 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The subject invention relates generally to the field of medical devices and
methods for monitoring physiological parameters of the body and, more
particularly, such devices and methods which are capable of monitoring on
a long-term basis various physiological constituents present in the
bloodstream.
Six million Americans have diabetes mellitus. On some schedule, all of
these patients need to monitor their blood glucose levels to keep their
disease under control. This monitoring is done by urine testing, which
indirectly reflects blood glucose, by intermittent blood glucose tests by
venipuncture or blood glucose monitoring by finger prick and strip
analysis. There are many inaccuracies associated with urine testing, and
many patients are reluctant to do an adequate number of blood tests
because of the pain involved. As a result many diabetics do not maintain
good glucose control.
Growing evidence proves that poor glucose control is a causative factor in
development of the secondary complications of diabetes. These secondary
complications take a great toll in morbidity and mortality. Twenty to
twenty five per cent of End Stage Renal Disease is caused by diabetes.
About 5000 diabetics become blind annually, and about 20,000 require
amputations. Diabetes increases the risk of cardiovascular disease and
diabetics also suffer from painful and sometimes disabling neuropathies.
The monetary cost for the treatment of the secondary complications of
diabetes is extremely high. The current cost of dialysis for End Stage
Renal Disease is about $25,000 per patient annually. The annual hospital
costs for amputations is presently about $250 million. Disability payment
and rehabilitation services for blind diabetics cost about $45 million
annually.
The best prospect for reducing the morbidity and mortality of diabetes lies
in technological developments which will provide better blood glucose
control. Insulin infusion pumps and finger-stick home blood glucose
monitoring are steps in that direction. The development of an implantable
glucose monitoring device would make home blood glucose monitoring
simpler, less painful and more acceptable for the diabetic population.
This would be a significant step toward improving blood glucose control
since it would provide much more information than available with any
current method of glucose monitoring. The greatly increased knowledge of
blood glucose levels obtained by an implantable monitoring system would
permit analysis of the basic kinetic parameters of insulin in each
patient, e.g. insulin sensitivity and half life. Also, a reliable long
term glucose sensor could be combined with automatic insulin infusion
systems already available to form an "artificial pancreas". With such a
device it would be possible to maintain blood glucose within normal limits
with little patient intervention.
The only hospital use instrument which has been on the market for constant
blood glucose monitoring is the Biostator, manufactured by Miles
Laboratories, Inc., Elkhart, Ind. This instrument constantly withdraws
blood to monitor blood glucose concentration and infuses insulin in
response to the blood glucose level. This type of instrument is very
expensive and is therefore available in only a relatively few hospitals.
Further, because the Biostator device requires a continuous flow of blood
without return to the patient there is a limit on the amount of time over
which the instrument can be used, and there are also problems and risks
involved with the vascular access.
A number of different approaches have been taken to develop an implantable
glucose sensor. Most approaches utilize a chemical reaction of glucose
which actually consumes the glucose in the process of measuring it. Thus
they are sensitive to the mass transfer coefficient of glucose to the
sensor. Fibrous tissue formation around the sensor changes the calibration
of the device. Secondly, those with enzyme components suffer degradation
of the enzyme after several days of use.
Enzymatic glucose electrodes utilize an immobilized enzyme, glucose
oxidase, which reacts selectively with glucose, in conjunction with an ion
selective electrode which measures the decrease of one of the reactants
(0.sub.2) as reported in Gough D.A. et al.: Progress toward a potentially
implantable, enzyme-based glucose sensor. Diabetes Care 5:190-198, 1982,
or the increase of one of the products (H.sub.2 O.sub.2). The change in
potential or current at the electrode can be used to make kinetic
measurements or the steady state current or potential can be used for
equilibrium measurements as disclosed in Guilbault G.G.: Enzymatic glucose
electrodes. Diabetes Care 5:181-183, 1982.
An electroenzymatic sensor disclosed in Clark et.al.: Implanted
electroenzymatic glucose sensors. Diabetes Care 5:174-180, 1982 involves
the enzymatic oxidation of glucose by glucose oxidase and the production
of H.sub.2 O.sub.2. The H.sub.2 O.sub.2 is measured voltametrically at a
Platinum electrode. The current produced by H.sub.2 O.sub.2 is directly
proportional to the glucose in blood, plasma or tissue fluid in the 0 to
100 mg/dl region. At higher glucose concentrations there is a non-linear
increase in current with increasing glucose concentration.
A glucose oxidase electrochemical sensor which detects the production of
H.sub.2 O.sub.2 has been made in a needle form and has functioned up to 3
days in subcutaneous tissue as reported in the publication Shichiri M. et
al.: Use of wearable artificial pancreas to control diabetes. Progress in
Artificial Organs 782-787, 1983. When this sensor was coupled with a
micro-computer and an insulin infusion system, glucose control was
achieved which was superior to that achieved with conventional treatment.
After three days there was a fixation of proteins and blood cells to the
membrane of the electrode, resulting in diminished function.
An O.sub.2 sensitive enzymatic glucose sensor which can be inserted into an
arterio-venous shunt is disclosed in the following publication: Kondo T.
et al.: A miniature glucose sensor, implantable in the blood stream.
Diabetes Care 5:218-221, 1982. This sensor can function 200 hours with a
10% loss in activity. Another publication, Ikeda et al.: Comparison of
O.sub.2 electrode type and H.sub.2 O.sub.2 electrode type as a glucose
sensor for the artificial B-cell. Prog. in Artificial Organs 773-777, 1983
compared the in vivo function of the O.sub.2 sensor with a H.sub.2 O.sub.2
electrode in a vascular access and found the O.sub.2 electrode responded
better to changes in blood glucose. However, this sensor is impractical
because of the amount of vascular surgery necessary to install the shunt.
Another approach to a glucose sensor is the catalytic electrode sensor,
which is based on the electrochemical oxidation of glucose on a platinum
electrode. Such sensors are reported in the following publications: Lerner
H. et al.: Measurement of glucose concentration with a platinum electrode.
Diabetes Care 5:229-237, 1982; Lewandowski J.J. et al.: Amperometric
glucose sensor: Short-term in vivo test. Diabetes Care 5:238-244, 1982.
The applied voltage is varied and the current response is measured. The
current-voltage curves vary with glucose concentration. Other substances,
such as amino acids and urea, can affect the output of this sensor, but
use of a compensated net charge method of evaluating the response improves
the sensitivity. Another problem is change in the loss in catalytic
activity over time. Overall, this type of sensor has not demonstrated the
selectivity or sensitivity necessary for a useful glucose sensor.
Several other technologies for glucose sensors depend on chemical or
physical properties of glucose such as its affinity for lectins described
in Schultz J.S. et al.: Affinity sensor: A new technique for developing
implantable sensors for glucose and other metabolites, Diabetes Care
5:245-253, 1982, its optical rotation in solution described in Rabinovitch
B. et al.: Non-invasive glucose monitoring of the aqueous humor of the
eye: Part I. Measurement of very small optical rotations, Diabetes Care
5:254-258, 1982; and March W. F., et al.: Non-invasive glucose monitoring
of the aqueous humor of the eye: Part II. Animal studies and the scleral
lens, Diabetes Care 5:259-265, 1982, or its osmotic effect, Janle-Swain E.
et al.: A hollow fiber osmotic glucose sensor, Diabetes 33: Supp. 1, 176A,
1984. These approaches do not consume glucose, but rather depend on the
concentration of glucose at the device site reaching an equilibrium with
tissue glucose. None of these devices have proven to be totally
satisfactory.
In the Schultz et al. study, a monitoring system is disclosed which
operates based on the ability of glucose and a fluorescein-labeled dextran
to bind competitively to the lectin Concanavalin A (Con A). Con A can be
bound to the inside of a hollow fiber through which glucose can diffuse.
Fluorescein labeled dextran is added to the inside of the fiber. The
amount of fluorescein labeled dextran is added to the inside of the fiber.
The amount of fluorescein-labeled dextran displaced from the Con A is
measured by an argon laser fiber optic system. This system has responded
to differences in glucose concentration in vitro, but less than the
theoretical response was obtained. In vitro tests have demonstrated that
the Con A can remain bound to the fibers for eight days.
Another study, Shichiri M. et. al.: Telemetry glucose monitoring device
with needle type glucose sensor: A useful tool for blood monitoring in
diabetic individuals, Diabetes Care 9:298-301, 1986, supports that
measurement of glucose in the subcutaneous tissue does provide an adequate
indication of blood glucose. This study discloses a correlation in
diabetic patients between tissue glucose and blood glucose, with
correlation coefficients ranging from 0.89 to 0.95. This work indicates a
five minute delay between changes in blood glucose and subcutaneous tissue
glucose, with tissue glucose being 6 to 22% lower than blood glucose. This
decrease in subcutaneous glucose versus blood glucose is due to the
metabolism of glucose by subcutaneous tissue. The level of glucose which
is obtained depends upon the metabolic rate of the subcutaneous tissue,
the blood flow, the degree of fibrous tissue in the space, and the rate of
fluid transfer across the capillary wall.
The goal of all of these studies was to develop sensors for permanent
subcutaneous placement. Currently there does not exist an implantable
glucose sensor which will function for an extended period of time in vivo.
There are a number of sensors which function in vitro and some which
function well for a few days in vivo, but none have proved effective over
long periods of time. Measurement of blood glucose is done by diabetic
patients at home by the finger-stick method. A drop of blood is placed on
a paper strip impregnated with glucose oxidase and a chromophore. The
color change produced by the glucose in the blood is determined visually
or with a small hand held reflectance meter. In hospitals blood glucose
may be measured at the bedside by the same finger-stick and strip method
used by patients at home or venous blood samples may be analyzed
automatically in the laboratory using glucose analyzers which are usually
based on spectrophotometric or electrocatalytic analysis methods.
SUMMARY OF THE INVENTION
A capillary filtration and collection system for long term sampling of an
ultrafiltrate of blood according to one embodiment of the present
invention comprises a porous filter made of a material compatible for
long-term patency inside the body in fluid communication with blood
capillaries and having pore sizes not greater than approximately 300,000
daltons. There is further provided an ultrafiltrate collector connected to
the filter in fluid communication therewith and adapted to extend
externally of the body with the filter implanted inside the body. A vacuum
generating means is also provided for withdrawing the ultrafiltrate
through the filter and into the ultrafiltrate collector.
The present invention also comprises a method for continuously monitoring
the blood level of a physiological constituent of blood. According to one
embodiment of the method, a porous fiber filter is implanted within the
body in fluid communication with blood capillaries. One desirable location
for placement of the filter is in the subcutaneous tissue, where it
removes fluid through capillary walls and through the gelatinous
subcutaneous tissues. Other possible locations include within the muscles
of the bowel wall or peritoneum. An ultrafiltrate of blood is filtered
from the capillaries by exerting a negative pressure within the porous
fiber filter so as to cause the ultrafiltrate to flow by convection from
the capillaries through the filter at a desired flow rate. The
ultrafiltrate is then collected in a reservoir located externally of the
body from which it is periodically removed to allow chemical measurement
of a physiological constituent, such as glucose, present in the
ultrafiltrate. Alternatively, fluid is removed and immediately transferred
to a measurement device.
ADVANTAGES
The device and technique of the present invention offers a unique approach
to the problem of continuous blood glucose measurement. Other attempts at
the development of an implantable glucose monitor have sought to make the
entire system implantable. Making the fluid removal portion internal and
the measurement system external eliminates many of the problems that arise
from a totally implantable system and increases the options for glucose
measurement. Among these options are optical paper strip methods for the
measurement of glucose. Such methods, which have been well developed,
require consumable supplies and are well suited for use in an external
system. There are also other options for external measurement systems,
such as the use of catalytic enzyme electrodes. These have a limited in
vivo lifetime which limit their usefulness in an implantable sensor, but
they can be easily replaced if used externally. Thus, when used with the
present invention, they could serve as a workable alternative to paper
strips for glucose measurement outside the body.
The implantable fiber system of the present invention has great potential
for use by diabetics in self-monitoring of glucose levels in their own
home. Patients may monitor their glucose levels at home using commercially
available strips and reflectance meters as often as six times an hour
without the necessity of sticking themselves and drawing blood.
Alternatively, the patient may wear the sample collector which will
collect samples over a 3-5 hour period. Fluid samples in the collector can
be analyzed at a convenient time either by the patient or at the doctor's
office. The collector may be used with an optical monitoring system which
we are developing which will automatically withdraw and analyze samples
and mathematically relate the sample to the collection time. Manual
analysis of samples from the collector would also be possible by
conventional methods. The painlessness of such a monitoring system offers
the possibility of better control for diabetics because it is more likely
to be faithfully used.
The subject invention could also be used in a hospital environment. An
intensive care unit version of the invention would perform glucose
measurements in a continuous manner using the same type of implantable
filter used in out-patient applications. Because only small amounts of
ultrafiltrate are needed for analysis of glucose levels using the present
invention and because the device is not within the blood vessels and
because red blood cells are not lost, the only limit on the length of time
that the device of the present invention could be used would be the useful
life of the filter unit in the body. Since the filtrate is a clear liquid
rather than blood, chemical analysis is easier. An automatic hospital
glucose monitor developed for use with the subject invention would be less
complex and therefore less expensive than vascular access/continuous blood
withdrawal glucose monitoring systems such as the Biostator device, thus
making such a system financially feasible for more hospitals. This machine
can be identical to the one used for intermittent analysis of samples from
the above mentioned collecting reservoir.
The subject invention would provide improved control of blood glucose over
the current hospital practice of venipuncture or finger-stick glucose
monitoring in diabetics. Glucose monitoring could be performed at
virtually any desired interval. The discomfort of repeated venipunctures
or finger sticks would be eliminated along with the necessity of having
lab personnel periodically visiting the patient to draw blood. In
addition, there would be none of the delays involved in obtaining glucose
concentration results from the laboratory. Treatment decisions could be
made much more rapidly and the time needed to obtain good glucose control
would be shortened. Improved control, in most cases, will hasten recovery
and shorten hospital stays. Effective storage and display capability of
the data would also mean that the information would be readily available
whenever needed.
In addition to monitoring glucose, the subject invention may also be useful
in monitoring and measuring the levels of other physiological or
pharmacological substances present in blood. Some preliminary testing has
indicated that sodium and potassium could be monitored using the capillary
filtration and collection techniques of the present invention. Measurement
of BUN, drug or hormone levels and capillary blood gases (CO.sub.2 and
O.sub.2) are other potential applications of a capillary filtrate
collector. Such systems would be highly cost effective and would also
provide nearly constant clinical monitoring without the need for
intermittent drawing of venous blood samples. The subject device may be of
particular use in neonates or pediatric patients where drawing blood is a
difficult and trying experience.
Accordingly, it is an object of the present invention to provide an
improved method and device for monitoring the blood level of a
physiological constituent of blood.
Related objects and advantages of the subject invention will become more
apparent by reference to the following drawings and detailed description.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a fragmentary elevational view, partially in section, showing the
capillary filtration device of the present invention.
FIG. 1a is an enlarged fragmentary perspective view showing a section of
the fibers depicted in FIG. 1.
FIG. 1b is an enlarged fragmentary perspective view showing an embodiment
of the filter of the present invention having two fibers in a loop
configuration.
FIG. 2 is a diagrammatic view of the capillary filtration device with the
filter portion inside a body and the device equipped with a coiled tube
collector and a syringe evacuating the collector.
FIGS. 3 and 4 are graphs illustrating the relationship between blood
glucose and tissue glucose found in test animals using the present
invention.
FIGS. 5-7 depict successive stages in performing the method of the
implantation.
FIG. 8 is a diagrammatic view of the capillary filtration device with the
filter portion inside a body and the device equipped with a roller pump.
FIG. 9 is a fragmentary perspective view showing an embodiment of the
filter of the present invention having two flat sheet membranes.
DESCRIPTION OF THE PREFERRED EMBODIMENT
For the purposes of promoting an understanding of the principles of the
invention, reference will now be made to the embodiments illustrated in
the drawings and specific language will be used to describe the same. It
will nevertheless be understood that no limitation of the scope of the
invention is thereby intended, such alterations and further modifications
in the illustrated device, and such further applications of the principles
of the invention as illustrated therein being contemplated as would
normally occur to one skilled in the art to which the invention relates.
The approach of the subject invention to the problem of constant long term
glucose monitoring is to apply a negative pressure to ultrafiltration
fibers implanted in interstitial spaces, such as subcutaneous tissue, and
withdraw a plasma ultrafiltration from blood capillaries located in the
region around the fibers. The fibers remove the plasma ultrafiltrate by
convection. The fluid quickly leaves the blood capillaries, transverses
the interstitial space between the capillaries and the fiber filter and
enters the filter.
In convective transfer of fluid across membranes, chemicals smaller than
the pore size of the membranes remain at the same concentration on either
side of the membrane. The pore size of a blood capillary is approximately
50,000 daltons. Therefore, if a fiber is used with a pore size of 30,000
daltons, then low molecular weight non-protein bound chemicals such as
glucose, sodium, potassium, urea, creatinine, O.sub.2 CO.sub.2 etc. will
appear in the ultrafiltrate at the same concentration as in the blood. The
rate of transfer of the fluid through the interstitial tissues is
sufficiently fast that metabolism of chemicals by interstitial tissue
cells will not cause a significant decrease in concentration of glucose or
other chemicals. Present studies indicate that fluid flow rates between
about 20 to 500 microliters per hour are acceptable for this purpose.
A first embodiment of the capillary filtration device 10 of the present
invention is illustrated in FIGS. 1 and 1a. Generally, the device 10
includes a filter 11 comprised of one or more implantable filter fibers 16
connected to a length of conducting tubing 12 which serves as a means to
collect ultrafiltrate fluid passing through filter 11 externally of the
body. The conducting tubing 12 is comprised of a short section 13 of
silicone or polyurethane tubing adapted to extend inside the body and a
longer section 14 of polyvinylchloride (PVC) tubing adapted to extend
externally of the body. A rigid metal tube 15 of relatively short length
extends inside silicone tubing section 13 and PVC tubing section 14 at
their juncture and serves to permit the silicone and PVC tubing sections
to be heat sealed together without allowing the lumen to collapse.
As seen clearly in FIG. 1a, in one configuration the filter 11 is formed of
three elongate ultrafiltration fibers 16 having a hollow tubular shape. An
alternative embodiment of the invention shown in FIG. 1b incorporates a
filter configuration 11' formed of two elongate ultrafiltration fibers 16
having both their ends secured by adhesive inside the lumen of tubing
section 13 so as to form a loop. This alternative configuration makes the
implanted portion of the device shorter and therefore increases patient
acceptability. The loop configuration also holds the fibers apart slightly
and therefore increases the surface area exposed to tissue. The fibers 16
are fixed at their proximal ends by an adhesive inside the lumen of the
tubing section 13.
In the FIG. 1a elongated filter configuration, the distal ends of the
fibers 16 are secured together by heat sealing while the proximal ends are
secured to the lumen wall of tubing section 13 by an epoxy or polyurethane
adhesive. A cuff 17 is secured around the silicone tubing section 13 at
the junction of the fibers 16 and tubing section 13. The purpose of cuff
17 is to allow tissue ingrowth into the cuff and seal off any path for
passage of skin bacteria around the device, and also to serve as an
anchoring means to anchor the device in the body after sutures are
removed. Cuff 17 is preferably made of material which promotes ingrowth of
fibroblast and fibrous tissue. Suitable materials are Dacron, PTFE or
textured polyurethane.
The purpose for the silicone (or polyurethane) and PVC tubing sections is
that it has been found that silicone tubing bonds better to the cuff 17
and fibers 16 than PVC tubing and has improved the biocompatibility of the
device in the body. The PVC tubing is preferred for the remainder of the
length of conducting tubing because it is sufficiently rigid so that it
will not collapse under negative pressure conditions during operation
while also possessing good flexibility and a low permeability to air when
placed under vacuum. It also serves as a pumping segment for a roller pump
which is one method of creating a vacuum for fluid removal.
At the proximal end of the tubing section 14 there is an 18 gauge hollow
needle 19 and hub 20 which serves to facilitate the transfer of the
ultrafiltrate fluid from the conducting tubing into a suitable reservoir
or a chemical analysis machine.
Negative pressure for ultrafiltrate fluid removal can be accomplished by
evacuating air from the conducting tubing 12 with a syringe (FIG. 2) or a
pre-evacuated rigid container such as a "vacutainer" tube. Alternatively,
the fluid may be removed by inserting a segment of the tubing section 14
into a roller pump 21 depicted in FIG. 8 which pumps the fluid into a
reservoir 21a or into a chemical analysis machine. A suitable roller pump
for this purpose is a perisaltic roller pump manufactured by Cormed, Inc.
of Medina, N.Y. This pump generates a vacuum of over 700 mmHg. For the
ambulatory hospital patient or for home use, however, a withdrawal system
based on vacuum alone is preferred because the need for pumps and an
associated power supply, such as electrical batteries, is eliminated.
At high negative pressures (-300 to -750 mmHg) the device removes
approximately 10 volumes of gas per volume of liquid from the subcutaneous
space. It has been found in animal tests of this device that if the lumen
of the conducting tubing 12 is sufficiently small in diameter, the
ultrafiltrate fluid passing through the tubing will remain in distinct
boluses separated by air. Tubing having an inner diameter less than about
0.115" should be practicable for this purpose, while an inner diameter of
0.045-0.055" is more preferred. This phenomena is useful in conducting
time studies of the fluid samples based on their position along the length
of the tubing. In addition, the gas speeds transfer of the liquid samples
through the conducting tubing.
A fluid collection device is shown in FIG. 2 which provides vacuum for
filtrate removal and preserves the separation of fluid samples which
occurs in the conducting tubing 12. A mathematical analysis can later
determine the relationship between time and sample position at the end of
the use of the collector. As illustrated in FIG. 2, the collector 22
includes a coil of flexible plastic tubing 23 having a needle hub 25 for
receiving needle 19 at one end, and an injection/sampling port 26 at the
other end. The hub 25 and port 26 are of conventionally known
construction. The tubing 23 is coiled on a layer of transparent adhesive
tape 28 and is covered on the opposing side with another layer of tape.
The tape stabilizes the tubing and also decreases air permeation into the
tubing. With the collector attached to the conducting tubing 12 by
insertion of needle 19 into hub 25, the collector can be evacuated with a
syringe 30 through port 26 to provide the negative pressure for fluid
withdrawal. When the collector 22 is filled the collector may be removed
and replaced with a new collector. The ultrafiltration samples may be
removed from the collector for sampling through injection port 26.
The relationship between the sample position along the length of coiled
tube 23 and time depends on the method of vacuum application. If a roller
pump is used on tubing section 14 with port 26 open, the negative pressure
is uniform over time. This results in a linear relationship between sample
position and time. If the sole source of vacuum pressure is from
evacuating air from the coiled tubing, then the vacuum pressure will
decrease with time. The small fluid samples will be collected more rapidly
when the collector is first attached and evacuated and more slowly as the
vacuum decreases logarithmically over time. However, the relationship
between time and sample position can still be determined mathematically.
Two different types of ultrafiltration fibers have been tested in
laboratory animals. The first type are AN-69 polyacrylonitrile (PAN)
fibers used in the Hospal Filtral dialyzer manufactured by Rhone Poulenc
Industries of Paris, France. These fibers have homogeneous walls with
30,000 dalton molecular weight cut off and a filtration rate of 20
ml./hr.*m..sup.2 *mmHg. The second type of fibers which have been tested
are P-100 polysulfone fibers manufactured by Amicon Corporation of
Danvers, Mass. The Amicon polysulfone fibers come in a variety of
dimensions with a variety of molecular weight cut offs. These fibers have
a porous thin 1-2 micron skin on the luminal side of the fiber and a
porous rough outer surface. Due to breakage problems with the Amicon
polysulfone fibers after implantation, the PAN fibers are presently
preferred. However, stronger polysulfone fibers than the Amicon
polysulfone fibers which have previously been used and fibers with a
smooth outer skin are manufactured by A/G Technology Corporation of
Needham, MA. It is perceived that other types of polymer hollow fibers
could also be suitably employed, such as for example either polypropylene
or polycarbonate fibers.
Preferably the molecular weight cut off of the pore size of the fiber needs
to be less than approximately 300,000 daltons to exclude fibrinogen
protein molecules from the ultrafiltrate which would otherwise promote
clot formation within the conducting tubing. If an optical monitoring
system is used to measure glucose, the molecular weight cutoff should be
less than about 60,000 daltons to exclude hemoglobin from the
ultrafiltrate. Because the molecular weight of glucose is approximately
180, the molecular weight cutoff of the fiber could be as small as
1000-2000 daltons without affecting the transmission of glucose through
the fiber. However, with molecular weight cutoffs of 30,000-50,000 daltons
other molecules such as uric acid, vitamin B.sub.12, insulin, etc. could
be measured in the filtrate. A molecular weight cutoff in the range of
30,000 to 50,000 daltons is suitable for these reasons.
It is well known in the field of blood filtration that if proteins are
excluded by a membrane, layers of protein will form on the blood side of
the membrane and diminish the flow of fluid. Stability of fluid removal
rate from the current device has been demonstrated for 3-39 days of
implantation. In the body interstitial spaces the lymphatic system serves
to remove excess protein from around the fibers to maintain constant flow
of filtrate.
The number, size and length of the fibers may also be varied within certain
design considerations. For example, in the case of the polysulfone fibers,
because of their thickness only one fiber has been used. When much thinner
fibers are used, such as the PAN fibers, it is possible to increase the
number of fibers which are employed to increase the ultrafiltrate flow
rate and still permit the fibers to be subcutaneously implanted through a
hollow needle. Polysulfone fibers have been used having lengths of about
20-25 cm. For the PAN fibers, a length of 17 cm. has been found to provide
acceptable flow rates.
It may also be possible to change the configuration of the filter from
hollow fibers to other configurations, such as flat sheet polymeric
membranes 16, shown in FIG. 9, glued or otherwise fixed together around
their circumferences to form two-sided membrane filters.
The flow rate of ultrafiltrate depends upon a number of factors, including
the size and configuration of the filter and the amount of vacuum applied.
The ultrafiltrate sample size necessary for the particular assay procedure
determines the desired flow rate and vacuum. To some extent, these factors
will in turn be dependent upon whether the device is being used at home
for periodic glucose monitoring or for continuous monitoring in hospital
applications. Since most glucose analysis methods require a sample size of
at least five microliters and since continuous glucose monitoring for
hospital patients coupled with automatic insulin infusion would require
sampling at least as frequent as every 15 minutes, it is perceived that a
minimum flow rate of 20 microliters per hour is desirable.
The present invention has been successfully tested in vivo in diabetic dogs
for periods as long as 39 days. Glucose levels in the ultrafiltrate were
compared to venous blood glucose levels and were very highly correlated.
FIGS. 3 and 4 are graphs illustrating the correlation between blood
glucose and filtrate glucose levels observed in two test dogs using the
device of the present invention. In both of the tests, PAN fibers were
used as the filtering medium. Filtrate fluid samples were collected over
half hour intervals. Blood samples were collected either at the end of the
filtrate sample collection period or at the mid-point of the collection
period. Glucose was analyzed using blood glucose test strips and a
reflectance meter. Regression analysis of the blood an ultrafiltrate
glucose yielded correlation coefficients of r=0.840-0.878, p=0.000.
One method for placement of the device of the present invention for long
term monitoring of blood glucose in a patient is performed as follows. Two
small scalpel incisions spaced apart approximately the length of the
fibers are made into the skin 31 of a locally anaesthesized patient at a
suitable subcutaneous location, such as on the leg, buttocks, abdomen or
arm of the patient. A hollow introducer needle 32 is advanced into the
skin through the proximal incision (FIG. 5) and tunneled through the
subcutaneous tissue space until the distal needle end is caused to exit
the skin through the distal incision (FIG. 6). The fiber filter 11 portion
of the device 10 is then inserted inside the hollow needle 32 until the
cuff 17 stops at the proximal end of the needle. As the needle is pulled
out, the distal end of the cuff is positioned beneath the s | | |
