Pediatric Cardiology
Pediatr Cardiol 13:141-145, 1992
9 Springer-VerlagNew York Inc. 1992
Left Ventricular Performance in Neonates on Extracorporeal Membrane Oxygenation Farhouch Berdjis, l* Masato Takahashi, 2 and Alan B. Lewis 2 ~Department of Pediatrics, University of Southern California School of Medicine and '-Division of Cardiology, Childrens Hospital of Los Angeles, Los Angeles, California, USA
SUMMARY. The evaluation of left ventricular systolic performance in infants undergoing extracorporeal membrane oxygenation (ECMO) using traditional ejection-phase indices is hampered by significant alterations in preload and afterioad. Therefore, a load-independent index, which relates heart-rate-corrected mean velocity of circumferential fiber shortening (VCFc) to afterioad, measured as end-systolic wall stress (ESS), was used to assess left ventricular function in 18 term neonates undergoing ECMO. The mean age at the onset of ECMO was 75.5 h and the duration of therapy was 171 +- 106 h. Left ventricular performance was highest before the onset of ECMO (VCFc = 1.65 - 0.49 circ/s) and decreased toward normal during (1.38 --+ 0.33 circ/s) and following ECMO (1.29 -_+ 0.16 circ/s). Initially, nine of 17 (53%) patients had enhanced performance for the degree of afterload but in only 16 of 48 (33%) studies during ECMO and none following ECMO was VCFc elevated beyond the normal range predicted for ESS. These changes in left ventricular performance may be the result of variations in exogenous, as well as endogenous, catecholamines rather than intrinsic alterations in myocardial contractility. It is concluded that the VCFc/ESS relation permits a meaningful assessment of ventricular performance in critically ill neonates undergoing ECMO. KEY WORDS: Ventricular function - - Neonate - - Extracorporeal membrane oxygenation Afterload
Extracorporeal membrane oxygenation (ECMO) has been used with increasing success as therapy to provide cardiopulmonary support for critically ill neonates with persistent pulmonary hypertension and/or meconium aspiration [1, 19]. Recent reports [17, 24] have documented transient decreases in left ventricular shortening fraction, velocity of circumferential fiber shortening (VCF), and ventricular output along with increases in blood pressure and wall stress. These ejection-phase indices of ventricular performance are dependent upon loading conditions and heart rate and, thus, unreliable in patients in whom preload and afterload are altered [16, 18]. Recently, several investigators have employed a load-independent index of left ventricular contrac-
* Present address: Deutsches Herzzentrum Berlin, 1000 Berlin,
Federal Republic of Germany. Address offprint reprints to: Dr. Alan B. Lewis, Division of Car-
diology, Childrens Hospital of Los Angeles, 4650 Sunset Blvd., Los Angeles, CA 90027, USA.
tility, which relates myocardial performance to afterload at end-systole [2, 4-6, 25]. The present study was undertaken to evaluate serial left ventricular performance in a group of critically ill neonates undergoing ECMO using the VCF-end-systolic wall stress (ESS) relation.
Methods The study group consisted of 18 term neonates who underwent ECMO therapy. Serial echocardiograms were performed by the same investigator (FB) immediately prior to the initiation of ECMO, on each of the 3 successive days during ECMO and 1 day after termination of ECMO. A Hewlett-Packard ultrasonograph with a 5-MHz phased-array transducer was used for the echocardiographic and Doppler examinations and recorded, along with the umbilical arterial pressure tracing, on a Honeywell strip chart and 1/2 inch VHS video tape. The axial resolution of the transducer was 0.7 mm. A complete echocardiographic examination was performed initially to rule out a congenital cardiac malformation.
142
Pediatric Cardiology Vol. 13, No. 3, 1992
Table 1. Left ventricular performance indices in the study population Group VCFc (circ/s) 1 2 3
LVSF
dVCFc (circ/s)
ESS (g/cm 2)
1.65 _+ 0.49 0.35 -+ 0.07 0.39 --- 0.48 27.7 -+ 6.0 1.38 -+ 0.33 ~ 0.29 -+ 0.08 c 0.17 + 0.29 b 35.3 -+ 10.6 1.29 + 0.16 ~ 0.34 _+ 0.04 0.13 -+ 0.17 41.6 -+ 20.1 c
VCFc, heart rate-corrected velocity of circumferential fiber shortening; LVSF, left ventricular shortening fraction; dVCFc, difference between observed VCFc and the mean predicted for a normal infant population; ESS, end-systolic wall stress. ~ -< 0.02; bp = 0.05; ' p < 0.01.
proximately m0.21 circ/s (95% confidence limits). Values beyond this range represent enhanced or depressed ventricular performance for a given afterload. The data are expressed as mean -+ standard deviation (SD). Statistical analysis between groups was performed by Student's t test and repeated measures analysis of variance (ANOVA). When significant differences between groups were observed by ANOVA, Scheffe's and Bonferroni's tests were applied to determine which specific groups differed. A value of p < 0.05 was considered significant. The study protocol was approved by the Committee on Clinical Investigations of the Childrens Hospital of Los Angeles and informed consent was obtained from the parents of each patient.
Results
The left ventricular minor axis dimension generally is used for the calculation of VCF in individuals in whom the crosssectional shape of the left ventricle is circular [9]. However, the interventricular septum in neonates, particularly in the presence of pulmonary hypertension, is flattened or even convex toward the left, thereby distorting the shape of the left ventricle. Thus, the minor axis end-diastolic and end-systolic dimensions are invalid measures of left ventricular cavity diameter in such patients [20]. In order to compensate for the noncircular configuration of the left ventricle, the end-diastolic circumference [Ced, defined as the circumference at the onset of the Q wave of the electrocardiogram (ECG)] and end-systolic circumference (Ces, defined as the smallest cross-sectional circumference) were measured directly from the two-dimensional echocardiogram rather than calculated from the minor axis dimensions and substituted for the respective diameters in the calculation of VCF. The left ventricular ejection time (ET) was measured from the pulsed Doppler recording of the aortic valve opening and closure [11]. The mean VCF was calculated as: VCF = (Ced - Ces)/(Ced)(ET) Correction of the ejection time to a heart rate of 60 beats/min was achieved by dividing it by the square root of the R-R interval. Left ventricular end-systolic meridional wall stress was calculated as described by Grossman et al. [12]: ESS = 1.35(Pes)(Des)/(4(hes)[l + (hes/Des)] ESS is left ventricular end-systolic wall stress (g/cm2). Pes is end-systolic left ventricular pressure obtained from the umbilical arterial pressure tracing by interpolation of peak systolic and diastolic blood pressure at the time of aortic valve closure. Des is the left ventricular end-systolic diameter, which in the presence of a noncircular, neonatal left ventricle, was calculated by dividing Ces by 7r. The end-systolic thickness of the left ventricular posterior wall (hes) was calculated from the M-mode tracing by the leading edge to leading edge method [8] using a digitizing pad. The average of 3-5 cardiac cycles was used for all measurements. The difference between observed VCFc and the value predicted from the regression of Colan et al. [5] (dVCFc) was calculated in order to compare observed left ventricular performance relative to expected normal values for a given afterload. The normal variability for dVCFc around the predicted mean is ap-
Eighteen patients were enrolled in the study. The mean age at the onset of ECMO was 75.5 h. The 5min Apgar scores ranged from 4-9, with half the patients having a 5-rain Apgar score below 7. Thus, while many infants were in significant distress from birth, none exhibited signs of profound perinatal asphyxia (i.e., Apgar scores of ~ 9
,oo,
,
~
i
. . . . 2 - . , " - -,~-" ..,..,,,.
1,5
1.5
: ~
1.0
0.5
-10
m
m
i
i 9
l
9~
. . . .
*
.
.
.
.
.
". ' '. ~.
....
I
I
I
I
,I
20
30
4(?
50
60
.........
......
1.0
0.5 10
""-:~.... i "
I
I
I
I
I
20
30
40
50
60
ESS (gm/cm ~)
Fig. 1. Left ventricular shortening velocity related to wall stress in patients prior to the initiation of ECMO (group l). Dashed lines represent the 95% confidence limits of the normal range. Mean rate-corrected velocity of shortening (VCFc) was elevated in nine of 17 patients. ESS, end-systolic wall stress.
was the lowest (ESS = 27.7 + 6.0 g/cm 2, range = 15.0-39.5) of the three observation periods (Fig. 1). The difference between the observed VCFc and the shortening velocity predicted from the regression line for normal infants (dVCFc) similarly was highest for group 1 (dVCFc = 0.39 + 0.48 circ/s). Indeed nine of 17 (53%) patients in group 1 had a VCFc above the 95% confidence limits of the normal range. Thus, patients tended to have significantly enhanced left ventricular systolic performance prior to ECMO. During ECMO, left ventricular performance tended to decline toward the normal range as afterload increased with improved systemic blood pressure. Group 2 VCFc decreased to 1.38 +_ 0.33 circ/s (p = 0.02), while dVCFc declined to 0.17 +_ 0.29 circ/s (p - 0.05). The VCF was elevated above the predicted normal range in only 16 of 48 (33%) studies (Fig. 2). ESS increased to 35.3 + 10.6 g/cm 2 but the change was not significant (p = 0.09). Following weaning from ECMO, VCFc was within the normal range (1.29 + 0.16 circ/s, dVCFc = 0.13 -+ 0.17 circ/s) in all patients (Fig. 3). Mean ESS increased (p < 0.01) following ECMO but this was affected by the presence of systemic hypertension in two patients, which resulted in exceedingly high wall stress values of 85 g/cm 2 in each. All but one of the other patients had ESS values between 20 and 40 g/cm 2 and were within normal limits for newborns [5]. Left ventricular shortening fraction (LVSF) was a poor indicator of ventdcular performance. Mean L V S F declined following the initiation of ECMO from 0.36 -+ 0.07 to 0.29 -+ 0.08 (p < 0.01). Prior to ECMO, four of 17 patients in whom L V S F
ESS (gm/cm)
Fig. 2. Left ventricular shortening velocity evaluated during days 1-3 of ECMO (group 2). Ventricular performance declined toward normal as afterload increased. One patient had mildly reduced left ventricular function relative to afterload. Abbreviations as in Fig. 1.
3.0
2.5
~2.0
>~-
1.5 9
"
'
O"
"1
'
9
"
n
i i
~
"
. . . .
"
.
.
.
.
.
.
.
.
'
'
9
4,
,
. . . .
,
.
.
.
,
,
.
,
1.0 .
0.5' 10
.
.
.
.
,
,
,
,
I
f
I
f
I
I
I
20
30
40
50
60
80
100
ESS (gmlcm)
Fig. 3. Following discontinuation of ECMO, left ventricular systolic performance was normal in all survivors (group 3). No patients exhibited reduced ventricular function. Afterload was normal in 10 infants but markedly elevated in two patients with systemic hypertension 9 Abbreviations as in Fig. I.
was measured had values
Pediatr Cardiol 13:141-145, 1992
9 Springer-VerlagNew York Inc. 1992
Left Ventricular Performance in Neonates on Extracorporeal Membrane Oxygenation Farhouch Berdjis, l* Masato Takahashi, 2 and Alan B. Lewis 2 ~Department of Pediatrics, University of Southern California School of Medicine and '-Division of Cardiology, Childrens Hospital of Los Angeles, Los Angeles, California, USA
SUMMARY. The evaluation of left ventricular systolic performance in infants undergoing extracorporeal membrane oxygenation (ECMO) using traditional ejection-phase indices is hampered by significant alterations in preload and afterioad. Therefore, a load-independent index, which relates heart-rate-corrected mean velocity of circumferential fiber shortening (VCFc) to afterioad, measured as end-systolic wall stress (ESS), was used to assess left ventricular function in 18 term neonates undergoing ECMO. The mean age at the onset of ECMO was 75.5 h and the duration of therapy was 171 +- 106 h. Left ventricular performance was highest before the onset of ECMO (VCFc = 1.65 - 0.49 circ/s) and decreased toward normal during (1.38 --+ 0.33 circ/s) and following ECMO (1.29 -_+ 0.16 circ/s). Initially, nine of 17 (53%) patients had enhanced performance for the degree of afterload but in only 16 of 48 (33%) studies during ECMO and none following ECMO was VCFc elevated beyond the normal range predicted for ESS. These changes in left ventricular performance may be the result of variations in exogenous, as well as endogenous, catecholamines rather than intrinsic alterations in myocardial contractility. It is concluded that the VCFc/ESS relation permits a meaningful assessment of ventricular performance in critically ill neonates undergoing ECMO. KEY WORDS: Ventricular function - - Neonate - - Extracorporeal membrane oxygenation Afterload
Extracorporeal membrane oxygenation (ECMO) has been used with increasing success as therapy to provide cardiopulmonary support for critically ill neonates with persistent pulmonary hypertension and/or meconium aspiration [1, 19]. Recent reports [17, 24] have documented transient decreases in left ventricular shortening fraction, velocity of circumferential fiber shortening (VCF), and ventricular output along with increases in blood pressure and wall stress. These ejection-phase indices of ventricular performance are dependent upon loading conditions and heart rate and, thus, unreliable in patients in whom preload and afterload are altered [16, 18]. Recently, several investigators have employed a load-independent index of left ventricular contrac-
* Present address: Deutsches Herzzentrum Berlin, 1000 Berlin,
Federal Republic of Germany. Address offprint reprints to: Dr. Alan B. Lewis, Division of Car-
diology, Childrens Hospital of Los Angeles, 4650 Sunset Blvd., Los Angeles, CA 90027, USA.
tility, which relates myocardial performance to afterload at end-systole [2, 4-6, 25]. The present study was undertaken to evaluate serial left ventricular performance in a group of critically ill neonates undergoing ECMO using the VCF-end-systolic wall stress (ESS) relation.
Methods The study group consisted of 18 term neonates who underwent ECMO therapy. Serial echocardiograms were performed by the same investigator (FB) immediately prior to the initiation of ECMO, on each of the 3 successive days during ECMO and 1 day after termination of ECMO. A Hewlett-Packard ultrasonograph with a 5-MHz phased-array transducer was used for the echocardiographic and Doppler examinations and recorded, along with the umbilical arterial pressure tracing, on a Honeywell strip chart and 1/2 inch VHS video tape. The axial resolution of the transducer was 0.7 mm. A complete echocardiographic examination was performed initially to rule out a congenital cardiac malformation.
142
Pediatric Cardiology Vol. 13, No. 3, 1992
Table 1. Left ventricular performance indices in the study population Group VCFc (circ/s) 1 2 3
LVSF
dVCFc (circ/s)
ESS (g/cm 2)
1.65 _+ 0.49 0.35 -+ 0.07 0.39 --- 0.48 27.7 -+ 6.0 1.38 -+ 0.33 ~ 0.29 -+ 0.08 c 0.17 + 0.29 b 35.3 -+ 10.6 1.29 + 0.16 ~ 0.34 _+ 0.04 0.13 -+ 0.17 41.6 -+ 20.1 c
VCFc, heart rate-corrected velocity of circumferential fiber shortening; LVSF, left ventricular shortening fraction; dVCFc, difference between observed VCFc and the mean predicted for a normal infant population; ESS, end-systolic wall stress. ~ -< 0.02; bp = 0.05; ' p < 0.01.
proximately m0.21 circ/s (95% confidence limits). Values beyond this range represent enhanced or depressed ventricular performance for a given afterload. The data are expressed as mean -+ standard deviation (SD). Statistical analysis between groups was performed by Student's t test and repeated measures analysis of variance (ANOVA). When significant differences between groups were observed by ANOVA, Scheffe's and Bonferroni's tests were applied to determine which specific groups differed. A value of p < 0.05 was considered significant. The study protocol was approved by the Committee on Clinical Investigations of the Childrens Hospital of Los Angeles and informed consent was obtained from the parents of each patient.
Results
The left ventricular minor axis dimension generally is used for the calculation of VCF in individuals in whom the crosssectional shape of the left ventricle is circular [9]. However, the interventricular septum in neonates, particularly in the presence of pulmonary hypertension, is flattened or even convex toward the left, thereby distorting the shape of the left ventricle. Thus, the minor axis end-diastolic and end-systolic dimensions are invalid measures of left ventricular cavity diameter in such patients [20]. In order to compensate for the noncircular configuration of the left ventricle, the end-diastolic circumference [Ced, defined as the circumference at the onset of the Q wave of the electrocardiogram (ECG)] and end-systolic circumference (Ces, defined as the smallest cross-sectional circumference) were measured directly from the two-dimensional echocardiogram rather than calculated from the minor axis dimensions and substituted for the respective diameters in the calculation of VCF. The left ventricular ejection time (ET) was measured from the pulsed Doppler recording of the aortic valve opening and closure [11]. The mean VCF was calculated as: VCF = (Ced - Ces)/(Ced)(ET) Correction of the ejection time to a heart rate of 60 beats/min was achieved by dividing it by the square root of the R-R interval. Left ventricular end-systolic meridional wall stress was calculated as described by Grossman et al. [12]: ESS = 1.35(Pes)(Des)/(4(hes)[l + (hes/Des)] ESS is left ventricular end-systolic wall stress (g/cm2). Pes is end-systolic left ventricular pressure obtained from the umbilical arterial pressure tracing by interpolation of peak systolic and diastolic blood pressure at the time of aortic valve closure. Des is the left ventricular end-systolic diameter, which in the presence of a noncircular, neonatal left ventricle, was calculated by dividing Ces by 7r. The end-systolic thickness of the left ventricular posterior wall (hes) was calculated from the M-mode tracing by the leading edge to leading edge method [8] using a digitizing pad. The average of 3-5 cardiac cycles was used for all measurements. The difference between observed VCFc and the value predicted from the regression of Colan et al. [5] (dVCFc) was calculated in order to compare observed left ventricular performance relative to expected normal values for a given afterload. The normal variability for dVCFc around the predicted mean is ap-
Eighteen patients were enrolled in the study. The mean age at the onset of ECMO was 75.5 h. The 5min Apgar scores ranged from 4-9, with half the patients having a 5-rain Apgar score below 7. Thus, while many infants were in significant distress from birth, none exhibited signs of profound perinatal asphyxia (i.e., Apgar scores of ~ 9
,oo,
,
~
i
. . . . 2 - . , " - -,~-" ..,..,,,.
1,5
1.5
: ~
1.0
0.5
-10
m
m
i
i 9
l
9~
. . . .
*
.
.
.
.
.
". ' '. ~.
....
I
I
I
I
,I
20
30
4(?
50
60
.........
......
1.0
0.5 10
""-:~.... i "
I
I
I
I
I
20
30
40
50
60
ESS (gm/cm ~)
Fig. 1. Left ventricular shortening velocity related to wall stress in patients prior to the initiation of ECMO (group l). Dashed lines represent the 95% confidence limits of the normal range. Mean rate-corrected velocity of shortening (VCFc) was elevated in nine of 17 patients. ESS, end-systolic wall stress.
was the lowest (ESS = 27.7 + 6.0 g/cm 2, range = 15.0-39.5) of the three observation periods (Fig. 1). The difference between the observed VCFc and the shortening velocity predicted from the regression line for normal infants (dVCFc) similarly was highest for group 1 (dVCFc = 0.39 + 0.48 circ/s). Indeed nine of 17 (53%) patients in group 1 had a VCFc above the 95% confidence limits of the normal range. Thus, patients tended to have significantly enhanced left ventricular systolic performance prior to ECMO. During ECMO, left ventricular performance tended to decline toward the normal range as afterload increased with improved systemic blood pressure. Group 2 VCFc decreased to 1.38 +_ 0.33 circ/s (p = 0.02), while dVCFc declined to 0.17 +_ 0.29 circ/s (p - 0.05). The VCF was elevated above the predicted normal range in only 16 of 48 (33%) studies (Fig. 2). ESS increased to 35.3 + 10.6 g/cm 2 but the change was not significant (p = 0.09). Following weaning from ECMO, VCFc was within the normal range (1.29 + 0.16 circ/s, dVCFc = 0.13 -+ 0.17 circ/s) in all patients (Fig. 3). Mean ESS increased (p < 0.01) following ECMO but this was affected by the presence of systemic hypertension in two patients, which resulted in exceedingly high wall stress values of 85 g/cm 2 in each. All but one of the other patients had ESS values between 20 and 40 g/cm 2 and were within normal limits for newborns [5]. Left ventricular shortening fraction (LVSF) was a poor indicator of ventdcular performance. Mean L V S F declined following the initiation of ECMO from 0.36 -+ 0.07 to 0.29 -+ 0.08 (p < 0.01). Prior to ECMO, four of 17 patients in whom L V S F
ESS (gm/cm)
Fig. 2. Left ventricular shortening velocity evaluated during days 1-3 of ECMO (group 2). Ventricular performance declined toward normal as afterload increased. One patient had mildly reduced left ventricular function relative to afterload. Abbreviations as in Fig. 1.
3.0
2.5
~2.0
>~-
1.5 9
"
'
O"
"1
'
9
"
n
i i
~
"
. . . .
"
.
.
.
.
.
.
.
.
'
'
9
4,
,
. . . .
,
.
.
.
,
,
.
,
1.0 .
0.5' 10
.
.
.
.
,
,
,
,
I
f
I
f
I
I
I
20
30
40
50
60
80
100
ESS (gmlcm)
Fig. 3. Following discontinuation of ECMO, left ventricular systolic performance was normal in all survivors (group 3). No patients exhibited reduced ventricular function. Afterload was normal in 10 infants but markedly elevated in two patients with systemic hypertension 9 Abbreviations as in Fig. I.
was measured had values
