Blood Pressure Measurement System for Filtering Low-Frequency, High-Amplitude Noise

This invention concerns a blood pressure measurement system. More particularly, the invention concerns a blood pressure measurement system that filters low-frequency, high-amplitude noise to accurately measure blood pressure.

Blood pressure measurement is used to generally indicate the health of a patient or the type of emergency treatment which is necessary for an accident victim or a battlefield casualty. Most blood pressure measurement systems involve listening for blood pulses of a patient with a stethoscope. Noises of an emergency vehicle or in a battlefield make these pulses difficult to hear, which makes accurate measurement of blood pressure very difficult to obtain.

FIG. 1 shows a schematic diagram of a blood pressure measurement system that determines systolic, diastolic, and mean arterial pressures without detecting blood pulses. The blood pressure measurement system of FIG. 1 is particularly useful in emergency vehicles and battlefield situations, where movement and noise drastically affect accurate measurement of blood pressure in systems that rely on the detection of blood pulses. The system of FIG. 1 filters out blood pulses and high frequency noise, and processes signals having frequencies on the order of 1 Hz.

The operation of the blood pressure measurement system of FIG. 1 is detailed in U.S. Pat. No. 4,649,928 to Samaras et al., which has been assigned to the same entity as this invention. This application incorporates the disclosure of U.S. Pat. No. 4,649,928 by reference.

FIG. 1 shows an occlusion bladder 10 on the upper arm 8 of a patient, which is inflated to a pressure substantially higher than systolic blood pressure of the patient to stop blood flow in the upper limb 8. A lower sensing bladder 12 on the forearm 6 of the patient is inflated to a nominal value, such as 70 mmHg. Pressure in the occlusion bladder 10 is gradually decreased over a period of approximately thirty seconds. A microprocessor 28 controls valves 16, 18, and 20 and controls pump 14 to inflate and deflate the occlusion bladder 10 and the sensing bladder 12. Transducers 22 and 24 monitor pressures in the occlusion bladder 10 and the sensing bladder 12, respectively. These transducers 22 and 24 respectively produce an occlusion pressure signal and a sensing pressure signal which are received by a microprocessor 28. The microprocessor 28 filters out pulses corresponding to a patient's heart beat from the sensing pressure signal produced by the transducer 24.

An occlusion pressure value sensed by the transducer 22 on the occlusion bladder 10 corresponds to systolic pressure of a patient when the sensing pressure signal from the sensing bladder 12 reaches a minimum value. The microprocessor 28 derives first, second, and third time derivatives of the sensing pressure signal from the sensing bladder 12. An occlusion pressure value from the occlusion bladder 10 corresponds to mean arterial pressure when the third time derivative of the sensing pressure signal is a positive-going, zero-crossing signal. The microprocessor 28 derives diastolic pressure from the mean arterial pressure and the systolic pressure according to a well known relationship.

The present inventors have found that the system of FIG. 1 is somewhat responsive to very low-frequency, high-amplitude noise signals, which can cause the system of FIG. 1 to produce false values of systolic, mean arterial, and diastolic pressures. These noise signals are produced when a patient wearing the system is slowly transported by an emergency vehicle over a large bump, for instance. These noise signals are also produced when a patient wearing the system makes slow, large movements with his fingers, for instance. In an emergency situation, it is critical to accurately measure these pressures for correct treatment of a patient. Thus, a need exists for determining the presence of such slow, large noise signals and for filtering out these noise signals to accurately 10 measure blood pressure of a patient.

The invention concerns a blood pressure measurement system comprising an occlusion bladder attachable to a patient's upper limb, a sensing bladder attachable to a patient's lower limb, and a means for monitoring pressures in the occlusion bladder and sensing bladder. The system also comprises a means for filtering out blood pulse signals from the sensing pressure signal to produce a static pressure signal, a means for recognizing and storing a plurality of relative minimums of the static pressure signal, and a means for relating the smallest relative minimum with a corresponding occlusion pressure to derive systolic pressure of the patient.

The blood pressure measurement system of this invention is immune to high frequency noise and does not involve listening for blood pulses of a patient. The system is also immune to slow, large noise signals.

It is an object of this invention to recognize slow, large noise signals to which a blood pressure measurement system responds.

It is another object of this invention to filter out these slow, large noise signals from a blood pressure measurement system for accurate measurement of blood pressures.

FIG. 1 is a schematic diagram of a blood pressure measurement system, comprising the environment of the present invention.

FIGS. 2(a)-(f) illustrate waveforms related to the system of FIG. 1 when responding to slow, large noise signals before sensing systolic pressure, as recognized according to this invention.

FIGS. 3(a)-(f) illustrate waveforms related to the system of FIG. 1 when responding to slow, large noise signals after sensing systolic pressure and before sensing mean arterial pressure, as recognized according to this invention.

FIGS. 4(a)-(f) illustrate waveforms related to the system of FIG. 1 when responding to slow, large noise signals when sensing mean arterial pressure, as recognized according to this invention.

FIGS. 5a and 5b show a flow chart illustrating steps for determining and filtering slow, large noise signals out of the system of FIG. 1 according to this invention.

The present inventors have identified waveforms related to the system of FIG. 1, when that system is responsive to slow, large noise signals that occur before systolic pressure of a patient is reached. FIGS. 2(a)-(f) illustrate these waveforms. Pressure in mmHg is plotted on the vertical axis and time in seconds is plotted on the horizontal axis in FIGS. 2(a), 2(b), and 2(f). FIGS. 2(c), 2(d), and 2(e) respectively illustrate first, second, and third time derivatives of the curve of FIG. 2(b). Negative and positive values are plotted on the vertical axis and time is plotted on the horizontal axis in FIGS. 2(c), 2(d), and 2(e).

FIG. 2(a) shows a sensing pressure signal produced by the transducer 24 on the sense bladder 12. As an occlusion bladder 10 is slowly deflated, blood fills the limb of a patient and pressure in the limb changes according to the curve of FIG. 2(a). The present inventors have recognized that a rise 40 occurs in the curve when the blood pulse measurement system of FIG. 1 is responsive to a slow, large noise signal that occurs before systolic pressure of a patient has been reached. Such a slow, large noise signal is on the order of 0.5 Hz and results when a patient makes large and slow finger movements, for instance. These noise signals result if a patient continually and gradually clenches his relaxed hand once or spreads the fingers of his relaxed hand once during 10-30% of the time necessary to bleed pressure from the occlusion bladder 10, for example. Somewhat smaller finger movements during this time will also cause the blood pressure measurement system of FIG. 1 to produce a rise 40 in response to slow, large noise signals. Slow, large noise signals also result when a patient wearing the system of FIG. 1 is slowly transported in a vehicle over a bump, for instance. These noise signals result if a patient is transported over a speed-bump at 10-15 mph, for example.

The curve of FIG. 2(a) would follow the dashed portion of the curve and the rise 40 would not be present in the curve, if no slow, large noise signals are present. The transducer 24 senses blood pulses as a limb fills with blood, which causes oscillations in the curve of FIG. 2(a).

FIG. 2(b) shows a static pressure curve. Though pressure changes in the limb of a patient, and the curve of FIG. 2(b) changes accordingly, the curve of FIG. 2(b) is static in that this curve shows no oscillations. This static pressure results after blood pulses are filtered in a manner as described in U.S. Pat. No. 4,649,928 from the sensing pressure signal of FIG. 2(a). FIG. 2(b) shows two relative minimums 41 and 42, which result when the system of FIG. 1 is responsive to a single slow, large noise signal 40. A single noise signal is shown as an example, though numerous slow, large noise signals might occur during measurement of a patient's blood pressure.

FIG. 2(c) shows two positive-going, zero-crosses, 43 and 45, corresponding to the relative minimums 41 and 42 in the static pressure curve of FIG. 2(b). According to the present invention, the positive-going, zero-crosses 43 and 45 are recognized when negative values of the first derivative of FIG. 2(c) change to positive values. For each of these positive-going, zero-crosses 43 and 45, a corresponding value of static pressure and occlusion pressure is determined and stored.

As recognized by the inventors, the value of occlusion pressure corresponding to the smallest value of static pressure equals systolic pressure. FIG. 2(f) represents an occlusion pressure signal from the transducer 22 on the occlusion cuff 10. FIG. 2(f) shows an occlusion pressure value 47 corresponding in time to the zero-cross 43 of the first derivative of FIG. 2(c) and an occlusion pressure value 48 corresponding to the zero-cross 45 of the first derivative of FIG. 2(c). The static pressure value 42 is smaller than the value 41. The occlusion pressure value 48 corresponds to the smaller static pressure value 42 and equals systolic pressure.

Systolic blood pressure is the highest level of blood pressure occurring in a patient and is measured first, as pressure in the occlusion bladder 10 is decreased. After systolic pressure has been determined, a region 54 of the static pressure curve of FIG. 2(b) is defined to determine mean arterial pressure. This region 54 of the static pressure curve begins when systolic pressure occurs at 42 and ends with a maximum value 53 occurring in the static pressure curve after systolic pressure 42 has occurred.

The third derivative of FIG. 2(e) illustrates two positive-going, zero-crosses 49 and 50. The zero-cross 49 is caused by the slow, large noise signal. Zero-cross 49 is a false indication of mean arterial pressure and is identified as such because the zero-cross 49 occurs before systolic pressure 48. The zero-cross 50 occurs within region 54. The present inventors have recognized that the occlusion pressure value 52 corresponding to the zero-cross 50 equals mean arterial pressure.

Diastolic pressure is derived based on the values for systolic pressure and mean arterial pressure according to a known relationship: diastolic equals 1/2(3.multidot.mean arterial 10 pressure-systolic pressure).

The present inventors have identified waveforms related to the system of FIG. 1 when that system is responsive to slow, large noise signals that occur after systolic pressure of a patient is reached. FIGS 3(a)-(f) illustrate these waveforms. Pressure in mmHg is plotted on the vertical axis and time is plotted in seconds on the horizontal axis in FIGS. 3(a), 3(b), and 3(c). FIGS. 3(c), 3(d), and 3(e) respectively illustrate first, second, and third time derivatives of the curve of FIG. 3(b). Negative and positive values are plotted on the vertical axis and time is plotted on the horizontal axis in FIGS. 3(c), 3(d), and 3(e).

FIG. 3(a) shows a pressure sensing signal produced by the transducer 24 on the sensing bladder 12. The present inventors have recognized that a rise 55 occurs in the curve when the system of FIG. 1 is responsive to a slow, large noise signal that occurs after systolic pressure of a patient has been reached and before mean arterial pressure has been reached. The curve of FIG. 3(a) would follow the dashed portion of the curve and the rise 55 would not be present in the curve, if no slow, large noise signals are present. The transducer 24 senses blood pulses as a limb fills with blood, which causes oscillations in the curve of FIG. 3(a).

FIG. 3(b) illustrates a static pressure curve after filtering of blood pulses in a manner as described in U.S. Pat. No. 4,649,928. The static pressure curve of FIG. 3(b) two relative minimums 42 and 56, which result when the system of FIG. 1 is responsive to a single slow, large noise signal 55.

FIG. 3(c) shows two positive-going, zero-crosses 45 and 57 corresponding to the relative minimums 42 and 56 of the static pressure curve of FIG. 3(b). According to the present invention, the positive-going, zero-crosses 45 and 57 are recognized when negative values of the first derivative of FIG. 3(c) change to positive values. For each of these positive-going, zero-crosses 45 and 57, a corresponding value of static pressure is determined and stored. The relative minimums 42 and 56 of FIG. 3(b) respectively correspond to the zero-crosses 45 and 57 of FIG. 3(c).

As recognized by the inventors, the value of occlusion pressure corresponding to the smallest value of static pressure equals systolic pressure. FIG. 3(f) represents an occlusion pressure signal from the transducer 22 on the occlusion cuff 10. FIG. 3(f) shows an occlusion pressure value 48 corresponding in time to the zero-cross 45 of the first derivative of FIG. 3(c)and an occlusion pressure value 46 corresponding to the zero-cross 57 of the first derivative of FIG. 3(c). In this case, the static pressure value 42 is less than the value 56. The occlusion pressure value 48 corresponds to the smaller static pressure value 42 and equals systolic pressure.

After systolic pressure has been determined, a region 54 of the static pressure curve of FIG. 3(b) is defined to determine mean arterial pressure. This region 54 of the static pressure curve begins when systolic pressure occurs at 42 and ends with a maximum value 53 occurring in the static pressure curve after systolic pressure 42 has occurred.

The third derivative of FIG. 3(e) shows two positive-going, zero-crosses 58 and 59. The zero-cross 58 is caused by the slow, large noise signal. Both zero-crosses 58 and 59 occur within region 54.

According to the invention, negative values 60 and 61 of the second derivative of FIG. 3(d) are stored in a temporary register, which correspond to positive-going, zero-crossings 58 and 59 of the third derivative of FIG. 3(e). The present inventors have discovered that the most negative value of the second derivative stored in the temporary register, corresponds to an occlusion pressure value that equals mean arterial pressure. In this example, the more negative value of the second derivative is at 61. The occlusion pressure value 52, corresponding to the more negative value 61 of the second derivative, equals mean arterial pressure. Diastolic pressure is derived based on the values for systolic pressure and mean arterial pressure.

The present inventors have identified waveforms related to the system of FIG. 1 when that system is responsive to slow, large noise signals at the same time mean arterial pressure of a patient is reached. FIGS 4(a)-(f) illustrate these waveforms. Pressure in mmHg is plotted on the vertical axis and time in seconds is plotted on the horizontal axis in FIGS. 4(a), 4(b), and 4(f). FIGS. 4(c), 4(d), and 4(e) respectively illustrate first, second, and third time derivatives of the curve of FIG. 4(b). Negative and positive values are plotted on the vertical axis and time is plotted on the horizontal axis in FIGS. 4(c), 4(d), and 4(e).

FIG. 4(a) shows the sensing pressure signal from the transducer 24 of the sense bladder 12. A rise 64 occurs in the curve in response to a single slow, large noise signal. FIG. 4(b) shows a waveform illustrating static pressure sensed by the transducer 24 on the sensing bladder 12 after blood pulses have been filtered in a manner as described in U.S. Pat. No. 4,649,928 from the curve of FIG. 4(a). The static pressure signal of FIG. 4(b) has two relative minimums, 42 and 65, which result when the system of FIG. 1 is responsive to a single, slow, large noise signal 64 occurs at the same time mean arterial pressure is reached. FIG. 4(c) illustrates one positive-going, zero-cross 45 corresponding to a minimum static pressure value 42 and an occlusion pressure value 48. The occlusion pressure value 48 equals systolic pressure, as discussed concerning FIGS. 2(b), 2(c), and 2(f), for instance.

FIG. 4(e) shows the third derivative of the static pressure curve of FIG. 4(b). FIG. 4(e) illustrates two positive-going, zero-crosses 66 and 67 which result when the slow, large noise signal 64 occurs. Negative values 68 and 69 of the second derivative of 4(d) respectively correspond to these positive-going, zero-crosses 66 and 67. These negative values 68 and 69 of the second derivative of FIG. 2(d) can be equal when the slow, large noise signal 64 occurs at the same time mean arterial pressure is reached. In this situation, two values of occlusion pressure 70 and 71, corresponding to the two equal negative values 68 and 69 of the second derivative, are averaged to yield an occlusion pressure value 52. According to this invention, occlusion pressure value 52 equals mean arterial pressure. Diastolic pressure is derived based on the values for systolic pressure and mean arterial pressure as discussed above.

FIG. 5(a) shows a flow chart illustrating steps performed by circuitry in the microprocessor 28 of FIG. 1 according to this invention. The microprocessor 28 regulates the pump 14 and valve 16, 18, and 20 to inflate the sense bladder 12 and occlusion bladder 10 at steps 75 and 76, respectively. At step 77, pressure in the occlusion bladder 10 is bled off slowly at 6 mmHg per second, for instance. As pressure in the occlusion bladder 10 reduces, blood begins to fill the limb of the patient. At step 78, the microprocessor 28 monitors the output of the transducer 22 and transducer 24 and records the resulting sensing pressure signal and the occlusion pressure signal. At step 79, the microprocessor 28 filters blood pulses from the output of the sensing pressure signal in a manner as described in U.S. Pat. No. 4,649,928 to produce a static pressure signal corresponding to FIG. 2(b), for instance. At step 80, the microprocessor 28 takes first, second, and third derivatives of the static pressure curve. At step 81, the microprocessor 28 searches for positive-going, zero-crosses in the first derivative of the static pressure curve, such as 43 and 45 in FIG. 2(b), for example.

At step 82, the microprocessor 28 stores in a temporary register a static pressure value for each corresponding zero-cross of the first derivative. FIG. 2(c) shows two positive-going, zero crosses 43 and 45. Thus, the microprocessor 28 stores corresponding static pressure values 41 and 42 of the curve of FIG. 2(b). At step 83, the microprocessor 28 determines if pressure in the occlusion bladder 10 has dropped to 20 mmHg, for instance. If the pressure in the occlusion bladder 10 has not yet dropped to 20 mmHg, the monitoring and recording of pressures in the occlusion bladder 10 and sense bladder 12 continues, from step 78 and pressure in the occlusion bladder 10 continues to decrease. If the pressure in the occlusion bladder 10 has dropped to 20 mmHg, the microprocessor 28 has accumulated enough information to derive systolic, mean arterial and diastolic pressures.

At step 84, the microprocessor 28 determines if any values representing static pressure values, such as 41 or 42, in FIG. 2(b) have been stored in the temporary register. If no values have been stored in the register, the initial pressure of the occlusion bladder 10 might have been lower than the systolic pressure of the patient. Thus, at step 85, the microprocessor 28 controls the pump 14, valves 16, 18, and 20 to reinflate the occlusion bladder 10 to a pressure which is higher than 70 mmHg above a typical systolic pressure and steps 77-84 repeat. If static pressure values have been stored in the register, the microprocessor 28 finds the smallest of these values at step 86.

At step 87, the microprocessor 28 relates this smallest static pressure value to an occlusion pressure value to derive systolic pressure. FIG. 2(b) shows the smallest stored instantaneous value of static pressure value at 42. FIG. 2(f) shows the corresponding occlusion pressure at 48. The occlusion pressure value 48 is equal to systolic pressure.

FIG. 5(b) shows a continuation of a flow chart of FIG. 5(a) and illustrate more steps performed by circuitry in the microprocessor 28 according to this invention. At step 88, a microprocessor 28 defines a region 54 of the static pressure curve of FIG. 2(b). Mean arterial pressure is found within this region 54. The region 54 begins with systolic pressure 42 of the static pressure curve of FIG. 2(b) and extends to a maximum point 53 of the static pressure curve.

At step 89, the microprocessor 28 examines the third derivative of FIG. 2(e) for positive-going, zero-crosses, for instance. The microprocessor 28 then stores negative values of the second derivative of FIG. 2(d), which correspond to these zero-crosses, in another temporary register. For example, the third derivative of FIG. 2(e) shows two positive-going, zero-crosses 49 and 50. A value 44 of the second derivative of FIG. 2(d), corresponding to the zero-cross 49, is not stored in the register, because the zero-cross 49 does not fall within the defined region 54. A value 51 of the second derivative of FIG. 2(d) corresponding to the zero-cross 50 is stored in the temporary register, because the zero-cross 50 of the third derivative of FIG. 2(e) falls within the defined region 54.

At step 90, the microprocessor 28 examines the temporary register to see if any negative values of the second derivative are present. If not, at step 91, the microprocessor reinflates the occlusion bladder to 40 mmHg above the systolic pressure which has been determined at step 87. The occlusion bladder pressure is reduced at a slower rate equaling 4 mmHg per second, approximately, and the steps 77-89 of FIGS. 5(a) and 5(b) repeat. If values are present in the temporary register according to step 90, the most negative value of the second derivative in the temporary register is found by the microprocessor 28 at step 92. An occlusion pressure value that corresponds to this most negative value of the second derivative in the register is equal to mean arterial pressure. For instance, FIG. 2(d) shows the most negative value of the second derivative at 51. The corresponding occlusion pressure value is 52 in FIG. 2(f), which is the mean arterial pressure of the patient.

Steps 75-80 of FIG. 5(a) are the same as described above when filtering out slow, large noise signals that occur after systolic pressure and before mean arterial pressure.

At step 81, the microprocessor 28 searches for positive-going, zero-crosses in the first derivative of FIG. 3(c). At step 82, the microprocessor 28 stores in a temporary register a static pressure value for each corresponding zero-cross of the first derivative. FIG. 3(c) shows two positive-going, zero-crosses 45 and 57. Thus, the microprocessor 28 stores corresponding static pressure values 42 and 56 of FIG. 3(b). At step 83, the microprocessor 28 determines if the pressure in the occlusion bladder 10 has dropped to 20 mmHg, and the microprocessor 28 has accumulated enough information to derive systolic, mean arterial and diastolic pressures, as described above.

At step 84, the microprocessor 28 determines if any static pressure values, such as 42 or 56 in FIG. 3(b), have been stored in the temporary register. If no values have been stored in the register, possibly the initial pressure of the occlusion bladder 10 was lower than the systolic pressure of the patient, and steps 77-84 repeat. If values of static pressure have been stored in the register, the microprocessor 28 finds the smallest of these values at step 86. At step 87, the microprocessor 28 relates the smallest value of static pressure to an occlusion pressure value. FIG. 3(b) shows the smallest stored static pressure value at 42. FIG. 3(f) shows the corresponding occlusion pressure value at 48. The value of the occlusion pressure 48 is equal to systolic pressure.

At step 88 of FIG. 5(b), the microprocessor 28 defines a region 54 of the static pressure curve of FIG. 3(b) in which mean arterial pressure is found. At step 89, the microprocessor 28 examines the third derivative curve of FIG. 3(e) for positive-going, zero-crosses, for instance. The microprocessor 28 then stores negative values of the second derivative of the static pressure curve, which correspond to these zero-crosses, in another temporary register. For example, the third derivative of FIG. 3(e) shows two positive-going, zero-crosses 58 and 59. Values 60 and 61 of the second derivative of FIG. 3(d) correspond to the zero-crosses 58 and 59. These values 60 and 61 of the second derivative of FIG. 3(d) are stored in the temporary register.

At step 90, the microprocessor 28 examines the temporary register to see if any negative values of the second derivative are present. If not, at step 93, the microprocessor 28 repeats steps illustrated in FIGS. 5(a) nd 5(b) from step 77. If values are present in the temporary register, the most negative value in the temporary register is found by the microprocessor 28 at step 91. An occlusion pressure value that corresponds to this most negative value in the register is equal to mean arterial pressure. For instance, FIG. 3(d) shows the most negative value at 61. The corresponding occlusion pressure value is 52 in FIG. 3(f), which is the mean arterial pressure of the patient.

If the second derivative of FIG. 4(d), for instance, has two equal most negative values 68 and 69, the microprocessor 28 averages their corresponding occlusion pressure values 70 and 71 at step 94. An average of the occlusion pressure values 70 and 71 yields occlusion pressure value at 52, which is equal to mean arterial pressure. At step 93, the microprocessor 28 derives diastolic pressure in light of the mean arterial pressure and systolic pressures. Diastolic pressure equals 1/2(3.multidot.mean arterial pressure-systolic pressure).

The microprocessor 28 of this invention comprises an INTEL microprocessor, Model Nos. 8088 or 8086, for instance. Code for the microprocessor has been written on a word processor in ASCII format and assembled by an INTEL ASM86 assembler. Resulting hexidecimal code from the assembler is then programmed into an erasable, programmable, read-only memory. The erasable, programmable, read only memory is used by the microprocessor 28 to relate the occlusion pressure curve, static pressure curve, and first, second, and third time derivatives of the static pressure curve. The code is attached as an Appendix.

The invention improves the accuracy of a blood pressure measurement system. The invention recognizes the effects of slow, large noise signal on the system, and filters out those effects to yield an accurate measurement of blood pressure. ##SPC1##