DE19600983C1 - Non-invasive brain pressure determination system for human beings - Google Patents

Abstract

The determination system includes a unit for the measuring and recording of the continuous blood pressure and a unit for the simultaneous measuring and recording of the blood flow speed using transcranial Doppler sonography (TCD-flow speed curve). Also a processor, which using the measured TCD and blood pressure values determines characteristic values, and from these values, by means of a copying prescription, determined by multilinear regression analysis, computes the weighting function between blood pressure and brain pressure. Also with the help of the so computed weighting function, determines the brain pressure, from the blood pressure values, using methods of the system theory. Whereby as characteristic values, Fourier coefficients or coefficients of the weighting function between blood pressure and the TCD flow speed curve or coefficients of the weighting function between the TCD flow speed curve and the blood pressure, are used and the blood pressure multiplied with the Pourcelot index of the TCD flow speed curve.

Description

The present invention relates to a device for non-invasive Determination of brain pressure in humans.

Acute cerebral affections often lead to an increase in the brain pressure. Monitoring intracranial pressure can be vital to come, for example, in foudroyanter brain pressure development to control the efficiency of therapeutic measures. Quantitative Are intracranial pressure values, especially continuous intracranial pressure curves so far only available through invasive measures. This is where the Measurement of intracranial pressure (ICP) using a special Ka theters or pressure transducer passing through a hole in the skull in the brain ventricle, the epidural space or subarachnoid space is introduced. It would of course be desirable not to measure the intracranial pressure to be able to carry out invasive measures.

US-A-5 388 583 describes a method and an apparatus for noninvasive research into an intracranial medium based on the Timing of ultrasound signals is based on the brain tissue sent. A proportionality between the change in the speed of the ultrasound signal and changes against the brain pressure. Thus, using this method, Än changes in brain pressure, the absolute value of the brain pressure, which is of crucial importance for clinical application is, however, cannot be grasped with it.

US-B: Webster John G .: Encyclopedia of Medical Devices and Instru mentation, Volume 3, John Wiley & Sons, New York, 1988, pp. 1961-1983, contains a comprehensive description of clinical relevance and current practice of brain pressure measurement, taking anatomical considerations are included. The methods described reflect the current medical status. They are presented accordingly procedures for intracranial pressure measurement exclusively invasive procedures.

The US-Z: IEEE Transactions on Biomedical Engineering, Volume 42, 1995, Pp. 529-540, describes a mathematical model for the non-invasive Determination of intracranial pressure. Here, an electrical analogue to Be  writing of the intracranial dynamics, in which, as far as possible the brain pressure functions in the form of an electrical circuit be replicated. It was measured invasively in 18 patients his brain print used four Mo characterizing this model to calculate dent parameters specifically for the patient. The result of this stu accordingly, it provides patient-specific modeling, see above that not from a general, non-invasive procedure for determining brain pressure can be spoken.

The present invention is therefore based on the object to provide the direction with which the non-invasive, patient-independent pending intracranial pressure measurement, especially the measurement of continuous Brain pressure curves, is made possible.

This object is achieved by a device according to An spell 1 solved. Preferred or particularly expedient embodiment gene of the subject matter of the invention are specified in the subclaims ben.

The invention therefore relates to a device for non-invasive Determination of intracranial pressure in humans, comprising a device for Measurement and recording of continuous blood pressure, an on Direction for simultaneous measurement and recording of blood flow speed by means of transcranial Doppler sonography (TCD-Strö mation speed curve), and a processor that is used determination of the measured TCD and blood pressure values from these parameters using a multilinear regression analysis lysis determined the mapping function between the weight function Blood pressure and cerebral pressure calculated and with the help of the Ge calculated in this way Weight function from the blood pressure values using methods of the system theory determines the intracranial pressure, using Fourier coefficients as characteristic values or coefficients of weight function between blood pressure and TCD flow rate curve or coefficient of weight function between TCD flow velocity curve and blood pressure multiplied by the Pourcelot index of TCD flow rate speed curve can be used.  

According to the invention, it has surprisingly been found that the tracranial commiment can be understood as a transmission system which receives the blood pressure curve as an input signal and as a reak tion then delivers the brain pressure as an output signal. The system ver hold, that is the relationship between blood pressure curve and intracranial pressure curve, is determined by the system parameters weight and transfer function described. An essential aspect of the invention be is a processor in the device according to the invention add the characteristic values, such as the Fourier coefficients, the measured TCD flow rate curve used to measure the Create weight function.

The simula that can be carried out by means of the device according to the invention tion process yields intracranial pressure values that are conventional compared to invasive measures measured intracranial pressure values only a small Ab show softening, so that the device according to the invention or there with feasible simulation methods for non-invasive brain pressure determination are clinically practicable.

Below are the theoretical foundations of the system theory explained.

Systems theory deals with the description of signals that are in a system, the so-called transmission system, arrive and cause an outgoing signal there in response. Such are common Transmission systems, regardless of whether it is information technical or physiological systems are too complicated to to describe them purely analytically, such as in the form of differences equations. Systems theory does not examine the structure or the realization of a system, but describes the system through be ne reaction to any given input signals. The plays Transfer function, which is the relationship between Input signal and output signal describes a central role. The System behavior is due to the relationship between the input signal  and fully characterized the output signal. For the purposes of The present invention uses the intracranial compartment as an over system with the input signal blood pressure and the output signal brain pressure understood.

The theory used for such a model applies to a transmission system stem with the following characteristics:

• 1. There is an input signal and a system response Output signal. With y (t) = T {x (t)} let the system reaction to the incoming designated signal x (t), where the parameter t always stands for time. x (t) → system → y (t) = T {x (t)}
• 2. Linearity
  For a signal x (t), which is composed of the signals x₁ (t) and x₂ (t) in the form
  x (t) = a _* x₁ (t) + b _* x₂ (t), a, b ∈ IP (real numbers), should apply
  T {x (t)} = a _* T {x₁ (t)} + b _* T {x₂ (t)}
• 3. Time invariance
  T {x (t-t₀)} = y (t-t₀) that means the system response is independent of the time of an incoming signal. A signal that arrives in the system postponed by the time t₀ generates the same system response as the original signal, but also postponed by the time t₀.
• 4. Stability
  From | (t) | <M <∞ for all t follows | y (t) | <N <∞; M, N, ∈ IN (natural numbers). If the signal x (t) is restricted for all tents, then y (t) is also restricted.
• 5. Causality
  From x (t) = O for all t <t₀ it follows that y (t) = 0 for all t <t₀. A reaction in the form of the output signal only takes place when its cause x (t) has arrived.

In a transmission system with the properties described above 1-4 can be the relationship between the input signal x (t) and the output signal y (t) generally by convolution with a function f (t) in the form

express.

If one additionally assumes the causality of the system, then for τ <0 f (τ) = 0, because otherwise y (t) of signal values of x that only after the Time t to reach the system (t-τ <t) would be influenced. With that becomes the equation (1) the equation

The function f is called the weight function. With the weight function it is System behavior fully described. As a transfer or trans The Laplace transform becomes the function

designated.

The Laplace transforms of y (t) and x (t) are denoted by Y (s) and X (s). If all Laplace transforms exist, i.e. the corresponding integrals converge, then Y (s) = F (s) _* X (s). The Laplace transform of the output signal y (t) is thus the product of the Laplace transform of the weight function f (t) and the input signal x (t). Since the transfer function can be clearly transformed back into the weight function, the transfer function and weight function are equivalent descriptions of the system.

In practice, the signals become a sequence of discrete samples points described. Equation (1) provides the in discrete notation approach

that is, the kth value of the output signal _k is determined by a number n of previous input values x _k , x _k-1 , x _k-2,. . .x _{k-n + 1} . The f _j are called coefficients of the (discrete) weight function f. For the best approximation of the calculated values to the actual signal values over a period of time
the expression takes from p measurements

a minimum of. Substituting equation (2) into equation (3), we get one

If G is minimal, the partial derivatives must

be.

This results in the system of equations

or

The system of equations (4 ') can be uniquely solved (that is, the relationship between the input and output signals expressed by the coefficients f _j can be calculated) if the (n × n) matrix

is regular.

The following are used in the device according to the invention facilities described.  

The continuous blood pressure is measured and recorded using standard equipment. Because of the greater accuracy a radial or arterial artery is particularly suitable implanted pressure transducer (e.g. Gould Statham 23 ID). Of the Blood pressure recorded by means of such a device becomes after amplification continuously and simultaneously with the measured blood flow speed values transferred to a microprocessor.

Also suitable for continuous non-invasive measurement commercially available blood pressure monitor, which on the finger pressure measurement based (e.g. Ohmeda 2300 / Finapress).

For the simultaneous measurement of the blood flow rate using Transcranial Doppler sonography is suitable for commercially available pulsed Doppler devices. For example, by using a pulsed Doppler device at a transmission frequency of suitably 2 MHz circumscribed vascular sections in defined steps of for example, 5 mm can be examined. To create the TCD For example, the flow rate curve can be the cerebral artery media can be used. The commercially available Doppler devices sometimes even simultaneous bilateral derivation.

For recording and evaluating the TCD flow rate Speed curves, blood pressure and cerebral pressure curves comprise the fiction device a microprocessor, suitably with built-in A / D converter card for data acquisition and implementation of the simula tion algorithm. This recording and evaluation can, for example wise by means of two personal computers (for example of the type PC / i486), where one can be a portable device, each with one A / D converter card are equipped, for example of type DAP 2400 (available from Microstar Laboratories, USA). The necessary mathematical and statistical calculations can be carried out using an im Commercially available software tool (e.g. Real Time Graphics & Measurement Tools / Quinn Curtis).  

The special make of the devices mentioned above is irrelevant as long these devices or devices have the functionalities described above exhibit.

As described above, the non-invasive brain pressure determination the intracranial commparti by means of the device according to the invention ment as a transmission system, the blood as the input signal receives pressure curve and in response to it as the output signal Brain pressure supplies. From the TCD flow velocity curves as well The continuous blood pressure curves are the appropriate ones TCD parameters and the weight function between blood pressure and TCD Curve calculated. A statistical evaluation then takes place for egg First a linear approach. Investigations of the correlations between different TCD characteristics and the coefficients of the Ge weight function. In this way, the relationship between the TCD parameters and the weight function as patient-independent Illustration instructions are formulated.

The invention is based on the drawings and the following be written example explained in more detail. The drawings show:

Figure 1 is a schematic representation in the form of a flow chart for non-invasive brain pressure determination by means of the device according to the invention.

FIG. 2a, the conventional invasive intracranial pressure curve measured (ICP) of patients with traumatic brain injury (Patient A);

FIG. 2b shows the specific means of the inventive device intracranial pressure curve (simulated intracranial pressure curve) (sim ICP.) Of the patient A;

. Fig. 2c, the difference between the invasively measured intracranial pressure curve in accordance with Figure 2a and the invention measured intracranial pressure curve in Fig. 2b;

3a according to the invention measured TCD Strömungsgeschwindig keitskurve (Doppler) of another patient with a traumatic brain injury (patient B).

FIG. 3b according to the invention measured blood pressure curve (RR) of the Pa tienten B;

Fig. 3c the conventionally invasively measured intracranial pressure curve of the patient B;

Fig. 3d according to the invention certain brain pressure curve of the patient B;

Fig. 4a, 4b, 4c and 4d to Figs. 3a, 3b, 3c and 3d analog Kur ven, but the measurement over a period of 320 seconds he followed.

As is apparent from Fig. 1, the non-invasive brain pressure determination by means of the device according to the invention by deriving the blood pressure curve (RR curve) and the TCD blood flow rate curve (TCD curve), analysis of the TCD curve at fixed time intervals, calculation of the weight function the TCD parameters and conversion of the blood pressure curve into the intracranial pressure curve (ICP curve) using the calculated weight function.

example

From the data of intensive care patients with transkra derived simultaneously niello-Doppler sonographic (TCD) flow velocity curves, as well as continuous blood pressure and invasively measured brain pressure curves were the TCD parameters and the weight function between blood and brain pressure calculated.

The intracranial pressure was measured for the invasive brain pressure measurement neither through a ventricular drainage inserted into the side ventricle and connected pressure transducers as ventricular pressure or with Miniature pressure transducers recorded as epidural pressure. The last ge Miniature pressure transducers are specially designed for long-term measurements constructed, electromechanical pressure transducers with miniaturized strain gauges work and are implanted epidurally. Sol che pressure transducers have a zero point drift of less than 2 mm Hg / 24 h.  

A statistical evaluation then followed for an initially linear one Approach studies of the correlation between different TCD Characteristic values and the coefficients of the weight function. So that was the relationship between the TCD parameters and the weight formulated as a patient-independent mapping rule.

The procedure for non-invasive brain pressure determination was carried out using following steps:

Step 1 Calculation of the weight function for a given blood pressure and brain pressure curve

For deriving a relationship between characteristic values of the TCD curve and the transfer function between the blood pressure (RR) - and the brain pressure curve (ICP) is the blood pressure as an input signal and the brain pressure is regarded as the output signal of the hemodynamic system. How described in the general case (theoretical basis, weight and transfer function) leads the approach

to the system of equations

As can be seen from equation (2 ′), the value n describes which period of the blood pressure influences the current intracranial pressure value (or is correlated with the current intracranial pressure). In the simulation process, n = 25 is used. The step size (the time interval between the k-1 tem and the k-th measured value) was selected as a function of the length of the cardiac cycle, so that the 25 measured values RR _k , RR _k-1 RR _{k-n + 1} equidistant over 3 cardiac cycles are distributed. This means that in the Modellvor position the current ICP value is influenced by the blood pressure over a period of 3 cardiac cycles. From equation (3) it can be seen that the value p determines the signal period over which the ICP curve calculated from the blood pressure curve corresponds as closely as possible with the actual (measured) ICP curve. p is chosen so that the comparison period is 14 cardiac cycles (that is, the time interval between RR _k and RR _{k + p} is 14 cardiac cycles). It follows

p = (14 cardiac cycles / 3 cardiac cycles) _* 25 = 116.

The parameters f _{j of} the weight function are calculated from equation (4 ′ ′) using a software tool using mathematical standard methods. The (n × n) matrix

has always proven to be regular. The solution of the system of equations provides the (discrete) weight function f = (f₀, f₁,... F _n-1 ).

2nd step Calculation of the TCD parameters

When determining brain pressure using the device according to the invention at least the three types of Pa parameters can be used as TCD parameters.

In the measurements shown in FIGS . 2 to 4, the simulation method was carried out with the characteristic value type described below under b).

a) Fourier coefficients

In general, every unique periodic function f (x) = f (x + kT₀), the piecewise monotonous and continuous, is clearly represented as a Fourier series adjustable. Such a Fourier decomposition transforms the function into its different frequency ranges "broken down".

When calculating the Fourier coefficients of the TCD curve, the duration of K cardiac cycles (K = 1, 2, 3) is assumed as the period T₀. The first m Fourier coefficients (DOP₀, DOP₁,.. DOP _m-1 ) are calculated using standard thematic software (Real Time Graphics & Measurement Tools / Quinn Curtis). With TCDN as the first and TCD _{N + P} as the last TCD sample point of the period T₀ is approximate

Φ _j indicates the phase shift of the different frequency components to each other. The jth summand is a sine function with j times the frequency of the decomposed oscillation (TCD curve), the coefficient DOP _j is a measure of the size of this frequency component of the TCD curve. DOP₀ corresponds to the DTMF (mean value of the TCD curve). DOP₁ is a measure of the pulse sarity of the curve. In previous studies, the number of Fourier coefficients evaluated was between 3 and 6 (m = 3-6).

b) The coefficients of the weight function from blood pressure to TCD Curve

Analogous to the calculation of the weight function between blood pressure and brain pressure curve explained in step 1, the weight function between blood pressure and TCD curve is calculated here. The TCD parameters are the coefficients of this weight function and are referred to as (a) with (DOP₀,... DOP₁ DOP _m-1 ). It has proven to be practically advantageous to work with weight functions with only 6 coefficients (in contrast to the weight function between blood pressure and brain pressure in step 1, which has 25 coefficients), that is called m = 6. The TCD curve is also sufficient with this described in detail and the complexity of the statistical analysis (step 3) is reduced.

c) The coefficients of the weight function from TCD to blood pressure Curve multiplied by the Pourcelot index of the TCD curve

Here the weight function is calculated, which converts the TCD curve into the blood pressure curve, in contrast to b), where the weight function, which (in the opposite direction) maps the blood pressure curve into the TCD curve, is used. The Pourcelot index R of the TCD curve, also called the index of ze rebrovascular resistance, is defined by R = (systolic flow velocity - diastolic flow velocity) / systolic flow velocity. The coefficients of the weight function, each multiplied by the Pourcelot index form the TCD parameters (DOP₀, DOP₁, ... DOP _m-1 ). As in b), m = 6 is chosen here.

3rd step Statistical evaluation of patient data for determination the relationship between the transfer function and the TCD Characteristic values

Starting point: At least one number vector (f₀, f₁,..., F _n-1 ) was described for each patient, which describes the transfer function between RR and ICP and simultaneously a number vector (DOP₀, DOP₁,..., DOP _m-1 ) of TCD characteristics that describe the TCD curve. In the presented method, n = 25 and m = 6. In the first approach, these two vectors are checked for a linear relationship. That is, it was checked whether there is an (n × m) matrix A and an n-dimensional vector B, so that regardless of the patient (₀, ₁,..., _N-1 ) = A _* (DOP₀, DOP₁, DOP _m-1 ) + B represents a "sufficient" approximation for (f₀, f₁,..., F _n-1 ). This can be reformulated in (1) ₁≈ _j = A _{j * *} (DOP₀, DOP₁, ..., DOP _m-1 ) × B _j , j = 0.1,. . ., n-1, A _{j *} is the jth line of A, Bj the jth coefficient of B. or explicitly written out

For m = 1 (only 1 TCD characteristic value) and fixed j, this would be the question of a straight line g (x) = m _* x + b, so that _j = m _* DOP₀ + b (m = A _jo , b = B _j ) the best linear approximation for the value pairs (f). DOP) across all patients. The method used here is called linear regression. A generalization of this question appears in equation (1 ′ ′). We are looking for a straight line in several variables g (x₁, x₂, ... x _m-1 ) = m₁ _* x₁ + m₂ _* x₂ +. . . m _{m-1 *} x _m-1 + b, so that) = m₁ _* DOP₀ + m₂ _* DOP₁ +. . .m _{m-1 *} DOP _m-1 + b represents the best linear approximation for the vectors (f _j , DOP₀, DOP₁,..., DOP _m-1 ) across all patients. The solution method is a generalized linear regression and is called multilinear regression. For each coefficient f _{j of} the transfer function, a multilinear regression is carried out over all patient data, that is a total of n multilinear regressions, so that

is minimal. This is done using standard mathematical software (Real Time Graphics & Measurement Tools / Quinn Curtis) and leads to the searched (n × m) matrix A and the n-dimensional vector B.

Result

To validate the procedure, data from 12 patients were used (age 46 ± 16 years) using multilinear regression analysis of the Zu correlation between TCD parameters and the weight function of Blood pressure after intracranial pressure to the individually independent illustration writing formulated. When reviewing the procedure on 5 new patients With 15 ICP leads there was a mean deviation of the simu calculated ICP (averaged over 5 seconds) from the measured values 4.27 ± 1.69 mm Hg. Also rapid ICP fluctuations, e.g. B. during a drug ICP reduction, could be recorded well. Between the ge measured and simulated ICP values there was a significant, linear Correlation (r = 0.79, p <0.001). These findings and especially the ge Rings deviation of 4.27 mm Hg show that by means of the Invention simulation method performed according to the device clinically practical tikabel is.

In Figs. 2 to 4, the measured and simulated brain pressure curves are exemplary juxtaposed in two patients with craniocerebral trauma at the age of 63 years (Patient A) and 46 years (patient B). Fig. 2 herein shows the comparison between measured and simulated brain pressure in patient A over a period of 8 seconds. In FIGS. 3 and 4 for the patient B are in addition the TCD (Doppler) - to see and the blood pressure curve (RR). In Fig. 3, the curve over 10 seconds, and in Fig. 4 is the curve shown over 320 seconds.

Claims (3)

1. Device for non-invasive brain pressure determination in humans, comprising

• a device for measuring and recording the continuous blood pressure,
• - A device for the simultaneous measurement and recording of the blood flow rate by means of transcranial Doppler sonography (TCD flow rate curve), and
• - A processor which determines the characteristic values using the measured TCD and blood pressure values and calculates the weight function between blood pressure and cerebral pressure from these characteristic values by means of a mapping rule determined by multilinear regression analysis and, using the weight function thus calculated, the cerebral pressure from the blood pressure values by means of systems theory methods determined, Fourier coefficients or coefficients of the weight function between blood pressure and TCD flow rate curve or coefficients of the weight function between TCD flow rate curve and blood pressure multiplied by the Pourcelot index of the TCD flow rate curve are used as characteristic values.

2. Device according to claim 1, wherein the means for measurement blood pressure into the radial artery or femoral artery implan tated pressure transducer. 3. Device according to claim 1 or 2, wherein the means for measuring solution of the TCD flow rate curve a pulsed Doppler gene advises.