- Email: [email protected]
Decreased incidence of intracranial hemorrhage using cephalic jugular venous drainage during neonatal extracorporeal membrane oxygenation
Decreased Incidence of Intracranial Hemorrhage Using Cephalic Jugular Venous Drainage During Neonatal Extracorporeai Membrane Oxygenation By Timothy A. O'Connor, Barbara M. Haney, Gary E. Grist, John C. Egelhoff, Charles L. Snyder, and Keith W. Ashcraft Kansas City, M i s s o u r i 9 Intracranial hemorrhage (ICH) remains one of the more common serious complications of extracorporeal membrane oxygenation (ECMO) in neonates. In 1990 this center began routine use of cephalic jugular venous drainage during neonatal ECMO to augment blood return to the ECMO pump and potentially decrease the incidence of ICH by decreasing cerebral venous pressure, Thirty-four ECMO cases utilizing cephalic jugular venous drainage were compared with the previous 34 ECMO cases. The incidence of ICH decreased from 35% (12/34) to 6% (2/34) when neonates without cephalic jugular venous drainage are compared with those being subject to this technique (P < .01). No differences were found between the two groups in gestational age, birth weight, duration of ECMO, survival, platelet counts, activated clotting times, or incidence of other bleeding complications. Cephalic jugular venous drainage during neonatal ECMO appears to be safe and may decrease the incidence of ICH.
Copyright 9 1993 by W.B. Saunders Company INDEX WORDS: Extracorporeal membrane oxygenation (ECMO), intracranial hemorrhage.
INCE THE introduction of extracorporeal membrane oxygenation (ECMO) for therapy of severe respiratory and cardiac insufficiency in neonates, intracranial hemorrhage (ICH) has remained a common complication with high morbidity and mortality. For infants < 35 weeks' gestation the ICH rate during ECMO has been found to be as high as 100%. 1 One series reported in 1986 showed an ICH rate of 52% in neonates >35 weeks' gestation who underwent ECMO therapy. 2 More recently the incidence of ICH or infarction has been reported to be in the range of 16%, according to the national ECMO registry. 3 However, not all reporting programs perform computed tomography (CT) on all ECMO patients, potentially underestimating the true incidence of ICH.
S
From the Sections of Neonatology, Surgery, and Radiology, The Children's Mercy Hospital and University of Missouri, Kansas City, MO. Presented at the 1992 Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, San Francisco, California, October 9-11, 1992. Address reprint requests to Timothy A. O'Connor, MD, Children's Hospital at the University of Missouri, Columbia, MO 65212. Copyright 9 1993 by W.B. Saunders Company 0022-3468/93/2810-0022503.00/0 1332
The etiology of the high ICH rate is probably multifactorial, involving reperfusion of previously damaged areas, abrupt changes in local and systemic blood pressure, systemic anticoagulation, abnormalities of clotting function induced by illness and the ECMO circuit, and the effect of increased cerebral venous pressure after ligation of the internal jugular vein. Since 1990 this center has utilized drainage of the cephalic portion of the right internal jugular vein during neonatal ECMO using a technique described previously. 4 We postulated that use of this technique would decrease the incidence of ICH seen in our ECMO population. MATERIALS AND METHODS Data were reviewed on 34 consecutive neonatal ECMO cases performed since the institution of cephalic jugular venous drainage. These cases were compared with the previous 34 neonatal ECMO cases in which cephalic jugular drainage was not utilized. These 68 cases represent approximately the middle third and last third of ECMO experience at our center. ICH was diagnosed in these infants by cranial ultrasound (US), CT, or at autopsy. At our center, cranial US is performed prior to ECMO cannulation in all infants and an ICH of grade II or greater is considered a contraindication to ECMO therapy. All surviving infants undergo cranial CT prior to discharge. Sonographic and CT films before and after the ECMO courses were reviewed by a pediatric radiologist (J.C.E.) who was blinded regarding the presence of a cephalic jugular drain during the ECMO course. Platelet counts are maintained above 100,000/mL and activated clotting times (ACTs) are maintained in the range 210 to 225 seconds, in our center during neonatal ECMO. Cannulation of the cephalic portion of the right jugular vein is accomplished at the time of initial ECMO cannulation. The internal jugular vein is surrounded by three silk ligatures, the central one which is tied. The vein is then cannulated through two separate venotomies, one proximal and one distal to the central ligature. The right atrial venous cannula is placed first. Typically a 9.6F vasocath (GESCO International, San Antonio, TX) modified with 2 side ports 1 cm from the distal end is utilized for the cephalic drain. This was advanced to its limit in the cephalad portion of the jugular vein and then withdrawn 1 cm. The catheters are then secured and the cephalic jugular cannula connected to the venous return tubing. Catheter position in the jugular bulb is confirmed radiographically. Figures 1 and 2 demonstrate conventional venous ECMO cannulation and the cephalic jugular venous cannulation. Patient demographics and variables regarding clinical course and outcome were compared using the • Fischer Exact, or Student's t test where appropriate. A level of significance of < .05 was used for all tests.
Journal of Pediatric Surgery, Vol 28, No 10 (October), 1993: pp 1332-1335
CEPHALIC JUGULAR DRAINAGE DURING ECMO
1333
Table 1. Demographic and Pre-ECMO Data No venous drainage via
R.InternalJugular
~ ( t e ~ o u ~
fSkufl
Venous
Jugular Ligagon
Gestational age (wk) Birth weight (kg) No. < 36 wk gestation Prior to ECMO Last measured pH No, with pH <_7.25 Last measured PaO2 Oxygenation index
Without Cephalic Cannula
With Cephalic Cannula
P Value
39.1 (2.6) 3.35 (.52) 10/34 (29%)
38.4 (2.8) 3.42 (.58) 9/34 (26%)
NS NS NS
7.43 5/34 33.8 63.4
NS NS NS NS
7.39 8/34 29.3 78.3
(.18) (24%) (12.2) (45.6)
(.15) (15%) (14.1) (37.3)
Cannula
RightAtrium
~C
NOTE. Results are expressed as mean (SD) or no. (%) as appropriate. A level of significance of <.05 was used, NS = not significant. Oxygenation index = [(MAP)(FiO2)/PaO2] • 100.
Partial obstruction
r
Fig 1. Traditional venous cannulation for neonatal ECMO with ligation of the right internal jugular vein,
RESULTS
Demographic and clinical data regarding the study population are shown in Table 1. No differences between the cephalic jugular drain group and the group without cephalic jugular drainage were found in mean gestational age, birth weight, or percentage < 36 weeks' gestation. Likewise, no differences were found between the groups in mean pH, percentage with acidosis (pH < 7.25), PaO2, or oxygenation index prior to cannulation. Table 2 illustrates variables of the ECMO course and outcome data. No differences were found between the groups in ECMO duration, mean or lowest platelet counts during the ECMO course, mean low or high ACTs, or in the percentage of patients with bleeding complications other than ICH (defined as
~
Cephalic
~
BaseofSkull
bleeding requiring surgical intervention or removal from ECMO). The incidence of ICH in the group with cephalic jugular venous drainage was 6% (2/34) compared with 35% (12/34) in those infants without cephalic jugular drainage (P < .01). Two infants in the cephalic jugular drain group had CT evidence of ischemia without ICH as did one infant in the group without cephalic jugular drainage. If the patients with evidence of ischemia are added to those with ICH then the incidence of ischemia or ICH in the cephalic jugular catheter group was 12% (4/34) versus 38% (13/34) in the patients without cephalic jugular drainage (P < .05). Survival was 85% in the group with cephalic jugular drainage compared with 74% in the group without cephalic jugular drainage. This difference was not statistically significant; however, 5 of the 9 nonsurvivors in the latter group had a significant ICH. Table 3 illustrates the sites of ICH in the neonates in this series. As in other series of ICH during ECMO, the sites of ICH were varied and no distinct pattern of injury was evident. 5,6 Five of the 34 cases with cephalic jugular drainage were venovenous ECMO and all others venoarterial. One of the five venovenous cases had an ICH. All cases without cephalic jugular drainage were venoarterial. Table 2, ECMO Course and Outcome Data Without Cephalic With Cephalic Cannula Cannula P Value
. ~
Cannuta
•
~lntertmlJ
ular
RightAtrium~ Fig 2. Venous cannulation for neonates using drainage from the cephalad portion of the right internal jugular vein.
ICH Other bleeding complications Platelet count (• Lowest plateiet count (• Highest daily ACT (s) Lowest daily ACT (s) ECMO duration (h) Survival
12/34 4/34 122.3 68.1 244 168 122.8 25/34
(35%) (12%) (15.5) (24.6) (15) (16) (69.9) (74%)
2/34 (6%) 6/34 (18%) 121.6 (13.9) 63.4(15.1) 240 (14) 173 (14) 118.8 (84.1) 29/34 (85%)
.006 NS NS NS NS NS NS NS
NOTE. Results are expressed as mean (SD) or no. (%) as appropriate. A level of significance of .05 was used. Abbreviations: NS, not significant; ICH, intracranial hemorrhage.
1334
O'CONNOR ET AL
Table 3. Sites of ICH in the Study Population Without Cephalic Cannula L parietal-occipital IVH, R thalamic L germinal matrix R cerebellar 13germinal matrix IVH, L thalamic B temporal-occipital, IVH IVH, B germinal matrix Subarachnoid IVH, B germinal matrix R frontal, parietal, temporal Subarachnoid
With Cephalic Cannula Subarachnoid IVH, R thalamic
NOTE. The locations of the ICHs seen in 12 of 34 ECMO patients without cephalic jugular cannulae and 2 of 34 patients with cephalic jugular cannulae. Abbreviations: IVH, intraventricular hemorrhage; L, left; R, right; B, bilateral.
Complications of cephalic jugular venous drainage in these 34 infants included clotting of the cannula (n = 1) and dislodgement of the cephalic cannula (n = 2). None of these infants had an ICH or ischemia by CT. A Doppler flow probe (Transonic Systems Inc, Ithaca, NY) was used to measure blood flow in the cephalic jugular cannulae in most cases. The mean flow rate was 75 mL/min and this was independent of the ECMO pump flow rates. DISCUSSION
ICH remains a common and serious complication of ECMO therapy, particularly in the preterm infant. This is likely due to a number of factors including increased cerebral venous pressure. In 1989 a technique of cannulation of the cephalic portion of the right internal jugular vein during ECMO was described. This resulted in a decrease in the incidence of ICH from 30% to 8.3% in 34- to 36-week-gestation infants undergoing ECMO therapy. 4 In an experimental primate model the effects of cephalic jugular venous drainage during ECMO were evaluated. Pressures in the sagittal sinus and right and left ventricles were measured while the cephalic jugular cannula was clamped and unclamped. In both
animals studied the sagittal sinus pressure decreased when the cephalic catheter was open. While this difference was not statistically significant, only two animals were studied. 7 Duplex and color Doppler evaluations of the superior sagittal sinus have been performed during venoarterial ECMO. With a cephalic jugular catheter in place, superior sagittal sinus blood flow was measured with and without occlusion of the catheter. Mean flow velocities in the superior sagittal sinus decreased from 7.5 cm/s to 5.3 cm/s with occlusion of the cephalic jugular catheter (P = .01). It was speculated that the cephalic jugular catheter maintained a more normal cerebral venous drainage during ECMO. 8 In our series of patients, we found an ICH rate of 6% in patients with cephalic jugular drainage, significantly reduced from historical controls. Although the incidence of ICH in our historical control group was somewhat higher than one might expect, one possible explanation for the 35% incidence of ICH lies in the fact that these were historical controls, ie, patients who underwent ECMO during the late 1980s, a time when many centers throughout the country were reporting similar rates of ICH. However, the incidence of ICH in our patients with cephalic venous drainage catheters is still lower than that reported by the ECMO registry.3 Few complications from cephalic jugular venous drainage occurred. In addition, blood return to the ECMO pump was substantially increased. This is especially advantagous in venovenous ECMO, where the effects of recirculation are blunted by the extra source of deoxygenated blood. This augmented blood return has virtually elimated the problem of insufficient blood return (cutting out) to the ECMO pump. Since the institution of this technique at this center, we have found a decrease in the severity of scalp edema, presumably due to reduced venous congestion. Although scalp edema is difficult to quantitate objectively, this finding was also seen in a series of patients reported previously.4 Cannulation of the cephalic portion of the right internal jugular vein during ECMO can be accomplished with safety and appears to decrease the incidence of ICH. A prospective randomized trial of this technique is warranted.
REFERENCES 1. Cilley RE, Zwischenberger JB, Andrews AF, et al: Intracranial hemorrhage during extracorporeal membrane oxygenation in neonates. Pediatrics 78:699-704, 1986 2. Sell LL, Cullen ML, Whittlesey GC, et ah Hemorrhagic complications during extracorporeal membrane oxygenation: Prevention and treatment. J Pediatr Surg 21:1087-1090, 1986 3. Extracorporeal Life Support Organization: ECMO Registry Report, January 1992 4. Gangitano ES, Vogt JF, Muenchow SK, et al: Proposal for a
multicenter trial to study the effects of cannulation of the cephalad segment of the internal jugular vein during extracorporeal membrane oxygenation in the neonate. Presented at the meeting of the Extracorporeal Life Support Organization, Ann Arbor, MI, October 1989 5. Babcock DS, Han BK, Weiss RG, et ah Brain abnormalities in infants on extracorporeal membrane oxygenation: Sonographic and CT findings. A JR Am J Roentgenol 153:571-576, 1989 6. Taylor GA, Fitz CR, Glass P, Short BL: CT of cerebrovascu-
CEPHALIC JUGULAR DRAINAGE DURING ECMO
1335
lar injury after neonatal extracorporeal membrane oxygenation: Implications for neurodevelopmental outcome. AJR Am J RoentgenoI 153:121-126, 1989 7. Atkinson JB, Kitagawa H, McComb JG: Proximal and distal cannulation of the internal jugular vein for ECMO in a primate. Presented at the meeting of the Extracorporeal Life Support
Organization, Ann Arbor, MI, October 1989 8. Taylor GA, Walker LK: Doppler evaluation of the intracranial venous system following ligation of the right jugular vein in infants treated with ECMO. Presented at the 8th Children's National Medical Center ECMO Symposium, Breckenridge, CO, February 1992
Discussion L.R. Scherer (Indianapolis, IN): I think we're all aware of the importance of ECMO as a salvage therapy for newborns with severe respiratory failure. This paper also points out the real, but variable, complication of intracranial hemorrhage in these infants on ECMO. A number of investigators have looked, as already noted in this paper, at a number of causes of the cerebral vascular injury, including arterial inflow due to carotid artery ligation, the loss of autoregulation, and venous hypertension. The distribution of hemorrhagic lesions in the periventricular area of the medullary veins has led a number of authors to focus on the effects of intracranial venous hypertension. Using Doppler ultrasounds, Cowan, studying healthy newborns, has shown considerable anatomic variability in the drainage of the cranial venous system, and she has noted that shunting of the venous drainage during a compression of either jugular vein is dependent upon the contralateral jugular vein or the vertebral veins or can even be totally dependent on the compressed vein itself. Taylor has also shown, with the use of intracranial color Doppler ultrasound, the significant reduction of superior saggital sinus blood velocity with occlusion produced by the jugular bulb catheter in 8 of 16 infants. Further, he noted that persistent reduction of the flow velocity was associated with a higher risk of cerebrovascular injury. This present study, although retrospective, provides one of the largest clinical studies providing some substantial evidence that jugular venous drainage decreases the incidence of intracranial hemorrhage in these critically ill neonates. I have several questions. During the course of ECMO therapy was there an attempt to evaluate the superior saggital sinus blood flow velocity? If this was performed, did it correlate with the incidence of intracranial hemorrhage? Where in the time course of therapy were intracranial hemorrhages diagnosed? Was this at the beginning of therapy in grade 1 hemorrhages or were these later in the course of therapy? Could you explain your higher incidence of intracranial hemorrhage in your precephalic catheter drainage of 35%, when the national average is 16%?
As another marker of your drainage system, Walker at Johns Hopkins has reported the use of mixed venous 02 saturation from the jugular bulb catheter as a measurement of cerebral perfusion and oxygenation as well as to determine the need for increased pump flow. Have you had an opportunity to evaluate mixed venous oxygen saturation in your patient population? R. Arensman (Chicago, IL): I would like to ask if you had statistical significance between the group with the cephalic jugular drainage and the national published complication rate of intraventricular hemorrhage? The figure you have given is 2 to 3 times the national norm. I'm not sure that those of us who do not use cephalic drainage and have complication rates between 8% and 11%, would feel that there is statistically significant difference from your 6% figure. R. Connors (Grand Rapids, MI): It seems the use of your study design using historical controls significantly compromises your ability to draw conclusions. Why did you choose that design rather than prospectively randomize your patients? T. O'Connor (response): Cerebral blood flow velocities were not measured in those patients. As to why the control group had such a high ICH rate, this may be in part a learning curve phenomenon, or patient selection criteria which were overlooked in our review. Nonetheless, there was a dramatic reduction in ICH after institution of the cephalic drain. While our current rate of ICH appears to be less than the national average per the registry, statistical comparison to that heterogenous group I don't feel is valid, especially in light of variation in surveillance techniques used in different centers. No pattern existed as to timing of the ICHs. Normative data regarding jugular bulb venous 02 saturation will be published soon in an upcoming issue of the Journal of Extracorporeal Technology. In response to the question concerning choice of retrospective study design, this review was undertaken to see if performing a prospective analysis was indicated, and we believe it is.
Copyright 9 1993 by W.B. Saunders Company INDEX WORDS: Extracorporeal membrane oxygenation (ECMO), intracranial hemorrhage.
INCE THE introduction of extracorporeal membrane oxygenation (ECMO) for therapy of severe respiratory and cardiac insufficiency in neonates, intracranial hemorrhage (ICH) has remained a common complication with high morbidity and mortality. For infants < 35 weeks' gestation the ICH rate during ECMO has been found to be as high as 100%. 1 One series reported in 1986 showed an ICH rate of 52% in neonates >35 weeks' gestation who underwent ECMO therapy. 2 More recently the incidence of ICH or infarction has been reported to be in the range of 16%, according to the national ECMO registry. 3 However, not all reporting programs perform computed tomography (CT) on all ECMO patients, potentially underestimating the true incidence of ICH.
S
From the Sections of Neonatology, Surgery, and Radiology, The Children's Mercy Hospital and University of Missouri, Kansas City, MO. Presented at the 1992 Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, San Francisco, California, October 9-11, 1992. Address reprint requests to Timothy A. O'Connor, MD, Children's Hospital at the University of Missouri, Columbia, MO 65212. Copyright 9 1993 by W.B. Saunders Company 0022-3468/93/2810-0022503.00/0 1332
The etiology of the high ICH rate is probably multifactorial, involving reperfusion of previously damaged areas, abrupt changes in local and systemic blood pressure, systemic anticoagulation, abnormalities of clotting function induced by illness and the ECMO circuit, and the effect of increased cerebral venous pressure after ligation of the internal jugular vein. Since 1990 this center has utilized drainage of the cephalic portion of the right internal jugular vein during neonatal ECMO using a technique described previously. 4 We postulated that use of this technique would decrease the incidence of ICH seen in our ECMO population. MATERIALS AND METHODS Data were reviewed on 34 consecutive neonatal ECMO cases performed since the institution of cephalic jugular venous drainage. These cases were compared with the previous 34 neonatal ECMO cases in which cephalic jugular drainage was not utilized. These 68 cases represent approximately the middle third and last third of ECMO experience at our center. ICH was diagnosed in these infants by cranial ultrasound (US), CT, or at autopsy. At our center, cranial US is performed prior to ECMO cannulation in all infants and an ICH of grade II or greater is considered a contraindication to ECMO therapy. All surviving infants undergo cranial CT prior to discharge. Sonographic and CT films before and after the ECMO courses were reviewed by a pediatric radiologist (J.C.E.) who was blinded regarding the presence of a cephalic jugular drain during the ECMO course. Platelet counts are maintained above 100,000/mL and activated clotting times (ACTs) are maintained in the range 210 to 225 seconds, in our center during neonatal ECMO. Cannulation of the cephalic portion of the right jugular vein is accomplished at the time of initial ECMO cannulation. The internal jugular vein is surrounded by three silk ligatures, the central one which is tied. The vein is then cannulated through two separate venotomies, one proximal and one distal to the central ligature. The right atrial venous cannula is placed first. Typically a 9.6F vasocath (GESCO International, San Antonio, TX) modified with 2 side ports 1 cm from the distal end is utilized for the cephalic drain. This was advanced to its limit in the cephalad portion of the jugular vein and then withdrawn 1 cm. The catheters are then secured and the cephalic jugular cannula connected to the venous return tubing. Catheter position in the jugular bulb is confirmed radiographically. Figures 1 and 2 demonstrate conventional venous ECMO cannulation and the cephalic jugular venous cannulation. Patient demographics and variables regarding clinical course and outcome were compared using the • Fischer Exact, or Student's t test where appropriate. A level of significance of < .05 was used for all tests.
Journal of Pediatric Surgery, Vol 28, No 10 (October), 1993: pp 1332-1335
CEPHALIC JUGULAR DRAINAGE DURING ECMO
1333
Table 1. Demographic and Pre-ECMO Data No venous drainage via
R.InternalJugular
~ ( t e ~ o u ~
fSkufl
Venous
Jugular Ligagon
Gestational age (wk) Birth weight (kg) No. < 36 wk gestation Prior to ECMO Last measured pH No, with pH <_7.25 Last measured PaO2 Oxygenation index
Without Cephalic Cannula
With Cephalic Cannula
P Value
39.1 (2.6) 3.35 (.52) 10/34 (29%)
38.4 (2.8) 3.42 (.58) 9/34 (26%)
NS NS NS
7.43 5/34 33.8 63.4
NS NS NS NS
7.39 8/34 29.3 78.3
(.18) (24%) (12.2) (45.6)
(.15) (15%) (14.1) (37.3)
Cannula
RightAtrium
~C
NOTE. Results are expressed as mean (SD) or no. (%) as appropriate. A level of significance of <.05 was used, NS = not significant. Oxygenation index = [(MAP)(FiO2)/PaO2] • 100.
Partial obstruction
r
Fig 1. Traditional venous cannulation for neonatal ECMO with ligation of the right internal jugular vein,
RESULTS
Demographic and clinical data regarding the study population are shown in Table 1. No differences between the cephalic jugular drain group and the group without cephalic jugular drainage were found in mean gestational age, birth weight, or percentage < 36 weeks' gestation. Likewise, no differences were found between the groups in mean pH, percentage with acidosis (pH < 7.25), PaO2, or oxygenation index prior to cannulation. Table 2 illustrates variables of the ECMO course and outcome data. No differences were found between the groups in ECMO duration, mean or lowest platelet counts during the ECMO course, mean low or high ACTs, or in the percentage of patients with bleeding complications other than ICH (defined as
~
Cephalic
~
BaseofSkull
bleeding requiring surgical intervention or removal from ECMO). The incidence of ICH in the group with cephalic jugular venous drainage was 6% (2/34) compared with 35% (12/34) in those infants without cephalic jugular drainage (P < .01). Two infants in the cephalic jugular drain group had CT evidence of ischemia without ICH as did one infant in the group without cephalic jugular drainage. If the patients with evidence of ischemia are added to those with ICH then the incidence of ischemia or ICH in the cephalic jugular catheter group was 12% (4/34) versus 38% (13/34) in the patients without cephalic jugular drainage (P < .05). Survival was 85% in the group with cephalic jugular drainage compared with 74% in the group without cephalic jugular drainage. This difference was not statistically significant; however, 5 of the 9 nonsurvivors in the latter group had a significant ICH. Table 3 illustrates the sites of ICH in the neonates in this series. As in other series of ICH during ECMO, the sites of ICH were varied and no distinct pattern of injury was evident. 5,6 Five of the 34 cases with cephalic jugular drainage were venovenous ECMO and all others venoarterial. One of the five venovenous cases had an ICH. All cases without cephalic jugular drainage were venoarterial. Table 2, ECMO Course and Outcome Data Without Cephalic With Cephalic Cannula Cannula P Value
. ~
Cannuta
•
~lntertmlJ
ular
RightAtrium~ Fig 2. Venous cannulation for neonates using drainage from the cephalad portion of the right internal jugular vein.
ICH Other bleeding complications Platelet count (• Lowest plateiet count (• Highest daily ACT (s) Lowest daily ACT (s) ECMO duration (h) Survival
12/34 4/34 122.3 68.1 244 168 122.8 25/34
(35%) (12%) (15.5) (24.6) (15) (16) (69.9) (74%)
2/34 (6%) 6/34 (18%) 121.6 (13.9) 63.4(15.1) 240 (14) 173 (14) 118.8 (84.1) 29/34 (85%)
.006 NS NS NS NS NS NS NS
NOTE. Results are expressed as mean (SD) or no. (%) as appropriate. A level of significance of .05 was used. Abbreviations: NS, not significant; ICH, intracranial hemorrhage.
1334
O'CONNOR ET AL
Table 3. Sites of ICH in the Study Population Without Cephalic Cannula L parietal-occipital IVH, R thalamic L germinal matrix R cerebellar 13germinal matrix IVH, L thalamic B temporal-occipital, IVH IVH, B germinal matrix Subarachnoid IVH, B germinal matrix R frontal, parietal, temporal Subarachnoid
With Cephalic Cannula Subarachnoid IVH, R thalamic
NOTE. The locations of the ICHs seen in 12 of 34 ECMO patients without cephalic jugular cannulae and 2 of 34 patients with cephalic jugular cannulae. Abbreviations: IVH, intraventricular hemorrhage; L, left; R, right; B, bilateral.
Complications of cephalic jugular venous drainage in these 34 infants included clotting of the cannula (n = 1) and dislodgement of the cephalic cannula (n = 2). None of these infants had an ICH or ischemia by CT. A Doppler flow probe (Transonic Systems Inc, Ithaca, NY) was used to measure blood flow in the cephalic jugular cannulae in most cases. The mean flow rate was 75 mL/min and this was independent of the ECMO pump flow rates. DISCUSSION
ICH remains a common and serious complication of ECMO therapy, particularly in the preterm infant. This is likely due to a number of factors including increased cerebral venous pressure. In 1989 a technique of cannulation of the cephalic portion of the right internal jugular vein during ECMO was described. This resulted in a decrease in the incidence of ICH from 30% to 8.3% in 34- to 36-week-gestation infants undergoing ECMO therapy. 4 In an experimental primate model the effects of cephalic jugular venous drainage during ECMO were evaluated. Pressures in the sagittal sinus and right and left ventricles were measured while the cephalic jugular cannula was clamped and unclamped. In both
animals studied the sagittal sinus pressure decreased when the cephalic catheter was open. While this difference was not statistically significant, only two animals were studied. 7 Duplex and color Doppler evaluations of the superior sagittal sinus have been performed during venoarterial ECMO. With a cephalic jugular catheter in place, superior sagittal sinus blood flow was measured with and without occlusion of the catheter. Mean flow velocities in the superior sagittal sinus decreased from 7.5 cm/s to 5.3 cm/s with occlusion of the cephalic jugular catheter (P = .01). It was speculated that the cephalic jugular catheter maintained a more normal cerebral venous drainage during ECMO. 8 In our series of patients, we found an ICH rate of 6% in patients with cephalic jugular drainage, significantly reduced from historical controls. Although the incidence of ICH in our historical control group was somewhat higher than one might expect, one possible explanation for the 35% incidence of ICH lies in the fact that these were historical controls, ie, patients who underwent ECMO during the late 1980s, a time when many centers throughout the country were reporting similar rates of ICH. However, the incidence of ICH in our patients with cephalic venous drainage catheters is still lower than that reported by the ECMO registry.3 Few complications from cephalic jugular venous drainage occurred. In addition, blood return to the ECMO pump was substantially increased. This is especially advantagous in venovenous ECMO, where the effects of recirculation are blunted by the extra source of deoxygenated blood. This augmented blood return has virtually elimated the problem of insufficient blood return (cutting out) to the ECMO pump. Since the institution of this technique at this center, we have found a decrease in the severity of scalp edema, presumably due to reduced venous congestion. Although scalp edema is difficult to quantitate objectively, this finding was also seen in a series of patients reported previously.4 Cannulation of the cephalic portion of the right internal jugular vein during ECMO can be accomplished with safety and appears to decrease the incidence of ICH. A prospective randomized trial of this technique is warranted.
REFERENCES 1. Cilley RE, Zwischenberger JB, Andrews AF, et al: Intracranial hemorrhage during extracorporeal membrane oxygenation in neonates. Pediatrics 78:699-704, 1986 2. Sell LL, Cullen ML, Whittlesey GC, et ah Hemorrhagic complications during extracorporeal membrane oxygenation: Prevention and treatment. J Pediatr Surg 21:1087-1090, 1986 3. Extracorporeal Life Support Organization: ECMO Registry Report, January 1992 4. Gangitano ES, Vogt JF, Muenchow SK, et al: Proposal for a
multicenter trial to study the effects of cannulation of the cephalad segment of the internal jugular vein during extracorporeal membrane oxygenation in the neonate. Presented at the meeting of the Extracorporeal Life Support Organization, Ann Arbor, MI, October 1989 5. Babcock DS, Han BK, Weiss RG, et ah Brain abnormalities in infants on extracorporeal membrane oxygenation: Sonographic and CT findings. A JR Am J Roentgenol 153:571-576, 1989 6. Taylor GA, Fitz CR, Glass P, Short BL: CT of cerebrovascu-
CEPHALIC JUGULAR DRAINAGE DURING ECMO
1335
lar injury after neonatal extracorporeal membrane oxygenation: Implications for neurodevelopmental outcome. AJR Am J RoentgenoI 153:121-126, 1989 7. Atkinson JB, Kitagawa H, McComb JG: Proximal and distal cannulation of the internal jugular vein for ECMO in a primate. Presented at the meeting of the Extracorporeal Life Support
Organization, Ann Arbor, MI, October 1989 8. Taylor GA, Walker LK: Doppler evaluation of the intracranial venous system following ligation of the right jugular vein in infants treated with ECMO. Presented at the 8th Children's National Medical Center ECMO Symposium, Breckenridge, CO, February 1992
Discussion L.R. Scherer (Indianapolis, IN): I think we're all aware of the importance of ECMO as a salvage therapy for newborns with severe respiratory failure. This paper also points out the real, but variable, complication of intracranial hemorrhage in these infants on ECMO. A number of investigators have looked, as already noted in this paper, at a number of causes of the cerebral vascular injury, including arterial inflow due to carotid artery ligation, the loss of autoregulation, and venous hypertension. The distribution of hemorrhagic lesions in the periventricular area of the medullary veins has led a number of authors to focus on the effects of intracranial venous hypertension. Using Doppler ultrasounds, Cowan, studying healthy newborns, has shown considerable anatomic variability in the drainage of the cranial venous system, and she has noted that shunting of the venous drainage during a compression of either jugular vein is dependent upon the contralateral jugular vein or the vertebral veins or can even be totally dependent on the compressed vein itself. Taylor has also shown, with the use of intracranial color Doppler ultrasound, the significant reduction of superior saggital sinus blood velocity with occlusion produced by the jugular bulb catheter in 8 of 16 infants. Further, he noted that persistent reduction of the flow velocity was associated with a higher risk of cerebrovascular injury. This present study, although retrospective, provides one of the largest clinical studies providing some substantial evidence that jugular venous drainage decreases the incidence of intracranial hemorrhage in these critically ill neonates. I have several questions. During the course of ECMO therapy was there an attempt to evaluate the superior saggital sinus blood flow velocity? If this was performed, did it correlate with the incidence of intracranial hemorrhage? Where in the time course of therapy were intracranial hemorrhages diagnosed? Was this at the beginning of therapy in grade 1 hemorrhages or were these later in the course of therapy? Could you explain your higher incidence of intracranial hemorrhage in your precephalic catheter drainage of 35%, when the national average is 16%?
As another marker of your drainage system, Walker at Johns Hopkins has reported the use of mixed venous 02 saturation from the jugular bulb catheter as a measurement of cerebral perfusion and oxygenation as well as to determine the need for increased pump flow. Have you had an opportunity to evaluate mixed venous oxygen saturation in your patient population? R. Arensman (Chicago, IL): I would like to ask if you had statistical significance between the group with the cephalic jugular drainage and the national published complication rate of intraventricular hemorrhage? The figure you have given is 2 to 3 times the national norm. I'm not sure that those of us who do not use cephalic drainage and have complication rates between 8% and 11%, would feel that there is statistically significant difference from your 6% figure. R. Connors (Grand Rapids, MI): It seems the use of your study design using historical controls significantly compromises your ability to draw conclusions. Why did you choose that design rather than prospectively randomize your patients? T. O'Connor (response): Cerebral blood flow velocities were not measured in those patients. As to why the control group had such a high ICH rate, this may be in part a learning curve phenomenon, or patient selection criteria which were overlooked in our review. Nonetheless, there was a dramatic reduction in ICH after institution of the cephalic drain. While our current rate of ICH appears to be less than the national average per the registry, statistical comparison to that heterogenous group I don't feel is valid, especially in light of variation in surveillance techniques used in different centers. No pattern existed as to timing of the ICHs. Normative data regarding jugular bulb venous 02 saturation will be published soon in an upcoming issue of the Journal of Extracorporeal Technology. In response to the question concerning choice of retrospective study design, this review was undertaken to see if performing a prospective analysis was indicated, and we believe it is.