Estimation of Cerebral Blood Flow at Bedside by Continuous Jugular Thermodilution

Abstract

The Kety-Schmidt technique can be regarded as the reference method for the measurement of cerebral blood flow (CBF). However, the method is somewhat cumbersome for routine use in the intensive care unit (ICU) at the bedside. The continuous thermodilution technique developed many years ago for the measurement of coronary sinus blood flow can be applied for the measurement of jugular blood flow (JBF). However, the measurement of JBF by thermodilution has never been validated using the Kety-Schmidt reference method. We first validate the continuous thermodilution in vitro by comparison with a volumetric flow. The thermodilution method is accurate for flows between 50 and 900 ml min⁻¹ with a mean difference volumetric-thermodilution flow of −1 ± 18 ml min⁻¹ (mean ± SD), and precise with a coefficient of variability ranging between 1.21% and 2.50%. In vivo accuracy was assessed by comparing in 15 comatose patients CBF measured using the Kety-Schmidt (CBF_KS) method and estimated from JBF measured by thermodilution (CBF_TH) at four levels of arterial P_aCO₂ (25, 30, 35, and 40 mm Hg). The mean difference CBF_KS — CBF_TH is −0.9 ± 3.6 ml min⁻¹ 100 g⁻¹. In vivo precision of the method was good, with a coefficient of variability of 4.1% in mean. We conclude that jugular continuous thermodilution technique is a reliable method for estimating CBF at the bedside. This technique allows repeated measurements jugular bulb blood sampling for brain metabolic studies.

Keywords

Nitrous oxide Cerebral blood flow Jugular blood flow Continuous thermodilution Cerebral vasoreactivity Carbon dioxide

Although many critically ill patients suffer from neurological problems, their management does not routinely require monitoring of cerebral blood flow (CBF). Nevertheless, assessment of cerebral hemodynamics can be of great interest for diagnostic and therapeutic purposes in patients in coma resulting from head injury, subarachnoid hemorrhage, anoxia, or metabolic encephalopathy. In these clinical situations CBF is a key variable for better understanding the pathological process responsible for cerebral ischemia.

In 1945, Kety and Schmidt (1945) described a quantitative method for measurement of CBF, based on the classical Fick's principle following the administration of nitrous oxide (N₂O) as a tracer. Using this technique interesting and fundamental information has been obtained on the physiology and the metabolism of the cerebral circulation in normal man. Subsequent modifications of the technology have involved the use of radioisotopes such as krypton-85 or xenon-133 combined with compact scintillation detectors and microprocessors (Obrist et al., 1975; Edvinsson et al., 1993; Young and Ornstein, 1994). Even though these methods for CBF measurements have allowed a lot of investigations, particularly in patients with severe head injury (Obrist et al, 1984) and with cerebral ischemia after cardiac arrest (Cohan et al, 1989), a more widespread clinical application has been limited by several disadvantages, including elevated cost, exposure to radioisotopes, requirement for complicated analyses and calculations. Furthermore, these techniques cannot be performed easily at bedside in the intensive care unit (ICU) despite some recent attempts to simplify the original Kety-Schmidt method (Kozam et al, 1970).

More sophisticated techniques, such as single-photon emission computed tomography (SPECT), positron emission tomography (PET), angiography with magnetic resonance imaging (MRI), and stable xenon-enhanced computed tomography (sXe-CT), have been developed recently. These represent powerful research tools, but do ot yield bedside measurements (Edvinsson et al., 1993; Young and Ornstein, 1994).

The exploration of the cerebral circulation in the ICU requires fast, easy, and accurate bedside procedures with relatively low cost. The method must be reproducible, allowing repeated or continuous measurements. Recently, several techniques meeting these requirements have been developed: transcranial Doppler (TCD) (Aaslid et al., 1982), jugular bulb oximetry (SjO₂) using optical fibers (Prough, 1991), and near-infrared cerebral spectroscopy (MRS) (Prough, 1991; Simonson and Piantadosi, 1995). These techniques, however, do not measure true flows but indices that can be related to flow changes. Therefore, we were interested in continuous thermodilution, described many years ago for the measurement of coronary sinus blood flow (Ganz et al., 1971). This method meets the above-mentioned requirements and has been occasionally used to measure jugular blood flow (JBF) (Van der Linden et al., 1991; Wilson and Halsey, 1970). However, this technique has neither been automated using a computer nor validated by comparison with the Kety-Schmidt reference method.

MATERIALS AND METHODS

Patients

Fifteen consecutive comatose patients, seven women and eight men, were studied according to a protocol that was approved by the Committee of Medical Ethics of our institution. The etiology of the coma, the Glasgow coma scale (Teasdale and Jennett, 1974), the Glasgow outcome scale (Jennett and Bond, 1975), and the anthropometric data are summarized in Table 1. All the patients were mechanically ventilated through an endotracheal tube, using a time-cycled 900 C servoventilator (Elema, Sölna, Sweden). The patients were sedated with midazolam or morphine sulfate and were given pancuronium bromide to optimize ventilatory control during the protocol. Drugs and fluids were determined by the attending physician and were not modified during the study.

Table 1.

Anthropometric data of 15 patients in coma

						GCS
Patient	Age (yr)	Sex	Etiology of coma	Height (cm)	Weight (kg)	Day of study	Day after study	GOS, days after study	Cause of death	Time between initial event and study (days)
1	34	M	Encephalitis	168	72	3	3	1(74)	Status epilepticus	22
2	71	F	Head trauma	164	75	10	11	3(12)	—	6
3	74	F	Subarachnoid hemorrhage	156	50	7	4	1(12)	Recurrence of hemorrhage	1
4	79	F	Subarachnoid hemorrhage	152	60	4	7	1(15)	Status epilepticus	7
5	64	M	Hematoma	174	80	8	6	1(10)	Recurrence of hemorrhage	4
6	46	F	Poststatus epilepticus	171	65	4	4	1(3)	Status epilepticus	2
7	47	M	Head trauma	172	82	3	3	1(8)	Temporal herniation	6
8	38	M	Head trauma	167	75	9	11	1 (9)	Cerebral edema with herniation	16
9	35	F	Subarachnoid hemorrhage	184	77	3	3	1(10)	Brain death	3
10	51	F	Subarachnoid hemorrhage	165	60	4	3	1(2)	Recurrence of hemorrhage	1
11	61	M	Postanoxic	181	70	6	4	1(8)	Septic shock	7
12	75	M	Subarachnoid hemorrhage	167	60	7	7	1(24)	Septic shock	9
13	43	M	Hematoma	169	70	10	11	3(30)	—	3
14	81	M	Hydrocephalus (neoplasm)	189	80	7	9	1(8)	Neoplasia	1
15	30	F	Head trauma	168	65	7	7	3(26)	—	4
Mean	55			170	69	6	6			6
SD	18			10	9	3	3			6

Abbreviations: GCS, Glasgow coma score; GOS, Glasgow outcome scale (1: dead; 2: vegetative state; 3: severe disability; 4: moderate disability; 5: good recovery); M, male; F, female.

Procedures

A small polyethylene catheter was inserted into a radial artery for therapeutic purposes.

Arterial and jugular bulb blood gases and hemoglobin saturations were measured using the ABL500 blood gas system and the OSM3 Hemoximeter (Radiometer, Copenhagen, Denmark).

The jugular blood flow was measured using continuous thermodilution (Fig. 1). Briefly, a fluid miscible with blood and at a lower temperature was in injected into the jugular bulb at a constant rate and the resulting change in temperature detected in a short distance downstream (25 mm). Conventionally this fluid has been referred to as the indicator, although from a rigorous scientific standpoint it is the cold (negative heat) that is the indicator, the injectate being only the vehicle for its introduction into the bloodstream. A continuous thermodilution catheter (CCS-7U-90B 7 Fr or Pepine Reiz 2.5 cm 8 Fr; Webster Laboratories, Baldwin Park, CA, U.S.A.) was inserted percutaneously in the jugular bulb of the patient using the technique of Jakobsen and Enevoldsen (1989). An introducer was inserted retrogradely into the internal jugular vein and the catheter was subsequently advanced until no further progression due to a palpable resistance was possible. With this procedure the tip was estimated to be located in the jugular bulb. The correct location of the tip was checked by x-ray examination in the first five patients. Thereafter, the major part of the introducer was removed from the vein and the catheter fixed by a bandage. The inclination of the head was set at +30°, and the position of the head was not changed throughout the study. We automated the technique of computation by connecting the temperature recorder box (Triple Channel Flowmeter CF-300, Webster Laboratories) of the thermodilution catheter to a computer through an analog/digital converter (RTI 800, Analog Device, Norwood, MA, U.S.A.). A program was written to compute instantaneously the blood flow when the temperature variation is recorded during the infusion of the indicator (Fig. 1). With the computerized procedure, a CBF measurement takes at most 40 s and the volume of indicator infused is 20 ml. The computation of flow is based on the assumption that the heat lost from the system between the site of injection and the site of detection is negligible and, therefore, that the heat loss by the blood equals the heat gained by the indicator.

Fig. 1.

Principle of the continuous thermodilution. Upper diagram: the catheter inside the jugular vein with the two thermistors. Lower diagram: temperature curves recorded by the internal and external thermistors.

Let V_B and V_I, be the volumes (in milliliters) of blood and indicator, respectively, participating in the process of mixing and heat exchange during a finite period; T_B, T₁, and T_M, the temperature of blood, indicator, and mixture of blood plus indicator, respectively, in degrees Celsius; S_B and S_I, the density of blood and indicator, respectively, in grams per cubic centimeter; and C_B and C_I, the specific heat of blood and indicator, respectively, in calories per gram per degree Celsius. The heat exchanged between the blood and indicator can then be expressed by the following equation:

\begin{aligned} (T_{B} - T_{M}) \times V_{B} \times S_{B} \times C_{B} = \\ (h e a t l o s t b y b l o o d) \\ (T_{M} - T_{I}) \times V_{I} \times S_{I} \times C_{I} \\ (h e a t g a i n e d b y i n d i c a t o r) \end{aligned}

(1)

From Eq. 1, if time is taken into consideration, volumes V_B and V_I, in the formula can be transformed into flows F_B and F_I (in milliliters per minute) of blood and indicator respectively. F_B can be computed as:

F_{n} = F_{I} (\frac{S_{I} \times C_{I}}{S_{B} \times C_{B}}) (\frac{T_{M} - T_{I}}{T_{B} - T_{M}})

(2)

When saline is used as indicator:

\frac{S_{I} \times C_{I}}{S_{B} \times C_{B}} = \frac{1.005 \times 0.997}{1.045 \times 0.87} = 1.10

(3)

The product S_B × C_B is virtually constant over a range of hematocrit from 30% to 60% because the change in S_B is balanced by a proportional change in C_B in the opposite direction (Fronek and Ganz, 1960).

Since mixing of blood and indicator must be attained by the kinetic energy of the injectate, the rate of injection of indicator is of critical importance to the accuracy of measurements. The injection rate has been estimated to be at least 10% of the flow to be measured (i.e., 30 ml min⁻¹ for a flow rate of 300 ml min⁻¹) (Ganz et al., 1971). For JBF measurements, saline was used as indicator and was injected at a rate of 38 ml min⁻¹ using a heavy duty infusion pump (model 940, Harvard Apparatus, South Natick, MA, U.S.A.).

The in vitro accuracy and precision of the thermodilution technique were first checked by comparing flow measured by thermodilution with volumetric flow in a circuit made of a Tygon tube (internal diameter of 8 mm) with saline at 37°C flowing through the tube, using a peristaltic pump (Juglabo U3 pump, Juglabo Labortechnik, Seelbach, Germany). Each measurement was repeated six times at eight preset volumetric flows ranging from 50 to 900 ml min⁻¹. Saline at room temperature was used as indicator. The injection rate was 38 ml min⁻¹ (50–4% of the flow).

Cerebral blood flow (CBF) was also measured using the Kety-Schmidt reference method (CBF_KS) with nitrous oxide (N₂O) as the indicator (Fig. 2). For the measurement of CBF_KS, 10% of N₂O was introduced into the patient's inspired gases and 10 timed samples of arterial and jugular bulb blood were collected during the first 12 min of N₂O saturation. N₂O was measured in arterial and jugular blood samples by gas chromatography (Ladron de Guevara et al., 1987). Briefly, blood samples were equilibrated during 45 min with pure nitrogen in the syringe used for blood sampling to allow diffusion of N₂O in the gaseous phase. Thereafter, N₂O concentration was measured in the gas phase by gas chromatography with an electron capture detector (HP gas Chromatograph 5890, Hewlett Packard, Palo Alto, CA, U.S.A.). Knowing the blood-gas partition coefficient for N₂O (0.471) (Kety, 1951), the volume of blood sample (4 ml), and the volume of nitrogen (16 ml), we can compute the in vivo concentration of N₂O, Measurements for arterial and jugular bulb concentrations were fitted using a best-fit double-exponential curve defined by least squares to allow computation of the area between the two curve [(A — V)dt] CBF_KETY was computed in a customary manner:

C B F_{K S} (m l min^{- 1} 100 g^{- 1}) = \frac{S \cdot 100 \cdot V_{\infty}}{\int_{t = 0}^{\infty} (A - V) d t}

(4)

where S is the N₂O blood-brain tissue partition coefficient (1.06) (Kety et al., 1948), V_∞ the jugular venous blood N₂O concentration when the concentrations reach a plateau (at most at the 20th min), and (A — V)dt the area between the arterial and venous N₂O concentration curves. Figure 2 shows an example of CBF calculation.

Fig. 2.

Arterial and jugular bulb blood concentrations of nitrous oxide in a patient during the inhalation of 10% nitrous oxide in air/oxygen. Each dot represents the actual concentration of nitrous oxide in blood. The solid lines represent the least-square best fit using a double-exponential curve extrapolated up to 20 min to obtain a plateau for the venous concentration.

To allow comparison with CBF expressed in milliliters per minute per 100 g, JBF measured by thermodilution, which reflects the venous outflow of one hemisphere, was doubled and divided by brain weight/100:

\begin{aligned} C B F_{T H} (m l min^{- 1} 100 g^{- 1}) = \\ \frac{J B F (m l min^{- 1}) \times 2 \times 100}{b r a i n w e i g h t (g)} \end{aligned}

(5)

where the brain weight was estimated as k* height (centimeters) with k = 8.3 for male and k = 8.0 for female (Spann and Dustmann, 1965).

Protocol

When a unilateral hemispheric brain lesion was observed at the CT scan, the jugular catheter was placed on that side. If bilateral hemispheric brain lesions were observed, the catheter was placed on the right side. The catheter was placed on the left side in five patients, and on the right side in ten. CBF determined by the Kety-Schmidt method and CBF by the jugular thermodilution technique were measured in each patient at four levels of P_aCO₂ (25, 30, 35, and 40 mm Hg) by adding CO₂ in the inspiratory circuit of the ventilator with ventilation kept constant. At each level of P_aCO₂, hemodynamic variables and blood gases were also measured.

Statistical analysis

The results are expressed as mean ± standard deviation (SD). Statistical analysis consisted of an analysis of variance for repeated measures. When the F ratio reached the significance level (p < 0.05), pairwise comparisons were performed using modified t tests (Winer, 1971). Correlation and least squares linear regression were calculated when indicated. Comparison of two methods measuring the same variable were also compared using the graphical method proposed by Bland and Altman (1986). To take into account that the measures were repeated four times in each patient, we used the method described by Poon (1988) to pool repeated measurements pertaining to several subjects.

RESULTS

Validation of technique

In vitro measurements showed that continuous thermodilution was an accurate and precise method to measure flow (coefficient of variability between 1.21 and 2.50%) (Fig. 3).

Fig. 3.

In vitro accuracy and precision of the measure of flow by thermodilution compared with volumetric flow.

The two methods were also compared using the graphical technique proposed by Bland and Altman (1986). Briefly, this technique is based on graphical methods using the plot of the difference between the two measures versus the mean of the two measures. The mean difference for the entire set of measurements reflects the agreement between the two methods and allows detection of bias of one method when compared with a gold standard. In addition, a 95% confidence interval can be computed as the mean difference ± 2 SD (see Figs. 4 and 6). The mean difference (i.e., the bias) for flow varying from 50 to 900 ml min⁻¹ is −1 ± 18 ml min⁻¹ (Fig. 4). These results confirm the good agreement between the two methods.

Fig. 4.

Agreement between flow measured by thermodilution and volumetric flow using Bland and Altman (1986) diagram. The mean difference is −1 ± 18 ml min⁻¹ (mean ± SD).

Fig. 5.

In vivo coefficient of variability for 60 triplicate measurements of jugular blood flow.

Fig. 6.

Comparison of CBF measured by Kety-Schmidt and thermodilution methods at P_aCO₂ 25, 30, 35, and 40 mm Hg in 15 patients. The solid line is the line of identity and the dashed line is the regression line for the 60 measures.

In vivo precision was computed using the coefficient of variability (CV = 100 · SD/mean) for triplicate measurements performed at four levels of P_aCO₂ in the 15 patients (mean CV for 60 measures = 4.1%) (Fig. 5).

CBF measured at four levels of P_aCO₂ in each patient by the Kety-Schmidt and the thermodilution methods gave an excellent correlation with the regression line (dashed line) not different from the line of identity (solid line) (Fig. 6).

The two methods were also compared using the graphical technique proposed by Bland and Altman (1986). The mean difference (i.e., the bias) for CBF is −0.9 ± 3.6 ml min⁻¹ 100 g⁻¹ (Fig. 7). These results confirm the good agreement between the two methods.

Fig. 7.

Agreement between cerebral blood flow measured by thermodilution and Kety-Schmidt methods using Bland and Altman (1986) diagram. The mean difference is −0.9 ± 3.6 ml min⁻¹ 100 g⁻¹ (mean ± SD).

Cerebral metabolic changes in response to changes in CBF

CMRO₂ in our patients (between 1.68 ± 0.98 and 1.87 ± 0.97 ml min⁻¹ 100 g⁻¹) is lower than that measured in normal subjects (3.3 ± 0.2 ml min⁻¹ 100 g⁻¹ for CBF 56.8 ± 5.6 ml min⁻¹ 100 g⁻¹ in the awake state, and 2.46 ± 0.27 ml min⁻¹ 100 g⁻¹ during anesthesia) (Van der Linden et al., 1991). The lower value in the present study is probably due to the severe brain damage of the patients as assessed by the GCS (Table 1).

Reduction in CBF induced by the decrease in P_aCO₂ leads to no change in CMRO₂ due to an increased cerebral extraction for oxygen (Table 2). This increase in cerebral extraction for oxygen is reflected by the huge decrease in jugular bulb oxygen saturation from 75 ± 10 to 55 ± 13%.

TabLE 2.

Effects of P_aCO₂ on cerebral hemodynamics, and arterial and jugular blood gases in 15 comatose patients

Variables	1	2	3	4
pH_a	7.53 ± 0.04	7.48 ± 0.04^b,d	7.42 ± 0.03^b,d	7.37 ± 0.04^b,d
P_aO2 (mm Hg)	144 ± 51	159 ± 49	169 ± 54^a	174 ± 44^b
P_aCO₂ (mm Hg)	25 ± 1	30 ± 1^b,d	35 ± 1^b,d	40 ± 1^b,d
pH_j	7.45 ± 0.03	7.41 ± 0.03^b,d	7.36 ± 0.03^b,d	7.32 ± 0.04^b,d
P_jO2 (mm Hg)	30 ± 6	35 ± 6^b,d	40 ± 7^b,d	47 ± 8^b,d
P_jCO2 (mm Hg)	33 ± 2	38 ± 3^b,d	42 ± 3^b,d	47 ± 3^b,d
S_jO2 (%)	55 ± 13	62 ± 12^b,d	68 ± 11^b,d	75 ± 10^b,d
JBF_th (ml min⁻¹)	193 ± 85	255 ± 91^a,c	300 ± 94^b,c	385 ± 187^b,c
CBF_th (ml min⁻¹ 100 g⁻¹)	28 ± 13	37 ± 14^a,c	44 ± 15^b,c	56 ± 28^b,d
CBF_ks (ml min⁻¹ 100 g⁻¹)	29 ± 11	36 ± 14	41 ± 14^a	55 ± 28^b,d
D_a-jO2 ml/100 ml)	5.9 ± 2.0	5.1 ± 1.8^b,d	4.3 ± 1.6^b,d	3.3 ±1.6^b,d
CMRO₂ (ml min⁻¹ 100 g⁻¹)	1.68 ± 0.98	1.87 ± 0.97	1.85 ± 1.00	1.70 ± 0.80
O₂EXTc (%)	45 ± 13	39 ± 12^b,d	32 ± 11^b,d	26 ± 11^b,d
P_sa (mm Hg)	98 ± 29	98 ± 28	97 ± 20	99 ± 17
P_j (mm Hg)	17 ± 2	17 ± 3	18 ± 3^a,c	20 ± 4^b,d
VE (L/min)	10.7 ± 3.7	10.2 ± 3.5	10.2 ± 3.5	10.2 ± 3.5

Values are expressed in mean ± SD; n = 15 patients. pH_a, arterial pH; P_aO2, arterial P_O2; P_aCO₂, arterial P_CO2; pH_j, jugular bulb pH; P_jO2, jugular bulb P_O2; P_jCO2, jugular bulb P_CO2; S_jO2, jugular bulb saturation; JBF_th, jugular blood flow by thermodilution; CBF_th, cerebral blood flow by thermodilution; CBF_ks, cerebral blood flow by the Kety-Schmidt method; D_a-jO2, arteriojugular bulb O₂ contents difference; CMRO₂, cerebral metabolic rate for O₂; O₂EXTc, cerebral O₂ extraction; P_sa, mean systemic arterial pressure; P_j, jugular bulb pressure; VE, minute ventilation.

p < 0.05.

p < 0.01 compared with the first column.

p < 0.05.

p < 0.01 compared with the preceding column.

Effect of increased inspired PCO2 on arterial oxygenation

Arterial P_O2 increases from 144 ± 51 to 174 ± 44 mm Hg when P_aCO₂ increases with no changes in minute ventilation nor in cardiac output. Inspired CO₂ has been shown to improve arterial oxygenation mainly by reducing dead space ventilation and, to a smaller extent, the Bohr effect and improvement in ventilation-perfusion relationships (Domino et al., 1993; Swenson et al., 1994).

Cerebral CO2 vasoreactivity

In human cerebral circulation, an acute increase in P_aCO₂ leads to a 2–6% increase in CBF (Olesen et al., 1971; Tominaga et al., 1976) per millimeter of mercury change in P_aCO₂. The carbon dioxide reactivity usually is best described by an exponential curve in the physiological range of P_aCO₂ (Olesen et al., 1971; Tominaga et al., 1976). In the present study, a normal response of CBF to increase in P_aCO₂ was defined as a significant exponential In CBF:P_aCO₂ relationship (R > 0.900, p < 0.05 for four points, one-sided t test) with a slope above 0.02. This value was chosen as the lower bound of the 95% confidence interval for the normal values (0.06 ± 0.04, mean ± 2 SD) measured in 16 subjects (Tominaga et al., 1976). Among the 15 comatose patients, 11 had a significant exponential In CBF:P_aCO₂ relationship, among which 10 had a normal CO₂ vasoreactivity.

DISCUSSION

The results of the present study suggest that the jugular blood flow measured by continuous thermodilution is a precise and accurate method to estimate cerebral blood flow. Moreover, this technique allows frequently repeatable measurements at the bedside.

The percutaneous retrograde cannulation using the Seldinger technique (Seldinger, 1953) and the anterior method described by Jakobsen and Enevoldsen (1989) is an easy procedure and allows an indwelling catheter to remain in situ for the duration of the study. We evaluated prospectively the complications due to the catheter in our patients. In one patient the internal carotid artery was punctured, creating a small hematoma without other complication. In two other patients the internal jugular vein was punctured at the second attempt. In one patient, at the removal of the catheter a small clot was observed at the tip of the catheter without jugular vein thrombosis. The high mortality rate observed in our series cannot be attributed to the procedure but to the severity of the brain damage. Moreover, the mean Glasgow coma score (GCS) was similar the day after the study, and death occurred at a mean of eleven days after the study (Table 1).

The Kety-Schmidt technique has been validated against the bubble flowmeter (Kety and Schmidt, 1948). Interestingly, if we apply the Bland and Altman method to the data comparing the bubble flowmeter and the Kety-Schmidt method in rhesus monkeys, we obtain a mean difference of 3.1 ± 5.4 ml min⁻¹ 100 g⁻¹ (mean ± SD), indicating a slight underestimation of flow by the Kety-Schmidt procedure. If we assume a monkey brain weight of 90 g in mean (Dumke and Schmidt, 1943), these figures indicate that the thermodilution, when compared with a volumetric flow, is more precise and more accurate than the Kety-Schmidt method (Fig. 4).

The duration of administration of nitrous oxide to ensure blood-brain equilibrium has also been a source of controversy (Lassen and Klee, 1965; Sharples et al., 1991). It is clear from a theoretical standpoint that the period required is dependent upon the cerebral blood flow (Lassen and Klee, 1965; Sharples et al., 1991). In our patients we administered N₂O for 12 min. If the cerebral venous concentration of N₂O is found not to be equal to the arterial concentration at the end of the measurement period, the venous curve is extrapolated (Lassen and Klee, 1965; Sharples et al., 1991) to within 5% of the arterial value using a best-fit double-exponential curve defined by least squares. Nevertheless, the Kety and Schmidt method has been reported to be flawed for low CBF. On the other hand, if we refer to the in vitro calibration curve (Fig. 3), the continuous thermodilution is accurate even for very low flow.

Continuous thermodilution was first described for the measurement of coronary sinus blood flow (Ganz et al., 1971). Continuous infusion was chosen in preference to bolus injection because it facilitates the computation (measurement of temperature levels rather than an area under the thermodilution curve). The validity of equation 2 (see Methods) assumes: (a) that there is adequate mixing between the tip of the catheter where the indicator comes into bloodstream and the site of the measurement of the temperature of the blood-indicator mixture (external thermistor, Fig. 1), (b) that the blood temperature remains constant during the measure, (c) that the indicator temperature at entry into the bloodstream is known accurately, and (d) that there is negligible net heat exchange from blood mixture to vessel walls between the tip of the catheter and the site of the measurement of the temperature of the blood-indicator mixture. To obtain a good mixing with minimal heat loss through the vessel walls, the distance between the tip and the external thermistor must be adequate, i.e., sufficiently long to allow heat exchange between indicator (i.e., good mixing) and not too long to minimize heat loss through the vessel walls. The anatomical requirement of injection in the jugular bulb and the external thermistor above the junction with the common facial vein constrains the distance to about 3 or 4 cm maximum. In vitro experiments have shown that a distance of 25 mm (catheters used in the present study) is usually adequate when the injection rate is above 25 ml min⁻¹ to minimize heat loss and to obtain an adequate mixing (Wilson and Hasley, 1970). Ideally the injection rate should be at least 10% of the flow to be measured (Ganz et al., 1971). However, in vitro experiments of the present study showed that injection rates between 50% and 4% of the flow to be measured gave accurate measurements of flow ranging between 50 and 900 ml min⁻¹ (Fig. 3). These flows correspond to JBF equivalent to CBF between 7 and 140 ml min⁻¹ 100 g⁻¹, the extreme values measured in our patients. The constant temperature of the blood is a reasonable assumption for the short period of time for one measurement. The temperature of the indicator is measured accurately owing to the location of the internal thermistor.

JBF measures hemispheric CBF, and because of the confluence of sinuses, it is really a global measure of CBF (Young and Ornstein, 1994). There is contamination from the contralateral hemisphere and each bulb represents roughly one third of the contralateral hemisphere drainage in normal subjects (Shenkin et al., 1948). An average of 2.7% (range 0–6.6%) of the blood in the jugular bulb is derived from extracerebral structures (Shenkin et al., 1948). Wilson et al. (1972) measured jugular blood flow on both sides by thermodilution and found a good agreement within 12–15% when compared with total cerebral blood flow measured by the dye dilution method. These results are in agreement with the study of Meyer et al. (1963), in which internal jugular flow measured with an electromagnetic flowmeter and extrapolated to flow per unit of mass was compared with predicted values from the nitrous oxide method. These results suggest that venous outflow via the emissaries and vertebral venous plexus was relatively small. This is at variance with the study of Epstein et al. (1970), who, by angiographic studies, demonstrated a substantial outflow in the vertebral venous plexus in upright rhesus monkeys. Eckenhoff (1966) and Shenkin et al. (1948) made similar observations in man. However, Epstein et al. (1970) and Eckenhoff (1966) pointed out that this outflow was reduced in the supine position. Therefore, it can be concluded that the venous jugular blood flow is slightly contaminated by extracerebral blood and can be considered as a reliable indicator of CBF.

CBF is normally tightly coupled to the rate of cerebral metabolism, a phenomenon commonly referred to as metabolic autoregulation (Bouma and Muizelaar, 1992). In our patients in steady-state conditions, once CMRO₂ has been set to the metabolic demand, reduction in CBF induced by the decrease in P_aCO₂ leads to no change in CMRO₂ with an increased cerebral extraction for oxygen. However, outside physiologic ranges of P_aCO₂, a change in CBF induced by CO₂ vasoreactivity leads to changes in CMRO₂ (Yang and Krasney, 1965; Cold, 1989).

Cerebral vasoreactivity to carbon dioxide is defined as the physiological increase of cerebral blood flow to increases in arterial P_CO2. In human cerebral circulation, an acute increase in P_aCO₂ leads to a 2–6% increase in CBF (Olesen et al., 1971; Tominaga et al., 1976) or in the middle cerebral artery flow velocity (Markwalder et al., 1984) per millimeter of mercury change in P_aCO₂. The shape of the CBF:P_aCO₂ curve in animal experiments was found to be sigmoid when including rather extreme P_aCO₂ values (Reivich, 1964). Using only the interval 20–60 mm Hg, it has been found to fit a straight line with reasonable accuracy in normal man (Alexander et al., 1964). However, more recent data suggest that the carbon dioxide reactivity usually is best described by an exponential curve in the physiological range of P_aCO₂ in animals (Waltz, 1970) as well as in humans (Olesen et al., 1971; Tominaga et al., 1976; Markwalder et al., 1984). In severe head trauma, interesting information can be obtained from CO₂ vasoreactivity tests (Fieschi et al., 1974; Messeter et al., 1986; Schalen et al., 1991; Nordstrom et al., 1988). The clinical impact of hypocapnic CO₂ reactivity seems to be considerable, since the loss of cerebral CO₂ vasoreactivity is associated with a poor outcome in head trauma (Schalen et al., 1991). Furthermore, knowledge of the integrity of the CO₂ response is helpful in determining the potential effectiveness of hyperventilation and/or barbiturate therapy for control of raised intracranial pressure in individual patients (Nordstrom et al., 1988). In our patients, ten had normal CO₂ vasoreactivity while five had an altered response of CBF to increase in P_aCO₂. Among the ten subjects with normal response, two are still alive and one patient with altered response survived. However, no definite conclusion can be drawn from this small number of patients.

In conclusion, the jugular continuous thermodilution technique is a reliable method for estimating CBF at the bedside in patients in coma. This technique allows repeated measurements and is easy to perform over a short period of time in the intensive care setting. In addition, the catheter in the jugular bulb allows for blood sampling and brain metabolic studies.

Footnotes

Abbreviations used

Acknowledgment:

The authors are grateful to Mrs. Pascale Jespers for her technical assistance. They are also indebted to the nurses and the technicians of the intensive care department.

References

Aaslid

Markwalder

Nornes

(1982) Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 57:769–774.

Alexander

Wollman

Cohen

(1964) Cerebrovascular response to PaCO2 during halothane anesthesia in man. J Appl Physiol 19:561–565.

Bland

Altman

(1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307–310.

Bouma

Muizelaar

(1992) Cerebral blood flow, cerebral blood volume, and cerebrovascular reactivity after severe head injury. J Neurotrauma 9(suppl 1):S333–S348.

Cohan

Mun

Petite

(1989) Cerebral blood flow in humans following resuscitation from cardiac arrest. Stroke 20:761–765.

Cold

(1989) Does acute hyperventilation provoke cerebral oligaemia in comatose patients after acute head injury? Acta Neurochir (Wien) 96:100–106.

Domino

Eisenstein

Hlastala

(1993) Hypocapnia worsens arterial blood oxygenation and increases VA/Q heterogeneity in canine pulmonary edema. Anesthesiology 78:91–99.

Dumke

Schmidt

(1943) Quantitative measurements of cerebral blood flow in the macaque monkey. Am J Physiol 138:421–431.

Eckenhoff

(1966) Circulatory control in the surgical patient. Ann R Coll Surg Engl 39:67–83.

10.

Edvinsson

MacKenzie

McCulloch

(1993) Cerebral Blood Flow and Metabolism, New York, Raven Press, pp 497–523.

11.

Epstein

Linde

Crampton

(1970) The vertebral venous plexus as a major cerebral venous outflow tract. Anesthesiology 32:332–337.

12.

Fieschi

Battistini

Beduschi

(1974). Regional cerebral blood flow and intraventricular pressure in acute head injuries. J Neurol Neurosurg Psychiatry 37:1,378–1,388.

13.

Fronek

Ganz

(1960) Measurement of flow in single blood vessels including cardiac output by local thermodilution. Cire Res 8:175–182.

14.

Ganz

Tamura

Marcus

(1971) Measurement of coronary sinus blood flow by continuous thermodilution in man. Circulation 44:181–195.

15.

Jakobsen

Enevoldsen

(1989) Retrograde catheterization of the right internal jugular vein for serial measurements of cerebral venous oxygen content. J Cereb Blood Flow Metab 9:717–720.

16.

Jennett

Bond

(1975) Assessment of outcome after severe brain damage. Lancet 1:480–484.

17.

Kety

(1951) The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 3:1–41.

18.

Kety

Schmidt

(1945) The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 143:53–65.

19.

Kety

Schmidt

(1948) The nitrous oxide method for the quantitative determination of cerebral blood flow in man: Theory, procedure and normal values. J Clin Invest 27:476–483.

20.

Kety

Harmel

Brommell

(1948) The solubility of nitrous oxide in blood and brain. J Biol Chem 173:487–496.

21.

Kozam

Landau

Cubina

(1970) Solubility of nitrous oxide in biologic fluid and myocardium. J Appl Physiol 29:593–597.

22.

Ladron de Guevara

Adaya Godoy

Cortinas de Nava

(1987) Quantification of halothane, enflurane and nitrous oxide by means of a simple gas chromatographic method. J Chromatogr 403:350–354.

23.

Lassen

Klee

(1965) Cerebral blood flow determined by saturation and desaturation with krypton: An evaluation of the inert gas method of Kety-Schmidt. Circ Res 26:26–32.

24.

Markwalder

Grolimund

Seiler

Roth

Aaslid

(1984) Dependency of blood flow velocity in the middle cerebral artery on end-tidal carbon dioxide partial pressure. A transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 4:368–372.

25.

Messeter

Nordstrom

Sundbarg

(1986) Cerebral hemodynamics in patients with acute severe head trauma. J Neurosurg 64:231–237.

26.

Meyer

Ishikawa

Lee

(1963) Quantitative measurement of cerebral blood flow with electromagnetic flowmeters, recording internal jugular venous flow in monkey and man. Trans Am Neurol Assoc 88:78–83.

27.

Nordstrom

Messeter

Sundbarg

(1988) Cerebral blood flow, vasoreactivity, and oxygen consumption during barbiturate therapy in severe traumatic brain lesions. J Neurosurg 68:424–431.

28.

Obrist

Thompson

Jr Wang

(1975) Regional cerebral blood flow estimated by 133-xenon inhalation. Stroke 6:245–256.

29.

Obrist

Langfitt

Jaggi

(1984) Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 61:241–253.

30.

Olesen

Paulson

Lassen

(1971) Regional cerebral blood flow in man determined by the initial slope of the clearance of intra-arterially injected 133Xe. Stroke 2:519–540.

31.

Poon

(1988) Analysis of linear and mildly nonlinear relationships using pooled subject data. J Appl Physiol 64:854–859.

32.

Prough

(1991) Brain monitoring. In: Critical Care. State of the Art ( Taylor

Shoemaker

, eds), Anaheim, CA, The Society of Critical Care Medicine, pp 157–196.

33.

Reivich

(1964) Arterial PCO2 and cerebral hemodynamics. Am J Physiol 206:25–35.

34.

Schalen

Messeter

Nordstrom

(1991) Cerebral vasoreactivity and the prediction of outcome in severe traumatic brain lesions. Acta Anaesthesiol Scand, 35:113–122.

35.

Seldinger

(1953) Catheter replacement of the needle in percutaneous arteriography. Acta Radiol 39:368–376.

36.

Sharples

Stuart

Aynsley-Green

(1991) A practical method of serial bedside measurement of cerebral blood flow and metabolism during neurointensive care. Arch Dis Child 66:1326–1,332.

37.

Shenkin

Harmel

Kety

(1948) Dynamic anatomy of the cerebral circulation. Arch Neurol Psychiatr 60:240–252.

38.

Simonson

Piantadosi

(1995) Near infrared spectroscopy for monitoring tissue oxygenation in the critical care setting. Curr Opin Crit Care 1:217–223.

39.

Spann

Dustmann

(1965) Das menschliche Hirngewicht und seine Abhägigkeit von Lebensalter, Körperlange, Todesursache und Beruf. Dtsch Z Med 56:299–317.

40.

Swenson

Robertson

Hlastala

(1994) Effects of inspired carbon dioxide on ventilation-perfusion matching in normoxia, hypoxia, and hyperoxia. Am J Respir Crit Care Med 149:1,563–1,569.

41.

Teasdale

Jennett

(1974) Assessment of coma and impaired consciousness. A practical scale. Lancet 2:81–84.

42.

Tominaga

Strandgaard

Uemura

(1976) Cerebrovascular CO2 reactivity in normotensive and hypertensive man. Stroke 7:507–510.

43.

Van der Linden

Wesslen

Ekroth

(1991) Transcranial Doppler-estimated versus thermodilution-estimated cerebral blood flow during cardiac operations. Influence of temperature and arterial carbon dioxide tension. J Thorac Cardiovasc Surg 102:95–102.

44.

Waltz

(1970) Effect of PaCO2 on blood flow and microvasculature of ischemic and non ischemic cerebral cortex. Stroke 1:27–37.

45.

Wilson

Halsey

Jr Vitek

(1972) Validation of jugular venous flow as an index of total cerebral blood flow. Stroke 3:300–321.

46.

Wilson

Halsey

Jr (1970) Bilateral jugular venous blood flow by thermal dilution. Stroke 1:348–355.

47.

Winer

(1971) Statistical Principles in Experimental Design, New York, McGraw-Hill, pp 514–603.

48.

Yang

Krasney

(1995) Cerebral blood flow and metabolic responses to sustained hypercapnia in awake sheep. J Cereb Blood Flow Metab 15:115–123.

49.

Young

Ornstein

(1994) Cerebral and spinal cord blood flow. In: Anesthesia and Neurosurgery ( Cotterell

Smith

, eds), St. Louis, MO, Mosby, pp 17–58.