neck fullness and the diagnosis of typical AVNRT.1, 12 From this report, prominent text books have referred to neck fullness or “pounding” as a helpful symptom in making the diagnosis of AVNRT,8, 9 with one stating that the symptom is “practically pathognomonic” of AVNRT.10
In fact however, previous studies have never examined the utility of neck fullness when including subjects with AF or AFL. Because the management of AF/AFL patients is significantly different from that of AVNRT (and as the management of typical AVNRT is not largely different from the management of other SVTs such as AVRT or atrial tachycardia), the argument can be made that the AVNRT versus AF/AFL distinction is in fact the most crucial one. Unfortunately, it appears that this symptom has no value in making this distinction.
While cannon A waves or witnessed neck pulsations (sometimes called the “frog sign”5) during tachycardia have also been purported to be particularly helpful in making the diagnosis of typical AVNRT,1, 8, 9 this has, to our knowledge, never been formally studied. Although this historical point was again not helpful in excluding AF or AFL, it was helpful in distinguishing typical AVNRT from other types of SVT. Although the immediate management in clinic may not differ as a result of this finding, counseling the patient regarding the details, rates of success, and risks of invasive curative ablation procedures may in fact be influenced.13 For example, whereas ablation of AVRT may often involve a left sided approach (such as via a transseptal puncture), typical AVNRT can be addressed from the right sided circulation in the majority of cases. Similarly, the success of typical AVNRT ablation is likely higher than that of other arrhythmias such as atrial tachycardia.13
The reasons for the discrepancy between the previous literature and our findings are not clear. It may be that the textbooks have been guided primarily by clinical anecdote, which often relies on extreme and perhaps more obvious cases without a more objective measurement of the number of cases that do not fit the paradigm or the denominator of the total cases without a given sign or symptom. In addition, although the study by Gursoy et al. addressed this topic (as above),6 the main point of that paper involved understanding the hemodynamic mechanism of symptoms in AVNRT rather than a blinded, rigorous assessment of the test characteristics of the symptoms, without inclusion of patients with AF or AFL.
This study has several limitations. First, the neck pulsations were not directly observed, but rather reported as witnessed by the patient. While this might take away from the accuracy in determining whether or not obvious cannon A waves were indeed present, the point of interest in clinical practice is in fact the historical report of witnessed neck pulsations. In other words, the point is to test the sensitivity and specificity of the particular question (not necessarily the phenomenon itself) that can be asked in the clinic or emergency department when the patient is in sinus rhythm. Second, although the interviewer was blinded to the diagnosis amongst all of those with SVT, it was not practical to blind the interviewer in some cases where the patient presented in AF or AFL and a cardiac monitor was clearly visible. However, given the traditional teaching regarding the questions in regards to typical AVNRT, one would have expected a lack of blinding in interviewing the AF/AFL subjects to bias the test characteristics in favor of typical AVNRT. Finally, we did not include patients with ventricular tachycardia, another group that might be expected to have cannon a waves and therefore theoretically similar symptoms; indeed, this issue is further complicated by the fact that individuals with idiopathic ventricular tachycardia more often exhibit dual AV nodal physiology and AVNRT,14 making the interpretation of neck symptoms in these patients potentially quite difficult.
In conclusion, contrary to claims promulgated in prominent review articles and text books regarding the utility of neck fullness and witnessed neck pulsations during episodes of palpitations, neither helps to distinguish typical AVNRT from AF or AFL. Among subjects with known SVT (other than AF or AFL), witnessed neck pulsations are significantly more common in patients with typical AVNRT, with a high specificity and moderately high positive predictive value.
This work was made possible by Grant Number KL2 RR024130 from the National Center for Research Resources, a component of the NIH (G.M.M.), and the American Heart Association Western States Affiliate Beginning Grant-in-Aid Award (G. M. M.).