Noninvasive Measurement of Central Venous Pressure by Neck Inductive Plethysmography* Konrad E. Bloch, M.D.; Broce P. Krieger, M.D., F.C.C.P.; and Marvin A. Sackner, M.D., F.C.C.P.
Bedside estimation of the height at which the internal jugular veins collapse, referenced to a standard hemodynamic location, theoretically reOects central venous pressure. This method has never been demonstrated as accurate when compared to invasive CVP measurements because of the great clinical skills and time required to visually identify the internal jugular venous waveform. Since the principles of the bedside method are sound, we utilized them in conjunction with the neck inductive plethysmograph, a device which has the capability of recording internal jugular venous and carotid arterial waveforms. The respiratory distortion of these vascular waveforms was eliminated by employing a digital bandpass Glter, making it easy to identify the venous and arterial waveforms on the videoscreen of a personal computer. The upper torso was positioned while observing the videoscreen until the vascular waveform was seen as a mixed arteriovenous waveform, signifying intermittent internal jugular venous collapse. The height of the internal jugular vein above the phlebostatic axis was ob-
venous pressure is generally measured from Central a catheter placed within the superior vena cava near the entrance to the right atrium. It may also be estimated at the bedside by visual inspection of the internal jugular venous pulse according to the following principles. 1•2 Standing on the patient's right side, the examiner places his/her right hand while holding a flashlight on the patient's sternum and shines the light tangentially across the right side of the neck. Shadows of the right internal jugular venous pulsations are cast on the sheet behind the patient. The distance between the uppermost level of visibility of the internal (not external) jugular pulsations and a standard hemodynamic reference location, viz, manubriosternal angle, midchest at fourth intercostal space, etc, provides an estimate of CVP. Since the bedside method is dependent upon identifying the right internal jugular venous waveform, great clinical skills and a major time commitment are required. The purpose of the present investigation was to develop and evaluate a noninvasive method for esti*From the Division of Pulmonary Disease, University of Miami School of Medicine at Mount Sinai Medical Center, Miami Beach. Manuscript received November 13; revision accepted December
21.
Reprint requests: Dr. Saclmer, 300 Wist Rioo Alto Drive, Miami Beach 33139
tained by external measurement and recorded as CVPni in cm ff.O. In 43 patients, 86 percent of CVPni (NIP) values fell within 20 percent CVPi (invasive catheter measurements) over a range of CVP from 0 to 19 cm ff.O. In an additional seven patients, CVPi was greater than the height that the upper torso could be elevated and an arterial
waveform could not be obtained. Here, CVPni was recorded as the value at least exceeding the value measured. In two other patients, obstruction of an internal jugular vein gave spuriously low values of CVPni. Our study indicates that this new neck inductive plethysmographic method is accurate compared to invasive catheter measurements of CVP and should serve as a safe, noninvasive alternative in situations where such measurements are required. (Chae 1991; 100:371-75)
I
=:tnl -
preuure; NIP= neek induetive
pledays-1
mating CVP based upon principles of bedside assessment without requiring accurate visual identification of the right internal jugular venous waveform. MATERIALS AND METHODS
Princi11les
The method hinges upon the observation that a NIP transducer, which measures changes in cross-sectional area of the underlying part, 3 provides analog wavefurms of vascular pulsations with superimposed respiratory variations due to changes of intrapleural pressure•.s and respiratory accessory muscular contractions. These vascular waveforms, which depict volume representations of either arterial or venous pressure pulses, closely resemble pressure wavefurms in appearance and timing. Respiratory variations tend to distort their appearance but can be eliminated by breathbolding or by ensemble-averaging during tidal breathing. In patients, breathholding is impractical and ensemble-averaging does not accurately depict waveforms with concurrent cardiac arrhythmias. Because the spectral content of respiratory and vascular waveforms overlap, a real time digital <er is necessary to achieve beat by beat display of the vascular wavefurms with elimination of respiratory variability to decide whether venous or arterial wavefurms are present. When the neck veins are partially distended, as normally occurs in the supine posture, their compliance, which Is higher than the arteries, allows greater transmission of volume pulsations to the NIP transducer, and a typical jugular venous wavefurm is depicted (Fig I). The NIP reftects venous pulsations in all veins underlying its transducer, but since the internal jugular veins are responsible for the greatest portion of venous return, the NIP signal can be CHEST I 100 I 2 I AUGUST, 1991
m
SUPINE ECG
Pnlp-v
········································5E·i1:fieciJ"iiENT................................................. 0
ECG
4cm ABOVE PHLEBOSTATIC AXIS ECG F1cuRE 1. In the supine posture, NIP depicts a jugular venous pulse. In the semirecumbent posture above the collapse point of the internal jugular vein, NIP depicts a carotid arterial pulse with the upstroke ( U) and incisura (I) denoted. At the point of internal jugular venous collapse, 4 cm H10 above the phlebostatic axis, a mixed arteriovenous pulse is depicted, viz, both a prominent "a" wave and an incisura are present.
Pnl1Hv
u 0
considered to primarily reflect internal jugular venous waveforms. In semirecumbant and erect postures, the neck veins normally collapse to slit-like passages leaving only the arterial volume pulsations as a remainder (Fig 1). With the patient placed in a posture at a point between these two extremes, the NIP vascular waveforms transform from venous to mixed arteriovenous or arterial waveforms and vice-versa (Fig 1). Tiie hydrostatic pressure exerted by the column of blood between the collapse point of the venous pulse and the right atrium represents the mean right atrial or central venous pressure. Neglecting a small error due to the specific gravity of blood (c.1.o.55), the CVP can, therefore, be measured externally as the vertical distance between a hemodynamic reference point (eg, phlebostatic axis) and the intersection of the internal jugular vein with the NIP transducer. 11ie internal jugular vein projects to the surface of the skin as a line drawn from the medial end of the clavicle to the interval between the ramus of the mandible and the mastoid process.• Tiie phlebostatic axis is taken as a point on the midaxillary line at the fourth interoostal space.' 11ie precision of estimating CVP with NIP (CVPni) was tested in four normal men and one woman, ages 24 to 58 years. Tiie vertical distance in which it was still possible to observe an arteriovenous waveform was 0.3 cm, SD JYl cm. In five consecutive measurements in each of the subjects, with CVPni ranging from 2. 7 to 7 .1 cm H10, the standard error of the estimate fell between 0.1 and 0.2cm H.o. lbtientB
Fifty-two critically ill patients (30 men, 22 women, mean age 71 years, range 31 to f17 years) were studied. They required intensive care for cardiovascular surgery (20), abdominal surgery (9), respi-
372
ratory failure (8), cardiac failure (7), trauma (2), and other nonsurgical conditions (6). Twenty-seven patients were on mechanical ventilators, and 33 were comatose or sedated to a degree that excluded verbal communication. Cardiac arrhythmias were present as premature ventricular contractions in 12 patients, as atrial flutter or 6brillation in seven, and as second or third degree heart blocks in three. Either a central venous or a Swan-Ganz catheter had been inserted for clinical management via the subclavian (30 patients), the internal jugular (20 patients), or the femoral vein (2 patients) prior to requesting informed consent from the attending physician and the patient or the patient's family to perform the noninvasive CVP measurement.
Procedure Tiie patient was placed in the supine posture. Tiie phlebostatic axis and the course of the right or let\ internal jugular veins were marked on the skin with ink. An inductive plethysmograph transducer, 1.5 cm in height, containing 6ve inductive coils in a sinusoidal array, was placed around the neck. After connecting the transducer to the oscillator module and to the demodulator/ampli6er circuitry (Respitrak), the NIP analog signal along with an ele11int and its value recordt>d. The tubing was then pluc.-ed next to the neck adjusting the water level to the intersection between tht' course of the internal jugular vein and the center of the NIP trw1sdm.-er and its value recorded. The diflerem.-e between these two readings was taken as CVP in cm 11,0.
The Mann-Whitney r.mk sum test and the Wilcoxon paired rank test were applied to compare mean differences between the ..open" and ..blinded" series and between the two different examiners, respe17.7 >11.0 >21.0 >15.8 >18.4 >18.0
FIGURE 3. Identity plot of CVPni (NIP) vs CVPi (intravascular catheter). The open circles denote the "open" and the solid circles the "blinded'' series. The rectangles denote the two patients with obstruction of the right internal jugular vein who showed much lesser _ _ _ _ _ _ ___. pressure with CVPni than CVPi. The inserted table lists the values of CVPni in whom an arterial waveform (Pnip-a) was unobtainable; although exact measurements of CVP could not be recorded with the NIP 20 24 method, all values were greater than normal in keeping witil CVPi measurements.
ECG
CVPni was taken as greater than the maximal achievable vertical distance between the phlebostatic axis and the NIP transducer (Tuble in Fig 3). There was no difference between trends of CVPi vs CVPni resulting from 14of15 changes in PEEP (3 to 10 cm H20) in ten mechanically ventilated patients (Fig 5). In 13 patients, independent measurements of CVPni by two different examiners revealed a mean difference (CVPni - CVPi) relative to CVPi of 8 percent for the examiner who performed the first and 7 percent for the examiner for the second test (p =NS).
Pnlp-v
LINE OF IDENTITY
the values of CVPni and CVPi agreed closely, 11.0 and 9.5 cm H20, respectively. In seven patients in whom the NIP showed a venous waveform despite maximal elevation of the upper part of the body, INDWELLING INTERNAL JUGULAR CATHETER
AFTER REMOVAL OF INTERNAL JUGULAR CATHETER
:c E ,g,
ECG
•
2
0Cll
0
c
A.
(;
Pnlp-v
Bedside estimation of the height at which the internal jugular veins collapse, referenced to a standard hemodynamic location, theoretically reOects central venous pressure. This method has never been demonstrated as accurate when compared to invasive CVP measurements because of the great clinical skills and time required to visually identify the internal jugular venous waveform. Since the principles of the bedside method are sound, we utilized them in conjunction with the neck inductive plethysmograph, a device which has the capability of recording internal jugular venous and carotid arterial waveforms. The respiratory distortion of these vascular waveforms was eliminated by employing a digital bandpass Glter, making it easy to identify the venous and arterial waveforms on the videoscreen of a personal computer. The upper torso was positioned while observing the videoscreen until the vascular waveform was seen as a mixed arteriovenous waveform, signifying intermittent internal jugular venous collapse. The height of the internal jugular vein above the phlebostatic axis was ob-
venous pressure is generally measured from Central a catheter placed within the superior vena cava near the entrance to the right atrium. It may also be estimated at the bedside by visual inspection of the internal jugular venous pulse according to the following principles. 1•2 Standing on the patient's right side, the examiner places his/her right hand while holding a flashlight on the patient's sternum and shines the light tangentially across the right side of the neck. Shadows of the right internal jugular venous pulsations are cast on the sheet behind the patient. The distance between the uppermost level of visibility of the internal (not external) jugular pulsations and a standard hemodynamic reference location, viz, manubriosternal angle, midchest at fourth intercostal space, etc, provides an estimate of CVP. Since the bedside method is dependent upon identifying the right internal jugular venous waveform, great clinical skills and a major time commitment are required. The purpose of the present investigation was to develop and evaluate a noninvasive method for esti*From the Division of Pulmonary Disease, University of Miami School of Medicine at Mount Sinai Medical Center, Miami Beach. Manuscript received November 13; revision accepted December
21.
Reprint requests: Dr. Saclmer, 300 Wist Rioo Alto Drive, Miami Beach 33139
tained by external measurement and recorded as CVPni in cm ff.O. In 43 patients, 86 percent of CVPni (NIP) values fell within 20 percent CVPi (invasive catheter measurements) over a range of CVP from 0 to 19 cm ff.O. In an additional seven patients, CVPi was greater than the height that the upper torso could be elevated and an arterial
waveform could not be obtained. Here, CVPni was recorded as the value at least exceeding the value measured. In two other patients, obstruction of an internal jugular vein gave spuriously low values of CVPni. Our study indicates that this new neck inductive plethysmographic method is accurate compared to invasive catheter measurements of CVP and should serve as a safe, noninvasive alternative in situations where such measurements are required. (Chae 1991; 100:371-75)
I
=:tnl -
preuure; NIP= neek induetive
pledays-1
mating CVP based upon principles of bedside assessment without requiring accurate visual identification of the right internal jugular venous waveform. MATERIALS AND METHODS
Princi11les
The method hinges upon the observation that a NIP transducer, which measures changes in cross-sectional area of the underlying part, 3 provides analog wavefurms of vascular pulsations with superimposed respiratory variations due to changes of intrapleural pressure•.s and respiratory accessory muscular contractions. These vascular waveforms, which depict volume representations of either arterial or venous pressure pulses, closely resemble pressure wavefurms in appearance and timing. Respiratory variations tend to distort their appearance but can be eliminated by breathbolding or by ensemble-averaging during tidal breathing. In patients, breathholding is impractical and ensemble-averaging does not accurately depict waveforms with concurrent cardiac arrhythmias. Because the spectral content of respiratory and vascular waveforms overlap, a real time digital <er is necessary to achieve beat by beat display of the vascular wavefurms with elimination of respiratory variability to decide whether venous or arterial wavefurms are present. When the neck veins are partially distended, as normally occurs in the supine posture, their compliance, which Is higher than the arteries, allows greater transmission of volume pulsations to the NIP transducer, and a typical jugular venous wavefurm is depicted (Fig I). The NIP reftects venous pulsations in all veins underlying its transducer, but since the internal jugular veins are responsible for the greatest portion of venous return, the NIP signal can be CHEST I 100 I 2 I AUGUST, 1991
m
SUPINE ECG
Pnlp-v
········································5E·i1:fieciJ"iiENT................................................. 0
ECG
4cm ABOVE PHLEBOSTATIC AXIS ECG F1cuRE 1. In the supine posture, NIP depicts a jugular venous pulse. In the semirecumbent posture above the collapse point of the internal jugular vein, NIP depicts a carotid arterial pulse with the upstroke ( U) and incisura (I) denoted. At the point of internal jugular venous collapse, 4 cm H10 above the phlebostatic axis, a mixed arteriovenous pulse is depicted, viz, both a prominent "a" wave and an incisura are present.
Pnl1Hv
u 0
considered to primarily reflect internal jugular venous waveforms. In semirecumbant and erect postures, the neck veins normally collapse to slit-like passages leaving only the arterial volume pulsations as a remainder (Fig 1). With the patient placed in a posture at a point between these two extremes, the NIP vascular waveforms transform from venous to mixed arteriovenous or arterial waveforms and vice-versa (Fig 1). Tiie hydrostatic pressure exerted by the column of blood between the collapse point of the venous pulse and the right atrium represents the mean right atrial or central venous pressure. Neglecting a small error due to the specific gravity of blood (c.1.o.55), the CVP can, therefore, be measured externally as the vertical distance between a hemodynamic reference point (eg, phlebostatic axis) and the intersection of the internal jugular vein with the NIP transducer. 11ie internal jugular vein projects to the surface of the skin as a line drawn from the medial end of the clavicle to the interval between the ramus of the mandible and the mastoid process.• Tiie phlebostatic axis is taken as a point on the midaxillary line at the fourth interoostal space.' 11ie precision of estimating CVP with NIP (CVPni) was tested in four normal men and one woman, ages 24 to 58 years. Tiie vertical distance in which it was still possible to observe an arteriovenous waveform was 0.3 cm, SD JYl cm. In five consecutive measurements in each of the subjects, with CVPni ranging from 2. 7 to 7 .1 cm H10, the standard error of the estimate fell between 0.1 and 0.2cm H.o. lbtientB
Fifty-two critically ill patients (30 men, 22 women, mean age 71 years, range 31 to f17 years) were studied. They required intensive care for cardiovascular surgery (20), abdominal surgery (9), respi-
372
ratory failure (8), cardiac failure (7), trauma (2), and other nonsurgical conditions (6). Twenty-seven patients were on mechanical ventilators, and 33 were comatose or sedated to a degree that excluded verbal communication. Cardiac arrhythmias were present as premature ventricular contractions in 12 patients, as atrial flutter or 6brillation in seven, and as second or third degree heart blocks in three. Either a central venous or a Swan-Ganz catheter had been inserted for clinical management via the subclavian (30 patients), the internal jugular (20 patients), or the femoral vein (2 patients) prior to requesting informed consent from the attending physician and the patient or the patient's family to perform the noninvasive CVP measurement.
Procedure Tiie patient was placed in the supine posture. Tiie phlebostatic axis and the course of the right or let\ internal jugular veins were marked on the skin with ink. An inductive plethysmograph transducer, 1.5 cm in height, containing 6ve inductive coils in a sinusoidal array, was placed around the neck. After connecting the transducer to the oscillator module and to the demodulator/ampli6er circuitry (Respitrak), the NIP analog signal along with an ele11int and its value recordt>d. The tubing was then pluc.-ed next to the neck adjusting the water level to the intersection between tht' course of the internal jugular vein and the center of the NIP trw1sdm.-er and its value recorded. The diflerem.-e between these two readings was taken as CVP in cm 11,0.
The Mann-Whitney r.mk sum test and the Wilcoxon paired rank test were applied to compare mean differences between the ..open" and ..blinded" series and between the two different examiners, respe17.7 >11.0 >21.0 >15.8 >18.4 >18.0
FIGURE 3. Identity plot of CVPni (NIP) vs CVPi (intravascular catheter). The open circles denote the "open" and the solid circles the "blinded'' series. The rectangles denote the two patients with obstruction of the right internal jugular vein who showed much lesser _ _ _ _ _ _ ___. pressure with CVPni than CVPi. The inserted table lists the values of CVPni in whom an arterial waveform (Pnip-a) was unobtainable; although exact measurements of CVP could not be recorded with the NIP 20 24 method, all values were greater than normal in keeping witil CVPi measurements.
ECG
CVPni was taken as greater than the maximal achievable vertical distance between the phlebostatic axis and the NIP transducer (Tuble in Fig 3). There was no difference between trends of CVPi vs CVPni resulting from 14of15 changes in PEEP (3 to 10 cm H20) in ten mechanically ventilated patients (Fig 5). In 13 patients, independent measurements of CVPni by two different examiners revealed a mean difference (CVPni - CVPi) relative to CVPi of 8 percent for the examiner who performed the first and 7 percent for the examiner for the second test (p =NS).
Pnlp-v
LINE OF IDENTITY
the values of CVPni and CVPi agreed closely, 11.0 and 9.5 cm H20, respectively. In seven patients in whom the NIP showed a venous waveform despite maximal elevation of the upper part of the body, INDWELLING INTERNAL JUGULAR CATHETER
AFTER REMOVAL OF INTERNAL JUGULAR CATHETER
:c E ,g,
ECG
•
2
0Cll
0
c
A.
(;
Pnlp-v
