Irwue et al. September 1989

American Heart .kwrna~

onstration of the mechanism of transient entrainment and in- teruption of ventricular tachycardia with rapid atria1 pacing. J Am Co11 Cardiol 1984;3:422-30.

3. Brugada P, Wellens HJJ. Entrainment as an electrophysio- logic phenomenon. J Am Co11 Cardiol 1984;3:451-4.

4. Feld GK. Characterization of atria1 flutter reentry circuit in man by response to atria1 pacing [Abstract]. Circulation 1987; 76:TV-176.

8.

9.

5. Hirao K, Suzuki F, Kawara T, et al. Paradoxical atria1 capture in atria1 flutter. Cardiac Pacing 1988:4:157 ( in Jananese ). 10.

6. Almendral JM, Gottlieb CD, kosenihal ME, Stimato NJ, Buxton AE, Marchiinski FE, Miller JM, Josephson ME. En- trainment of ventricular tachycardia: explanation for surface electrocardiographic phenomenon by analysis of electrograms recorded within the tachycardia circuit. Circulation 1988; 77:569-80.

11.

12.

7. Henthorn RW, Okumura K, Olshansky B, Plumb VJ, Hess PG, Waldo AL. A fourth cirterion for transient entrainment:

the electrogram equivalent of progressive fusion. Circulation 1988;77:1003-12. Rsenblueth A, Garcia Ramos J. Studies on flutter and fibril- lation: the influence of artifical obstacles on experimental au- ricular flutter. AM HEART J 1947;33:677-84. Inoue H, Matsuo H, Takayanagi K, Murao S. Clinical and ex- perimental studies of the effects of atrial extrastimulation and rapid pacing on atria1 flutter cycle: evidence of macro-reentry with an excitable gap. Am J Cardiol 1981;48:623-31. Gallagher JJ, Kasell J, Sealy WC, Pritchett ELC, Wallace AG. Epicardial mapping in the Wolff-Parkinson-White syndrome. Circulation 1978;57:854-66. Wallenstein S, Zucker C, Fleiss J. Some statistical methods useful in circulation research. Circ F&s 1980;47:1-9. Boineau JP, Schuessler RB, Mooney CR, Miller CB, Wylds AC, Hudson RD, Borremans JM, Brockus CW. Natural and evoked atria1 flutter due to circus movement in dogs. Am J Cardiol 1980;45:1167-81.

Limitations of cough in maintaining blood flow during asystole: assessment by two-dimensional and Doppler echocardiography

It has been previously demonstrated that patients can maintain consciousness by vigorously coughing during episodes of prolonged ventricular asystole or ventricular fibrillation. Using two-dimensional Doppler echocardiography and arterial blood pressure recordings, we evaluated the changes that occur during coughing in nine patients who were dependent on pacemakers and in whom periods of asystole could be induced. During asystole in each p’atient the mitral valve stayed partially open and the left ventricular area comprised 87% and 90% (apical four-chamber view, short-axis view, respectively) of the area during diastole of paced rhythm. Peak arterial pressure during coughing and asystole was 80% of peak systolic pressure of paced rhythm (P = 0.001). Left ventricular area during coughing was 110% (both apical four-chamber view and short-axis views) of the area during asystole when patients were not coughing. Both the mitral and aortic valves showed no appreciable motion during coughing. During coughing the mean flow velocity across the mitral valve was 0.06 2 0.03 m/set and the flow velocity integral was 0.03 + 0.02 cm. Thus coughing during asystole produced minimal flow despite a rise in brachial arterial pressure. (AM HEART J 1989;118:474.)

Andrew Cohen, MD,* John Gottdiener, MD, Marc Wish, MD, and Ross Fletcher, MD. Washington, D. C.

Kouwenhoven et al.’ demonstrated in 1960 that ex- ternal chest compression was effective in maintaining

From the Veterans Administration Medical Center and Georgetown Uni-

versity, Washington, DC.

Received for publication Jan. 13, 1989; accepted Apr. 24, 1989.

*Current address and reprint requests: Andrew Cohen, MD, Director of Electrophysiology, Rose Medical Center, 4567 East 9th Ave., Denver, CO 80220.

4/l/13636

474

circulation during cardiac arrest. He postulated that sternal compression resulted in direct cardiac com- pression between the sternum and the spine. In the cardiac compression model, an arteriovenous pres- sure gradient is present in intrathoracic vessels and in the heart. However, canine experiments by Weale and Rothwell-Jackson in 1962 revealed that there was no arterial-venous pressure gradient during ex- ternal chest compression. Several subsequent obser- vations were also inconsistent with the idea that

Volume 118

Number 3 Cough-effected hemodynamics during asystole 475

chest compression was causing cardiac compression. Studying human subjects, MacKenzie et aL3 demon- strated high right atria1 pressures equal to systemic arterial pressure during chest compression. Further- more, Weisfeldt and HaIperin4 found that chest compression was ineffective in patients with flail chests, even though these patients would be most likely to experience cardiac compression during chest compression. Chandra et aL5 reported that simulta- neous chest compression and ventilation augmented mean arterial blood pressure and carotid flow velocity.

Criley et a1.6 noted that, with forceful, abrupt coughing, three patients remained awake in the catheterization laboratory despite ventricular fibril- lation that lasted up to 39 seconds. Later, Niemann et a1.7 reported on four additional patients in the catheterization laboratory or in the cardiac care unit who developed unstable rhythms (ventricular fibril- lation, asystole, or heart block). The patients main- tained consciousness by vigorous coughing for peri- ods of up to 92 seconds.

Niemann et aL7 also performed a series of animal experiments using angiography and hemodynamic measurements that showed that blood flow occurred because of changes in intrathoracic pressures. These observations led to the thoracic pump theory, which postulates that increases in intrathoracic pressure from coughing or chest compression are transmitted to intrathoracic vessels and to the heart. Because of high systemic venous compliance and because veins in the thoracic inlet have valves, the intrathoracic pressure is not transmitted through the venous sys- tem. The intrathoracic arteries transmit the pressure to the extrathoracic vessels and an arteriovenous pressure gradient that allows flow to occur is estab- lished. Hence the heart acts as a passive conduit for blood flow.

In the thoracic pump model, the mitral valve should stay open during chest compression or cough- ing and transmitral and transaortic blood flow should occur simultaneously. However, if cardiac compres- sion is the mechanism for blood flow, the mitral valve should remain closed during compression, and in this case flow across this valve would not be simultaneous with transaortic Aow.

In two echocardiographic studies of chest compres- sion in human subjects, 8-g the mitral valve remained open for at least part of the compression phase. However, none of the observed patients survived re- suscitative efforts. An animal study10 in which echocardiography was used to evaluate mitral valve motion during chest compression revealed mitral valve closure to be a determinant of successful resus- citation. A more recent animal study” showed that

the mitral valve closed rapidly during chest compres- sion with rapid chest compressions but not with slow, prolonged compressions. Because coughing may be a model for evaluating bloodflow during asystole, we used two-dimensional Doppler echocardiography to evaluate changes in cardiac chamber size, valve posi- tion, and flow velocity as well as arterial pressure in pacemaker-dependent patients during episodes of asystole before and during coughing.

METHODS

Patients consented to be studied under a protocol approved by the Institutional Review Board of Veterans Administration Medical Center. The study population consisted of nine patients (ages 59 to 77 years, mean 71 years) with complete heart block or a high degree of atri- oventricular block with slow or inapparent escape rhythms. Eight had permanent pacemakers and one had a temporary pacemaker wire in place pending the insertion of a perma- nent pacemaker. Echocardiographic evaluation was done with a commercially available phased array two-dimen- sional Doppler echocardiograph (Hewlett-Packard, Palo Alto, Calif.).

An indwelling catheter was placed in the brachial artery of each patient to record pressure measurements. Patients were instructed to cough forcefully once per second as ar- terial pressure was monitored. Pacemakers were inhibited (by direct programming or by skin pacing with an external temporary pacemaker) to induce asystole for periods of up to eight seconds. Echocardiographic recordings were then made during asystole and coughing. One to two seconds of asystole (in the absence of coughing) was recorded, and then a period of 3 to 5 seconds of forceful coughing was re- corded. Patients were allowed to rest for a least 1 minute between episodes of coughing during pacemaker inhibition. Coughing was considered adequate for subsequent analy- sis when peak brachial pressure during coughing was at least 90 mm Hg. Two-dimensional echocardiographic re- cordings were attempted in the short-axis, long-axis, and four- and five-chamber views in each patient.

Measurements of left ventricular cross-sectional area were taken in the parasternal short-axis view at the level of the papillary muscles and in the apical four chamber view. Measurements were done immediately before coughing, in the middle of the cough, and during diastole. Area mea- surements were done after the video image was digitally converted for analysis by a commercially available imaging system (Microsonics, Inc, Indianapolis, Ind.). Measure- ments of two beats were averaged for each view.

Mitral and aortic valve motion was observed by means of both two-dimensional and M-mode echocardiography in all planes of view. Care was taken to ensure that the echo sector intersected the valve at its maximum excursion which was determined just before asystole.

Pulsed wave Doppler recordings were made of transmi- tral blood flow during asystole and during coughing. An apical recording position was used, and a 400 Hz filter was used. The sample volume (gate length, 4 mm) was placed

476 Cohen et al. September 1989

American Heart Journal

Fig. 1. Three electrocardiographic leads, a brachial arterial pressure recording, and a phonogram from the chest wall. After three paced beats, the pacemaker is inhibited and aortic diastolic pressure falls to 36 mm Hg. A cough occurs (open arrow) causing a rise in arterial pressure to 98 mm Hg. A second cough is simul- taneous with an escape QRS followed by a second ventricular depolarization that causes a small rise in ar- terial pressure. A third cough soon after the QRS generates an arterial pressure of 108 mm Hg.

in the left ventricular inflow tract. Doppler recordings were made both on videotape and on a stripchart recorder at paper speeds of 50 to 100 mm/set. The Doppler recordings from the paper were measured with a planimeter on a dig- itizing tablet and calibrated for time and velocity. Analy- ses of mean flow velocity and flow velocity integral were performed with the Microsonics system. In addition, Dop- pler recordings were attempted for aortic blood flow with continuous-wave Doppler from either a suprasternal or an apical approach.

Data were analyzed with analysis of variance (ANOVA) and paired t tests with p < 0.05 considered significant. Values reported are means plus or minus standard devia- tion.

RESULTS

In all nine patients, asystolic periods of at least 3 seconds could be induced by pacemaker inhibition. The duration of asystole uninterrupted by escape beats was 5.6 f 1.5 seconds. All patients were able to cough forcefully enough to achieve peak brachial pressures of at least 90 mm Hg during asystole.

Values for left ventricular areas are reported for only those patients in whom the endocardium was adequately visualized for the two comparative views reported. The two sets of paired data points are: (1) change from diastole to asystole and (2) change from asystole to state induced by coughing.

Asystole before coughing. Mean aortic diastolic pressure before and during paced rhythm was 74 + 10 mm Hg compared with 50 f 10 mm Hg during asys- tole before the first cough in the sequence (Fig. 1) and 42 + 12 mm Hg before the third cough in the sequence. Both the mitral and aortic valves were vi- sualized during coughing and asystole. In all patients

the mitral and tricuspid valves floated in a partially open position during asystole. The mitral and tricus- pid valves opened during atria1 systole in patients who were in sinus rhythm and in one patient who had a chronic atria1 tachyarrhythmia. The aortic valve appeared to be closed throughout the asystolic period with minimal motion during atria1 systole.

The left ventricular cross-sectional area in the short-axis view measured at the papillary muscle level was 15.1 ? 7.1 cm2 in diastole and 14.2 rt 6.2 cm2 in asystole before coughing (not significant [NS]) (Six patients had paired data.) Similarly, left ven- tricular area measured in the apical four-chamber view was 31.5 t 5.8 sq cm during diastole, compared to 27.0 it 8.8 sq cm. during asystole. Six patients had paired data (not significant with ANOVA).

Cough during asystole. The peak brachial arterial blood pressure of all nine patients during the initial cough in asystole was 80 % of the systolic pressure during paced rhythm (mean 128 f 29 mm Hg during coughing, 159 f 22 mm Hg during paced rhythm; p =O.OOl). Cough-generated brachial artery pressures gradually fell after the initial cough in most patients. (The mean of the peak arterial pressure of the third cough was 80% of the pressure of the first cough [p =O.OOl]). Coughs occurring immediately after ventricular depolarization caused increments in bra- chial artery pressure equal to those during asystole (Fig. 1).

All patients experienced dizziness during the coughing stage of the study, even though episodes did not last more than 8 seconds. The mitral and tri- cuspid valves did not appreciably change positions during coughing. Similarly, the aortic valve did not

Volume 118

Number 3 Cough-effected hemodynamics during asystole 477

Fig. 2. A two-dimensional echocardiograph of the heart in a parasternal, long-axis view. In the bottom left-hand corner, the electrocardiogram is displayed showing two cough artifacts (the first after 4 seconds of asystole) with the frame stopped immediately after the second cough. The aortic valve appears closed.

open during coughing (Fig. 2). In one patient, the aortic valve opened immediately after coughing (Fig. 3).

The cross-sectional area of the left ventricle during coughing was 19.6 + 5.1 cm2 during coughing com- pared with 17.8 + 4.0 cm2 during asystole (NS) (four patients). Apical four-chamber measurement of the left ventricle during coughing was 31.3 + 7.7 cm,2 compared with 27.0 f 8.8 cm2 during asystole (NS, with ANOVA) (seven patients). Pulsed-wave Dop- pler recordings of flow across the mitral valve during coughing revealed a mean flow velocity of 0.06 + 0.03 m/set and a flow velocity integral of 0.03 + 0.02 cm (Fig. 4). Doppler recordings from the aorta could not be obtained during coughing because of artifact pro- duced by the coughing.

DISCUSSION

The mechanisms that cause forward blood flow during chest compression remain in dispute. Studies of both animals and human subjects have provided evidence for both the cardiac compression model and the thoracic pump model. A recent animal studyll provided support for direct cardiac compression model when high-velocity-high-impulse chest com- pression was used and support for the thoracic pump model when low-impulse compressions were em- ployed.

Cough-generated flow, first described by Criley et

Fig. 3. Parasternal short-axis m-mode recording through aorta. The aortic valve and the left atrium, with the elec- trocardiogram and brachial artery tracings are also shown. The third and fourth coughs of the sequence are shown. The first cough illustrated generates a brachial artery pressure of 152 mm Hg (first open arrow). After the fall in arterial pressure, the aortic valve is seen opening (first closed arrow). The second cough (second open arrow) gen- erates an arterial pressure of 119 mm Hg. The aortic valve is again seen opening (second closed arrow) only after the fall in arterial pressure. BA-Brachial artery.

a1.,6 was believed to be a model that may provide an explanation for blood flow during chest compression by means of manual cardiopulmonary resuscitation. In his original paper, he described three patients who maintained consciousness despite ventricular fibril- lation by coughing vigorously. Neimann7 reported an additional four patients who had undergone hemo- dynamic study during coughing and unstable rhythms. One patient had a prosthetic aortic valve that opened during cinefluoroscopy at the time of coughing. Two patients had Doppler velocity probes over their brachial arteries, which showed flow- velocity signals equivalent to those shown by ven- tricular contraction. Thus, of the seven patients re- ported, one had antegrade flow through the heart (documented by fluoroscopy of the aortic valve shown opening during cough) and one had flow that was perceived as the result of the patient’s remaining conscious for 92 seconds.

Niemann et a1.7 and Rosborough et a1.,i2 in a canine study, used vagal stimulation and one-way valves at- tached to the airways to induce coughing. Cough- generated aortic pressures averaged 100 mm Hg. Each cough caused antegrade aortic flow, as demon- strated by cineangiography,7p l2 and opening of both the aortic and mitral valves. In addition, carotid

478 Cohen et al September 1989

American Heart Journai

Fig. 4. Recordings of electrocardiogram and pulsed wave @TV) Doppler recording in left ventricular in- flow tract (from apical transducer site), and brachial artery pressure recording. In A, the pacemaker has been inhibited, and nonconducted p waves are seen (open arrowheads) causing flow across the mitral valve (closed arrows). In B, the atrially caused flow is still seen (closed arrows), but the cough-induced rise in brachial artery pressure to 140 mm Hg causes only minimal flow, as revealed in the Doppler sample.

blood flow l2 increased linearly with the pressure generated by each cough.

Little13 evaluated the hemodynamic effects of coughing in a group of patients in normal rhythm. Measurements were made of the effect of coughing on aortic and right atria1 pressures, brachial artery flow, and aortic valve motion (determined echocardio- graphically). Despite aortic peak pressures that were 50 mm Hg higher during coughing than during left ventricular systole and brachial artery flow during coughing that was 158% of that during ventricular systole, no aortic valve opening was noted in 200 coughing episodes. The authors speculated that blood was compressed out of the central aorta to the periphery without flowing through the heart. They believed that high aortic diastolic pressure relative to the left ventricular pressure precluded aortic valve opening. Their study (done in patients with normal sinus rhythm, mainly normotensive) contrasted with the previous reports of Niemann et a1.7 and Criley et a1.6 in which the subjects experienced asystole, high-

degree heart block, or ventricular fibrillation. In ad- dition, all were hypotensive.

As in the reports by Feneley et al.ll and Werner et a1.,g the present study shows that during the state with no flow (asystole in this study, ventricular fibrillation in the others) the mitral valve leaflets stay separated. During coughing there is minimal motion of the mitral valve leaflets, suggesting low flow or slow acceleration of blood through the left ventricular in- flow tract. Similarly, Doppler flow recordings across the mitral valve show minimal flow during coughing despite brachial arterial pressures as high as 200 mm Hg in one patient. The aortic valve appears to stay in a closed position during asystole. With coughing the valve shows no apparent motion, which is also con- sistent with a low-flow state.

Our data confirm those of Criley et a1.,6 Niemann et al.,7 and Little et a1.13 in documenting preservation of adequate arterial pressure during coughing. How- ever, we found, as did Little et al., that, despite the rise in arterial pressure, only minimal antegrade flow

Volume 118 Number 3 Cough-effected hemodynamics during asystole 479

occurs through the heart during vigorous coughing. It appears that coughing may cause antegrade flow

of blood from the pulmonic veins to the heart. There is minimal flow across the mitral valve, and the aor- tic valve stays closed, leading to an increase in left ventricular cavity dimensions. The rise in brachial artery pressure measured in this study and the bra- chial artery flow reported in other studies may rep- resent transmission of intrathoracic pressure through the arterial system, with a static blood column effect, causing to-and-fro motion rather than forward vol- ume flow only. Thus, although there is maintenance of systemic blood pressure during coughing, ante- grade flow may be minimal. Patients may be able to maintain consciousness because the coughing may preserve systemic arterial pressure. The factor that limits consciousness may be the oxygen content of the static blood volume.

There are two main differences between this study and the previous reports by Criley et a1.6 and Nie- mann et a17. First of all, the patients reported by Criley and Niemann had hypotension for prolonged periods (up to 92 seconds) in contrast to our patients who had asystole or hypotension of short duration ( up to only 8 seconds). Second several patients in the other reports and all of the animals studied were in ventricular fibrillation. For the subjects in this study, it is possible that the left ventricular pressure (which was not measured) is lower than the aortic pressure during these brief hypotensive episodes. Since cough- ing causes approximately equal rises in all intra- thoracic pressures (i.e., intracardiac and aortic), equal pressure rises in the left ventricle and aorta may cause a gradient insufficient to open the aortic valve; thus no transcardiac flow occurs. However, in the patient who had documented prosthetic valve opening during coughing, as reported by Niemann et al7 the period of asystole was less than 3 seconds long similar to the duration of asystole in our study. In the canine experiments of the same report, left ventric- ular and aortic pressures equalized within 2 seconds of the onset of ventricular fibrillation. However, there may be differences in the rate of change of left ven- tricular pressure during asystole as opposed to dur- ing ventricular fibrillation.

Limitations of this study. The patients were elderly and may have been unable to generate intrathoracic pressures equal to those of younger patients. How- ever, the mean of the peak arterial pressures achieved by our patients (128 mm Hg) was comparable to that reported by Criley et a1.6 (139 mm Hg).

Also, Doppler flow signals are dependent on the adequate positioning of the beam. Hence, because of chest motion during coughing, accurate flow velocity

measurements may not have been obtained despite care taken to aim the Doppler signal as precisely as possible during simultaneous imaging with two- dimensional echocardiography. However, consistent, good-quality Doppler flow signals caused by atria1 contractions were observed during asystole and coughing, and the lack of valve motion and the absence of symptoms were consistent with the poor flow recorded by the Doppler echocardiograph.

Another limitation is the possibility that measure- ments of the left ventricular cavity in various planes may no accurately represent left ventricular volume. Left ventricular area measurements may be decep- tive if coughing caused rotation or distortion of the ventricle. However, when internal landmarks were observed, only minor rotation of the heart was noted. Area measurements were taken from views where in- ternal landmarks such as papillary muscles and valves were present, and multiple planes of imaging were used.

REFERENCES

1. Kouwenhoven WB, Jude Jr., Knickerbocker GC. Closed chest cardiac massage. JAMA 1960;173:1064-7.

2. Weale FE, Rothwell-Jackson RL. The efficiency of cardiac massage. Lancet 1962;1:990-2.

3. MacKenzie G, Taylor S, McDonald A, Donald K. Hemody- namic effects of external cardiac compression. Lancet 1964; 1:1342-5.

4. Wesifeldt ML, Halperin HR. Cardiopulmonary resuscitation: beyond cardiac massage. Circulation 1986;74:443-8.

5. Chandra N, Rudikoff M, Weisfeldt ML. Simultaneous chest compression and ventilation at high airway pressure during cardiopulmonary resuscitation. Lancet 1980;1:175-8.

6. Criley JM, Blaufuss AH, Kissel GL. Cough induced cardiac compression; self L ‘ministered form of cardiopulmonary re- suscitation. JAM.: ! 976;236:1246-50.

7. Niemann JT, Roei iii[ ough J, Hausknecht M, et al. Cough CPR: documentation of. stemic perfusion in man and in an exper- imental model: a w-indow to the mechanism of blood flow in external CPR. Crit Care Med 1980;8:141-6.

8. Rich S, Wix HL, Shapiro EP. Clinical assessment of heart chamber size and valve motion during cardiopulmonary re- suscitation by two-dimensional echocardiography. AM HEART J 1981;102:368-3.

9. Werner JA, Greene HL, Janko CL, Cobb LA. Visualization of cardiac valve motion in man during external chest compres- sion using two-dimensional echocardiography. Circulation 1981;63:1417-21.

10. Deshmuhk HG, Weil MW, Rackow EC, Trevino R, Bisera J. Echocardiographic observations during cardiopulmonary re- suscitation: a preliminary report. Crit Care Med 1985;13: 904-6.

11. Feneley MP, Maier GW, Gaynor W, et al. Sequence of mitral valve motion and transmitral blood flow during manual cardiopulmonary resuscitation in dogs. Circulation 1987; 76:363-75.

12. Rosborough JP, Hausknech M, Niemann JT, Criley JM. Cough supported circulation. Crit Care Med 1981;9:371-2.

13. Little WC, Reeves RC, Coughlan HC, Rogers EW: Effects of cough on coronary perfusion pressure: does coughing help clear the coronary arteries of angiographic contrast medium? Circulation 1982;65:604-10.