Search

A narrow QRS complex is a hallmark of a normal beat, although this common assumption is questioned in the findings of the present case. This electrocardiographic challenge investigates the unique mechanism of a narrow QRS and the clinical importance of discovering this phenomenon. (Level of Difficulty: Intermediate.), Central Illustration

A 17‐year‐old woman presented with frequent palpitations and shortness of breath and was diagnosed with drug‐refractory ventricular parasystole. We predicted that the parasystole originated from the left anterior fascicle (LAF). Detailed activation maps of both conduction systems, including the LAF, during sinus rhythm and ventricular parasystole were obtained using a parallel mapping system. We confirmed the earliest fascicular potential of the parasystole and performed catheter ablation with no complications. This novel mapping algorithm for simultaneous acquisition of multiple maps aided effective treatment of ventricular parasystole originating from the LAF., Ventricular parasystole originating from the left anterior fascicle can be effectively eliminated by using this novel mapping algorithm of parallel mapping and LAT hybrid mapping. Parallel mapping makes it possible to obtain both activation maps during ventricular parasystole and sinus rhythm simultaneously, decreasing the procedure time while also increasing efficacy. LAT hybrid mapping also makes it possible for precise positioning during parasystole ablation.

Ventricular parasystole is a kind of cardiac disease that can induce sudden death. To diagnose this disease, cardiologists often spend a long time tracking dynamic electrocardiogram to identify parasystolic episodes. It is urgent to develop automatic detection based on artificial intelligence to help doctors effectively improve the diagnosis rate in the early stage. In this paper, an identification algorithm based on the Gaussian mixture distribution test is proposed. Firstly, the Gaussian mixture (GM) model is established to characterize the intervals of adjacent ectopic ventricular beats, modulated by sinus rhythms. Secondly, the characters of model parameters are summarized by analyzing the relative regularity of the ectopic beats. Then a clustering algorithm is proposed to detect the class number of ventricular intervals, and the expectation-maximization (EM) algorithm is used to obtain the expectation and variance of each sub-distribution. Lastly, the existence of ventricular parasystole is determined by hypothesis testing of the model parameters. Two cardiologists have marked manual annotations of ventricular parasystole to the open-access INCART 12-lead arrhythmia database as gold standards. On 75 records of the dataset, the proposed algorithm is tested using the five-fold cross-validation scheme and achieves a sensitivity of (100.00 ± 0.0)%, positive predictive (93.33 ± 11.93)%, specificity (98.46 ± 2.75)%, and overall accuracy (98.67 ± 2.39)%. These results reflect that our method is superior to the current inverted Y model of Lorenz scatterplot.

During a clinical trial assessing the efficacy of ranolazine for the suppression of atrial fibrillation, a patient was encountered who, aside from paroxysmal atrial fibrillation, had premature ventricular complexes (PVC) on his electrocardiogram (ECG). A close examination of the ECG rhythm strip revealed that these PVC’s had: 1) a regular interectopic interval or multiple thereof (approximately 1720 ms), 2) fusion complexes with the underlying sinus rhythm QRS complexes, and 3) variable coupling intervals. This constellation of findings defines a parasystolic focus, making this rhythm sinus rhythm with ventricular parasystole. A dose of ranolazine 1000 mg po bid satisfactorily reduced the patient’s atrial fibrillation and also suppressed virtually all of the ventricular ectopy.

BACKGROUND In 1974, Kinoshita reported a case of 'irregular parasystole' due to type I second-degree entrance block. Since then, many cases of such 'irregular' parasystole have been reported by us. To explain the mechanism of 'irregular' parasystole, two theories have been suggested, namely, 'electrotonic modulation' by Jalife and Moe, and 'type I second-degree entrance block' by us. On the contrary, in 1960, Kinoshita et al. reported a case of concealed bigeminy for the first time. The electrocardiographic findings in concealed bigeminy have suggested that there are dual re-entrant pathways with markedly long effective refractory periods in the re-entrant pathway. We have suggested that parasystole may be caused by re-entry in such re-entrant pathways. In this article, attempts are made to explain the mechanism of all the electrocardiographic findings in our cases of parasystole by 'parasystole due to re-entry'. METHODS Using 24 studies on parasystole and 21 studies on concealed extrasystoles that we have reported over 50 years, as well as three exemplary cases in this article, attempts are made to explain all electrocardiographic findings in parasystole by 'parasystole due to re-entry'. CONCLUSIONS The electrocardiographic findings in our previous clinical cases of parasystole and concealed extrasystoles, as well as exemplary cases and diagrams in the present article, strongly suggest 'parasystole due to re-entry' as the mechanism of ventricular parasystole with second-degree entrance block.

Electrocardiograms were taken from a 44-year-old man with irregular ventricular parasystole in whom pure parasystolic cycles without any intervening nonectopic QRS complexes were found. When a sinus impulse fell late in the parasystolic cycle, it hastened occurrence of the next parasystolic discharge. This suggested that type I second degree entrance block occurred in the re-entrant pathway containing the parasystolic focus. When a sinus impulse fell early in the parasystolic cycle, it delayed occurrence of the next parasystolic discharge. This suggested that electrotonic modulation occurred in the parasystolic focus. As a result, the difference in length between the short form and the long form of the parasystolic cycle became markedly great. When the length of two adjacent sinus cycles ranged between the short and the long parasystolic cycle, manifest parasystolic QRS complexes disappeared for a long time. In true ventricular parasystole with pure ectopic cycles, such long disappearance has never been reported before.

Mad honey poisoning syndrome has been reported in the Eastern Black Sea region and Southeastern regions of Turkey. Herein we report a case of 70‐y‐old man presented with syncope and severe hemodynamic instability following ingestion of one teaspoon of honey and his unusual electrocardiographic manifestations: nodal rhythm alternating with sinus bradycardia and intermittant ventricular parasystole. In this report, we also tried to explain the possible mechanism responsible for these electrocardiographic findings. Copyright © 2009 Wiley Periodicals, Inc.

A problem of identifying the phase response curve of the myocardial pacemaker was investigated using a simple dynamic system model of cardiac arrhythmia. A hybrid optimization method of a genetic algorithm and a local optimization technique was employed to obtain a solution of the model fitting. The method was applied to two categories of arrhythmias, ventricular parasystole and high-degree AV block. The present study offered a new insight into the mechanism of cardiac arrhythmias.

OBJECTIVE: To investigate the effect of standing on the parasystolic cycle length in cases of "true" ventricular parasystole. METHODS: Parasystolic cycle length and sinus cycle length were measured during lying and standing in eight men with true ventricular parasystole. These cycle lengths were also measured after exercise in the lying position. RESULTS: In all cases, parasystolic cycle length and sinus cycle length both shortened on standing, by a mean of 6.4% and 17.8%, respectively, compared to lying. In all cases, the rate of shortening of the parasystolic cycle length was less than that of the sinus cycle length. Parasystolic cycle length was prolonged after exercise, in contrast to a shortening of the sinus cycle length. CONCLUSIONS: Influences on the parasystolic cycle length are not always in the same direction as on the sinus cycle length. This suggests that the effect of autonomic changes on parasystolic rhythm is not always parallel to that on sinus rhythm.

One-dimensional circle maps are good models for describing the nonlinear dynamical behavior of two interacting oscillators. They have been employed to characterize the interaction between a periodic external forcing stimulus and an in vitro preparation of chick embryonic cardiac cells. They have also been used to model some human cardiac arrythmias such as modulated ventricular parasystole. In this paper, we describe several techniques involving engineering feedback control theory applied to a circle map model of human heart parasystole. Through simulations of the mathematical model, we demonstrate that a desired target phase relationship between the normal sinus rhythm and an abnormal ectopic pacemaker can be achieved rapidly with low-level external stimulation applied to the system. Specifically, we elucidate the linear, self-tuning, and nonlinear feedback approaches to control. The nonlinear methods are the fastest and most accurate, yet the most complex and computationally expensive to implement of the three types. The linear approach is the easiest to implement but may not be accurate enough in real applications, and the self-tuning methods are a compromise between the other two. The latter was successful in tracking a variety of period-1, period-2, and period-3 target phase trajectories of the heart model.

Objective —A mathematical model of modulated ventricular parasystole based on the relation between the coupling interval and the preceding RR interval was developed in an attempt to distinguish between parasystolic automaticity and other mechanisms. Mathematical model —The relation between the coupling interval and the preceding RR interval was examined by plotting the coupling interval of each extrasystole against the preceding RR interval (coupling interval/RR diagram). The coupling interval/RR diagrams obtained from simulations with various modulation modes suggested that the parasystolic mechanism was likely when the dots representing extrasystoles appeared as discrete clusters. In contrast, a linear horizontal accumulation of dots indicated a non-parasystolic mechanism. Clinical observation —To verify the validity of the simulations, 24 hour electrocardiographic recordings from 60 patients with frequent ventricular extrasystoles (>1000/day) were analysed to determine whether the extrasystoles showed intrinsic periodicity. Intrinsic periodicity indicative of a parasystolic mechanism was seen in 14 (93%) of 15 patients in whom the coupling interval/RR diagram was characteristic of a parasystolic mechanism. When the coupling interval did not change (variability 700 ms) intrinsic periodicity was never identified (0/17). Parasystolic automaticity was the likely mechanism in 11 of the remaining 28 patients (39·3%) in whom coupling interval/RR diagrams were not definitive. Conclusion —These data indicate that definite patterns of coupling interval/RR diagrams can be used to distinguish between parasystolic and non-parasystolic mechanisms.

A 45-year-old patient free of any heart disease was admitted to the hospital with an electrocardiographic pattern of ventricular parasystole. The parasystolic rhythm was relatively fast, such that several consecutive ectopic complexes manifested. A later tracing reflected only isolated parasystolic complexes with long and fixed coupling intervals. The interectopic intervals, however, were once more in multiples of the parasystolic cycle as directly measured during the phases of undisturbed parasystolic rhythm. In the latter tracing, several scheduled parasystolic impulses did not yield a response, despite calculation suggesting that these impulses occurred outside the refractory period. In other words, an exit block was present. Analysis of the tracing suggests that the exit block was caused by concealed penetration of the sinus impulses into the ectopic-ventricular junction. That is, any sinus impulse penetrates into the junction and renders it refractory, in such a way that only parasystolic impulses that are relatively late within the sinus cycle may be conducted to the surrounding myocardium and result in a parasystolic complex.

A rare case of parasystolic ventricular tachycardia with variable exit block concurring with runs of reentrant ventricular tachycardia is described. The reentrant ventricular tachycardia or isolated ventricular extrabeats modified the exit block patterns of the parasystolic tachycardia.

A 50-year-old man with intermittent ventricular parasystolic bigeminy is reported in whom the parasystolic focus was protected from late intervening sinus impulses. This is the first report to suggest the presence of protection due to interference in parasystolic bigeminy. The findings in this case suggest that when a sinus impulse falls in a late period of the parasystolic cycle, it travells so slowly along the ventricular-ectopic junction that it is unable to reach the parasystolic focus before the spontaneous occurrence of the next parasystolic impulse; as a result, the sinus impulse interferes with the next parasystolic impulse in the ventricular-ectopic junction. Thus it is suggested that the parasystolic focus is protected from the sinus impulse because of the interference and not because of an entrance block. This reinforces the concept of a second-degree entrance block as a mechanism of parasystole.

This electrocardiogram of a 90-year-old man shows repeating similar cycles, each consisting of 4 complexes (Figure). Every cycle begins with 2 sinus-initiated complexes followed by an early, wide QRS (ventricular premature complex) that has a right-bundle-branch-block configuration and precedes a pause that is terminated by a slightly different, wide QRS. The cycle then repeats itself. What is not readily apparent is that the interval between each ventricular premature complex and the following wide QRS (1.10 seconds) is exactly half the interval between this second wide QRS of each cycle and the ventricular premature complex of the next cycle (2.20 seconds). Furthermore, an interval of 1.10 seconds from the beginning of the last QRS of each cycle lies in the ventricular refractory period of the first sinus-initiated complex of the next cycle (vertical lines above the V1 rhythm strip). Thus, one can make a good case for a left ventricular, parasystolic focus discharging every 1.10 seconds, with every third impulse failing to capture the ventricles because it arrives in the absolute refractory period. Figure Electrocardiogram in a 90-year-old man shows a repetitive pattern, or group beating. See text for explication. Why is the second wide QRS complex of each cycle different from the first wide QRS? Close inspection again is the key and reveals that immediately following each premature wide QRS complex, there is a retrograde P wave that resets the sinus node and causes the ensuing P′-P interval to be longer than the usual P-P interval. Before the next sinus P wave can be conducted to the ventricles, the atrioventricular junction escapes, causing a fusion beat with the simultaneously discharging parasystolic focus. Parasystole results from a protected, spontaneously discharging focus in the atria, atrioventricular junction, or, as in this case, the ventricles. This protection from the dominant rhythm, here sinus rhythm, is demonstrated by interectopic intervals that are multiples of a common denominator (1.10 seconds in this patient) and by variable coupling of the ectopic impulses with the dominant rhythm (1). Fusion beats are seen frequently, and parasystolic complexes fail to appear when their exit from the parasystolic focus is blocked by refractoriness of the surrounding tissue, here ventricular refractoriness following the first sinusinitiated complex of each cycle. Although this electrocardiogram is too short to make a definitive determination, the first wide QRS complex of each cycle appears to have fairly constant coupling to the preceding sinus- initiated QRS: 0.41, 0.40, and 0.36 seconds for the 3 complexes seen. Even fixed coupling would not exclude parasystole in this case. Because the sinus rhythm is fixed to the parasystolic rhythm by the retrograde P waves that reset the sinus node, as long as the parasystolic rate and the sinus rate each remain constant, which is not unusual for short intervals, the parasystolic impulses will appear to be coupled to the sinus impulses (1). Other electrocardiograms in the same patient often answer many questions. This patient's prior and subsequent electrocardiograms frequently reveal ventricular parasystole of the more typical variety with the parasystolic QRSs not fixedly coupled to the sinus rhythm because no retrograde P waves couple the sinus rhythm to the parasystolic rhythm. Furthermore, the parasystolic QRS complexes in those electrocardiograms all have the identical right-bundle-branch-block morphology seen here. Parasystole usually is considered a rare phenomenon, but its prevalence depends on how it is sought and by whom. Routine electrocardiograms have yielded a prevalence of 0.13% (2), while the prevalence among patients, who may have >1 electrocardiogram to review, is greater (0.3%) (3). When Holter monitoring is used for the search, the prevalence is much higher (1). The parasystolic focus is more often in the ventricles (>50%) than in the atria (20%) or atrioventricular junction (20%) (2). Although most examples of parasystole have been reported in patients with heart disease, many persons with apparently normal hearts have these rhythms (1). Parasystolic rhythms per se appear to carry no significant risk. A repetitive pattern, or group beating, of which a bigeminal rhythm is the simplest example, has many causes, which include most of the abnormalities of impulse formation and impulse conduction (4). Because parasystolic rhythms usually are independent of the dominant cardiac rhythm, they do not often present as group beating. In this example, group beating occurs not because the parasystolic rhythm is coupled to the dominant rhythm, but because sinus rhythm is coupled to the parasystolic rhythm through the mechanism of ventriculoatrial conduction resetting the sinus node.

Software designed to meet the varied and changing medical requirements of real-time arrhythmia analysis, data storage, and scientific reanalysis was developed. The incorporated define function allows the specification of various rhythm patterns and relationships (e.g. ventricular parasystole), which can be analyzed by the system. The results suggest that the program is capable of detecting complex clinical arrhythmias. >

This report is a brief survey of some results using feedback to control two diverse models of cardiac electrical activity. A chaotic quadratic map has been used as a model of ventricular fibrillation. Here we describe simulations to control this map using a self-tuning feedback controller. We also demonstrate control of a one-dimensional circle map which has been used as a model of the cardiac arrhythmia known as modulated ventricular parasystole. We show that feedback control offers a straightforward and viable alternative for stabilization of these model systems in comparison to OGY or other control methods.

Ventricular parasystole has been known to be a benign rhythm. We present a case of a 53-year-old man with chest pain and ventricular parasystole on electrocardiogram upon initial presentation. He was admitted and found to have normal serial cardiac enzymes and nuclear stress testing. He was discharged home with primary care follow-up. We highlight this case, as ventricular parasystole has been shown to be associated with cardiovascular disease and is a rhythm that should not be ignored. Our review of the literature shows that a thorough cardiovascular investigation should take place when this rhythm is identified.

This article explains the mechanism of “irregular” parasystole. Two theories have been suggested: “electrotonic modulation” and “type I second degree entrance block.” This study attempts to clarify the mechanism of irregular parasystole in cases of true ventricular parasystole associated with ventricular parasystolic couplets. Cases associated with ventricular parasystolic couplets were selected from 37 clinical cases of true ventricular parasystole in which one or more pure parasystolic cycles with no intervening nonectopic QRS complexes were found. Of the 37 cases of true ventricular parasystole, ventricular parasystolic couplets were found in 4 cases. In none of the other 33 cases, ventricular parasystolic couplets were found. In all the cases coexisting with ventricular parasystolic couplets, the latter ectopic QRS complex of the couplet failed to reset the parasystolic rhythm. The above findings suggest that the latter ectopic QRS complex of the parasystolic couplet originated not in the parasystolic pacemaker but in the pathway between the ventricle and the parasystolic pacemaker. It seems that when a sinus impulse fell late in the parasystolic cycle, it passed through the site of second degree entrance block and that the parasystolic couplets originated from the reentrant pathway between the ventricle and the pacemaker. This strengthens our previous suggestion that the mechanism of irregular parasystole is governed by “type I second degree entrance block” and not by “electrotonic modulation.”

Tachographic Discrimination of Parasystole. Introduction: To discriminate ventricular parasystole from fixed coupling interval ventricular premature complexes (VPCs), we developed a new diagnostic method using a dot distribution pattern corresponding to VPCs recorded on a heart rate tachogram using ambulatory ECG monitoring data. We tested our hypothesis that widely scattered VPC dots on instantaneous heart rate tachograms indicate a constant VPC-VPC interval compatible with parasystole. Methods and Results: Patients with frequent VPCs > 5,000/day) were divided into two groups depending on the tachogram dot distribution patterns: group F (n = 10, aged 61 ± 16 years) showed widely scattered VPC dot distribution, whereas group F (n = 10, 60 ± 17 years) showed fixed VPC dot distribution limited to a narrow zone. Using digitized R-R interval data, full-day heart rate tachograms and VPC-VPC intervals were depicted simultaneously. Group S demonstrated constant basic VPC-VPC intervals (1,285 to 2,052 msec, mean 1,738 ± 219), with a coefficient of variation (CV) of 0.061 ± 0.018. Their VPC coupling intervals were markedly variable (651 ± 113 msec; CV = 0.193 ± 0.034). Each patient's basic VPC-VPC intervals showed small diurnal alterations (minimum -13%± 3% to maximum +15%± 6%). VPC-VPC intervals in group F were not constant and showed marked variation. Group F VPC coupling intervals were shorter and constant (480 ± 30 msec, P = 0.0002; with CV = 0.076 ± 0.013, P < 0.0001). Conclusion: Ventricular parasystole with constant VPC-VPC intervals consistently became evident based on VPC dot patterns recorded on heart rate tachograms.

In recently reported cases of ventricular parasystole, it was shown that after exercise the parasystolic cycle length is prolonged, in contrast to a shortening of the sinus cycle length, whereas during standing the parasystolic cycle length and the sinus cycle length both shortened. In this report, to explore whether the same features as occur in ventricular parasystole are seen in atrial parasystole, effects of exercise and standing on the parasystolic cycle length were investigated in two men with atrial parasystole. The atrial parasystolic cycle length was prolonged after exercise, whereas it shortened during standing, similar to what occurs in ventricular parasystole. This is the first report to show such changes of cycle length in atrial parasystole. These findings suggest that in atrial parasystole, as in ventricular parasystole, influences on the parasystolic cycle length do not always act in the same direction as those on sinus cycle length.

A report of modulated ventricular parasystole is described in a patient presenting with acute myocardial infarction and classical (nonmodulated) ventricular parasystole. After infusion of intravenous lidocaine, the pattern of QRS distribution changed to one of modulated parasystole. This observation suggests that lidocaine altered the receptivity to electrotonic depolarizations and changed the slope of phase 4 depolarization.

A 53-year-old patient had a complex ventricular arrhythmia, which most likely was an intermittent pure (nonmodulated) parasystole, due to protection limited to the first part of the cycle coexisting with exit block. However, other interpretations of the observations were possible. Foremost among these was modulated parasystole with partial protection or with an attenuated or nondetectable early delaying phase, with exit block. Interestingly, the parasystole appeared to have fulfilled the dynamic rules regarding the number of sinus beats interposed between manifest parasystolic beats, as has been reported for pure or weakly modulated parasystole. This case corroborates unusual manifestations of an arrhythmia, which because of its newly found complexities and various possible interpretations seems to be discussed with decreasing frequency in most textbooks on general cardiology.

In two patients, ventricular parasystole (VP) was associated with ventricular tachycardia (VT), and in one patient, catheter ablation was successful. In patient 1, with dilated cardiomyopathy, VP led to VT, which converted to ventricular fibrillation. In patient 2, VP led to symptomatic nonsustained polymorphic VT. The origin of parasystolic focus was determined by endocardial mapping, and a radiofrequency current was delivered to patient 2. Both VP and VT disappeared immediately, and no recurrence has been observed during a follow-up of 8 months. Catheter ablation to the parasystolic focus was effective and a relationship between VP and VT was strongly suggested.

The cases of two patients with ventricular parasystole revealed a new depressive effect of adenosine on ventricular parasystolic activity. To the authors' knowledge, this is the first report that clearly demonstrates the disappearance of "true" parasystole.

Until recently, it had not been recognized that predictions regarding the number of sinus beats interposed between two consecutive parasystolic beats could be made. In a case of perfect, pure parasystole resulting from unintentional fixed rate ventricular pacing, the following was observed: there were consistently three different values (0,2,3) for the number of interposed sinus beats; only one of these values was odd, and the sum of the two smaller values was one less than the larger value. Our findings, which are in keeping with those obtained in an mathematical model, may be of additional help in the diagnosis of this elusive arrhythmia.

This presentation reflects the analysis of an electrocardiographic recording obtained from a patient with hypertensive heart disease. In the initial section of the tracing, fixed coupled monomorphic ectopic ventricular beats occurred in regular trigeminal rhythm. The pattern changed following an atrial extrasystole and several ventricular ectopic beats of various configuration occurred, often in sequence. Analysis demonstrated the presence of three independent parasystolic rhythms, two of which manifested with the character of intermittency, namely they were occasionally reset by extraneous impulses. The interplay of the sinus pacemaker with three parasystolic foci resulted in a very complex arrhythmic pattern. In some periods, however, two out of three ectopic rhythms were inapparent, and the third one manifested with fixed coupled complexes, so that a regular extrasystolic trigeminy ensued, and parasystole was not recognizable.

Electrocardiograms taken from 11 patients in sinus rhythm with ventricular ectopic rhythms from two different foci were analyzed to find the number of sinus beats, S, between the ectopic rhythms (S values). Three out of 11 patients had the S values typical for concealed ectopic rhythms. One of them had concealed bigeminy of 2n-1 form that occaslonally shifted to 2n form. Following the shift, S values of 2n-1 form were always achieved by the occurrence of double ventricular ectopic rhythms in succession. Concealed trigeminy of 3n and 3n-2 form was seen in the other two patients. Double ventricular ectopic rhythms had bizarre abnormal QRS complexes of two different morphologies and were inscribed in opposite directions. Ectopic rhythms in each case had parasystolic characteristics. These observations suggest bifocal automaticity as a mechanism for bidirectional ventricular tachycardia.

Previous reports dealing with the dynamics of “pure” parasystole using mathematic and electronic (pacing-induced) models, respectively, showed that a number of predictions emerged naturally from mathematic analysis. 1,2 Therefore, it appeared of interest to determine whether they would also apply to a series of patients with clinical (apparently) “pure” ventricular parasystole who were monitored for relatively long periods of time. In contrast to previous studies, Glass et al 1 had the ingenious idea of analyzing the number of sinus beats interposed between 2 consecutive manifest parasytolic beats. They observed that, with the rates being practically constant, a detailed mathematic description could be made using 2 parameters: the ratio of ectopic cycle length to sinus cycle length and the ratio of ventricular refractory period to sinus cycle length. From a correlation of these parameters 3 basic rules emerged: (1) There are at most 3 different values for the number of sinus beats in between parasystolic beats. (2) Only 1 of these values is odd. (3) The sum of the 2 lesser values is 1 less than the larger value.

Communications for this section will he published as space and priorities permit. The comments should not exceed 350 words in length, with a maximum of five references; one figure or table can be printed. Exceptions may occur under particular circumstances. Contributions may include comments on articles published in this periodical, or they may be reports of unique educational character. Specific permission to publish should be cited in a covering letter or appended as a postscript.