CA1107467A - Polycarbonate membranes for use in hemodialysis - Google Patents

Abstract

POLYCARBONATE MEMBRANES FOR USE IN HEMODIALYSIS

ABSTRACT OF THE DISCLOSURE

Membranes for hemodialysis prepared from polycarbonate materials have superior transport properties for middle molecular weight molecules than do cellulosic membranes while maintaining approximately the same ultrafiltration rate as cellulosic membranes. A percentage of patients treated using dialysers containing polycarbonate membranes show improved hematocrit and neurobehavioral functions.

Description

Thi~ invention relateq to new and improved polycarbonate membranes which are especially useful for hemodialysis.
Hemodialysis membranes Eor use in the artificial kidney are at the present time generally made of cellophane materials. The best of these materials currently available for such purpose has been found to be a celluloseregenerated from a cuproammonium solution, plasticized with glycerol and identified by the trade mark "Cuprophan". Although Cuprophan membranes provide ultrafiltration rates and clearance of low molecular weight solutes within the desirable range~ for proper hemo-dialysls, they still have many deficiencies whlch prevent them from being completely satisfactory as hemodialysis meubrane~. Certain toxins which it is thought necessary to remove from the blood by hemodialysis are "middle molecules", i.e., molecules of molecular weights in the range of 300 to 5,000. Such middle molecules pass through Cuprophan membranes at rates much slower than is deslrable. Babb et al ("The Geneffl~ of the Square ~eter-Hour Hypothe~i~" TransASAI0, Vol. XVII, ~1971 Pages 81-91) advanced the hypothesls that higher molecular weight metabolites (middle molecules) are important uremic toxins. The blood from normal persons toes not show the presence of mlddle molecules while uremic patients exhibit a signiflcant amount of ~: .

~ 7467 ; ' '"

middle molecules, particularly in the range of 300 to 1,500 molecular weight. In testing Babb's hypothesis, it was found that metabolites hauing a molecular weight less than 300 or greater than 2000 were not believed to be causing uremic abnormalities and in fact, metabolites in the 300 to 1,500 molecular weight range were the predominant causes of uremic toxicity and neuropathy. (Babb et al "Hemodialyzer Evaluation By Examination of Solute Molecular Spectra" Trans.ASAIO, Vol XVIII (1972) pg. 98-105). Popovich et al, ("The Prediction of Metabolite Accumulation Concomitant With Remal Insufficiency: "The Middle Molecule Anomoly"
Trans.ASAI0, Vol XX(1974) p 377-387) discuss the results of numerous clinical investigators who explored the connection of neuropathy to middle molecule concentrations. Additionally, the burst and tear strengths of Cuprophan membranes are lower than is desirable in materials employed in hemodialysis and their shelf-life is low, apparently due to migration of plasticizer during storage. Further, the permeability of the Cuprophan membranes has been found to vary from batch to batch and to decrease on aging. Lastly, it is very difficult to cause adhesion between Cuprophan and other materials and between Cuprophan and itself. Thus, it is difficult to utilize improved hemodialyzer designs requiring leak-proof compartments which depend upon the membrane material for sealing off blood from dialysate solution and blood and dialysate solutions from the atmosphere.
The membranes prepared from the present invention are significantly improved over the state-of-the-art materials, e.g. Cuprophan in the following areas.
1. Polycarbonate membranes permit clearance of critical "middle molecules" up to 4 times greater than Cuprophan in comparable tests while exhibiting an ultrafiltration rate of l.25 to 2 times Cuprophan membranes.

.

X

74~7

2. The burst strength of polycarbonate membranes is 1.5-2 times that of Cuprophan.

3. The latitude of membrane properties achievable with poly-carbonates is considerable and can be arranged in accordance with clinical needs.

4. Polycarbonate membranes are stiffer than Cuprophan in the wet state. This property results in thinner blood layers in dialy-zers, more efficient dialysis and lower blood priming volume.

5. Polycarbonates are heat-sealable wet or dry permitting wide latitude in dialyzer design.

6. Due to greater efficiency of dialysis with polycarbonate mem-branes projections indicate a greatly reduced dialysis time (9 hrs/wk) compared with Cuprophan.

7. Dialysis procedures using polycarbonate membrane have resulted in the improved physical condition o~ dialyzed patients including increased hematocrit, clecreased blood pressure, improved motor nerve conduction velocity ancl reduction in symptoms of neuropathy.

8. Polycarbonate membrane~ are up to 36.6~ more compatible with blood than are Cuprophan membranes.
In attempting to develop hemodialysis membranes with mechanical and transport properties superior to those of Cuprophan it ilas previously been proposed, by two of the present co-inventors, to fabricate membranes of polyether-polycarbonate block copolymers contalning a balance of hydrophobic aromatic polycarbonate blocks, whict-l impart toughness, and hydrophilic polyether blocks~ which impart water and solute permeability.
The polycarbonate system was chosen for dialysis membrane development because of the outstanding mechanical properties shown by corlmercial ~7467 polycarbonate, the very low thrombogenicity exhibited by properly heparinized polycarbonate surfaces, the ease of forming this polymer type into various configurations such as films and fibers, and the many synthetic possibilities for chemical modification of the basic aromatic polycarbonate backbone structure to achieve desired membrane transport properties. As disclosed in the "Proceedings of the 5th Annual Contractors' Conference of the Artificial Kidney Program of the National Institute of Arthritis and Metabolic Diseases", U.S. Department of Health, Education and Welfare (1972), pages 32-33, gelled membranes were prepared from polyether-polycarbonate block copolymers by means of the phase inversion technique, i.e., casting a sGlution of the copolymer in a suitable solvent onto a substrate surface to form a layer which is allowed to dry only partially and which is then immersed in a liquld gelation medium in which the copolymer is insoluble but which is miscible with the solvent, employing chloroform as the casting solvent and methanol as the gela-tion medium. The gelled membranes resulting from such procedure, while exhibiting considerable superiority over Cuprophan membranes in their permeabilities to solutes in the middle molecule range, were found, however, to possess several drawbacks to their practical use as hemodialysis membranes. First of all, their ultrafiltration rates were 2 to 5 times that of Cuprophan membranes, which would be clinically unacceptable for hemodialysis as presently administered due to the possibility of dehydration of the patient occurring during treatment.
Secondly, their burst strength was no more, and in many cases, less than that of Cuprophan membranes. Thirdly, attempts at continuous casting of the membrane on production-type machinery in widths suitable for -e~
~,.~.

~ 7467 use in commercial hemodialyzers, presented further problems which rendered the methanol gelation procedure impractical for commercial hemodialysis membrane production.
The most serious problem encountered was the frequent occurrence of gross leakage of albumin through the membranes during ultrafiltration testing, and which was found to be attributable to holes or other imperfections in the ultrathin surface of the membrane which forms the barrier between the blood and the dialysate or flushing solution. All of these membranes are referred to as being "anisotropic"
or "skinned", which means that their two sicles are significantly differ-ent from each other, one side being relatively smooth and the other side being relatively rough and porous. The smooth side is the "barrier" layer which faces the blood during hemodialysis and is quite thin, on the order of 0.05 to 0.2 microns. The remainder of the membrane merely functions as a support structure and is about 25 to 30 microns in thickness. The integrity of the barrier layer is crucial to the performance of the membrane in dialysis. Any perforation, puncture or other compromise of the integrity of the barrier layer destroys the usefulness of the membrane and all materials in contact with the membrane merely leak through. It has now been proven by electron microscopy that the methanol-gelled polycarbonate membranes are formed with their barrier layer on the side of the membrane contacting the casting surface rather than the side of the membrane facing the air during drying. The significance of this fact is that continuous casting of these membranes on production-type machinery involves continuously peeling the delicate barrier layer off of the casting surface during tl-e process, making it almost impossible to ... . ..
~ . 5 -11~7467 .
maintain the integrity of the barrier layer and obtain a membrane suitable for use in hemodialysis. Also, it was found that long term exposure of the membrane to methanol affects the membrane properties, thereby necessitating the quick and extensive flushing or washing of the membrane to remove the methanol therefrom and replace it with water in order for the membrane to have adequate shelf-llfe. One ; additional problem presented was the impracticality of employing large volumes of methanol as the gelation medium due to the cost, toxicity and flammability of this material.
Membranes of polycarbona~e type have been made by other investigators such as suggested in British Patent Specification 1,395,530 but these membranes have been found unsuitable for hemodialysis purposes.
See also Kesting, J. Macromol? Sci. ~_hem), A4(3), pp. 655-664 (1970);
U.S. Patents Nos. 2,964,794, 3,031,328, 3,450,650, 3,526,588 and 3,655,591; and British Patent Specificatlon No. 1,059,945.
It is therefore an object of the present invention to provide hemodialysis membranes having improved permeability to solutes in the middle molecule range as compared with presently available hemodialysis membranes, while maintaining low molecular weight solutes.
Another object of the invention is to provide hemodialysis membranes having improved burst and tear strengths as compared with presently available hemodialysis rnembranes.
A further object of the invention is to provide hemodialysis membranes having improved shelf-life as compared with presently avail-able hemodialysis membranes. A further object of the present invention is to provide hemodialysis membranes having improved sealability over presently available hemodialysis . - , membranes making possible leak-proof hemodialyzer compartments through simple heat-sealing of the membranes.
Still another object of the invention is to provide a process for producing gelled polycarbonate membranes useful for hemodialysis and having the improved properties as set forth in the preceding objects, which is easily and economically adaptable to large scale machine produc-tion without impairing the integrity of the barrier layer of the membrane.
Thus, by one aspect of this invention there is provided a membrane of hydrophilic polycarbonate polymer adaptable for use in a hemodialysis apparatus as a means for preferential removing of middle molecular weight molecules from blood, said membrane having a diffusive permeability measured at 37C to sodium chloride of about 630 to 825 cm/min x 10 4, a permeability to urea of about 650 to 850 cm/min x 10 4, a permeability to vitamin B12 of greater than 90 cm/min x 10 4 and an ultrafiltration rate of less than 5.5 ml/hr/M /mm Hg.
By a preferred aspect of this invention there is provided a membrane as described above and which consists of a polyether-polycarbonate block copolymer containing 5 to 35% by weight of repeating alkylene ether carbonate units and from 95 to about 65% by weight of repeating bisphenol A - carbonate units.
As set forth in our copending Canadian Application 222,510 filed March 19, 1975 membranes according to the present invention can be produced from a polyether-polycarbonate block copolymer by the phase inversion technique employing an aqueous gelation system with water as the gelling medium and a water-miscible organic solvent as the casting solvent. More specifically, this process comprises casting on to a sub-strate surface having a smooth finish, a layer of casting solution com-prising a polyether-polycarbonate block copolymer containing from about 5 to about 35% by weight of the polyether component and a water-miscible organic solvent to~ether witll a co-solvent which acts as a swelling agent for the copolymer, drying the layer to partially evaporate the solvents ~..~

~ 7467 therefrom, immersing the partially dried layer in water to form a gelled membrane, and stripping the resulting gelled membrane from the substrate surface.
It has been found that gelled polycarbonate membranes produced in this manner, with water as the gelling medium, are formed wlth thelr barrier layer on the side of the membrane faclng the air during drying, rather than on the side of the membrane ~n contact with the casting sur-face as is the case with methanol-gelled polycarbonate membranes, which enables the gelled membrane to be readily stripped from the casting surface without impairing the integrity of the delicate barrier layer, thereby rendering large-scale machine production of such membranes practical.
The use of water as a gelling medium in place of methanol also facilitates large scale machine production in that water is, of course, less expensive, non-toxic and non-flammable, and also eliminates the necessity for the extensive flushing or washing of the membrane to remove the gelling medium therefrom as was required in methanol gelation. It has also been found that the water-gelled polycarbcnate membranes have considerably higher strength than either the methanol-gelled polycarbonate membranes or Cuprophan membranes.
Gelled polycarbonate membranes fabricated in accordance with the present invention have furthermore been found to be considerably superior to Cuprophan membranes in their permeabilities to solutes in the middle molecule range while maintaining ultrafiltration rates and clearance of low molecular weight solutes comparable to that of Cuprophan membranes. Moreover, it has been found that the ultra-filtration rates of the membranes fabricated in accordance with the present invention are controllable to levels comparable to those of Cuprophan membranes by proper selection of the molecular weight of the polyether-polycarbonate block copolymer used in fabricating the membrane.
The polycarbonate material from which the improved hemodialysis membranes are fabricated in accordance with the present invention is a polyether-polycarbonate block copolymer preferably containing from about 5 to about 35~ by weight of the polyether component. It has been found that this proportion of polyether blocks renders the normally hydrophobic poly-carbonate sufficiently hydrophilic so as to make it suitable for use as a ' ' ' ' '' ~7467 ,. , .

hemodialysis membrane. Certain of such block copolymers may be prepared, for example by the method of Goldberg (Journal of Pol~mer Science: Part C, No. 4, pp. 707-730 (1963) wherein a comonomer mixturc of from about 95 to 6S% by weight of 2, 2-(4, 4~-dihydroxydiphenyl) propane, generally known as bisphenol A, and correspondingly from about 5 to about 35% by weight of a polyether glycol such as polyethylene glycol, is reacted l with a carbonic acid derivative such as phosgene. A polyethylene glycol ¦ which is found to be particularly ~uitable is Carbowax~96000, which is a ¦ polyethylene glycol having an average molecular weight of 6700, altl~ough polyetl-ylene glycols o~ other molecular weights carl also be used, such as, l ~ 6~ (19 for example, Carbowax 600, Carbowax 1000 and Carbowax 4000, which are polyethylene glycols having molecular weights of 600, 1000 and 4000, re spectively.
In accordance with the abo~e the polyether-polycarbonate blocli copolylner consists of recurring units of the ~or~ula :.

~ ~ C ~C~ ~ ~ o (CHZC H2CI~--c~

: LC

~17967 wherein x is from about 10 to about 155 and preferably from about 12 to about 152 and a and b are chosen such that the bisphenol A carbonate unit (I) is about 95 to 65% of the weight of the recurring unit and the alkylene-ether carbonate unit (II) is about 5 to 35% of the weight of the recurring Ullit. Preferred values of a and b are about 80 and about 1 respectively.
Polyether glycols other than polyethylene glycols can also be used, such as, for example, polypropylene oxide-polyethylene oxide block copolymers as exemplified by members of the Pluronic diol series such as Pluronic@~68.
~ Polyether-polycarbonate block copolymers having molecular ; 10 weights ranging from about 50,000 to about 750,000 may suitably be pre-pared in the above manner. A preferred range of molecular weights is from about 200,000 to about 500,000, since it has been found that membranes fabricated in accordance with the present invention from polyether-poly-carbonate block copolymers having molecular weights within such preferred range exhibit ultrafiltration rates comparable to those of Cuprophan membranes and hence within the range clinically acceptable for use in hemodialysis.
Suitable casting solutions for use in producing membranes of the present invention may be prepared by dissolving the polyether-poly-carbonate block copolymer in a water-miscible organic solvent for the copolymer. '~he solvent preferably has a boiling point within the range of 50 to 85C for optimum room temperature casting. The preferred solvent is 1,3-dioxolane which has the appropriate combination of high solvent power for the copolymer, suitable vapor pressure at 25C, water-miscibility and a boiling point of 75 to 76C. Other suitable solvents which can be employed include 1,3-dioxan, 1,4-dioxan, tetrahydrofuran, butyrolactone, acetonitrile, cellosolve acetate, dimethylformamide, pyridine and mixtures thereof.

~7467 Chloroform, which was previously suggested for use as a casting solvent in the methanol~gelation of polycarbonate membranes, is not suitable since it is not water-miscible.
The casting solutions are generally formulated to have a total solids content of from about 1 to about 20 weight % to give dopes ranging in viscosity from about 5,000 to about 30,000 cps. Typically, solids contents range from about 10 to about 20 weight % to give viscosities of from about 7,000 to about 25,000 cps, the preferred range. A swelling agent, such as dimethyl sulfoxide, is advantageously added to the casting solution in amounts ranging from about 10 to about 75% by weight of the copolymer, the preferred range being from about 15 to about 25% by weight of the copolymer. The addition of the swelling agent has been found to enhance the permeability of the resulting membrane. Other swelling agents which have been employed include dimethylformamide, dimethylacetam~de, acetamide, formamide and pyridine.
Production of the polycarbonate membrane can be effected on a continuous basis by doctor blade casting of the casting solution onto a moving surface having a smooth finlsh, such as a coated release paper.
The well filtered (10~ m) casting solution is preferably supplied to a hopper placed in front of the doctor blade by means of a positive displacement metering pump. The hopper is provided with end guides for controlling the width of the membrane sheet. The thickness of the membrane sheet is controlled by adjusting the gap between the knife and the moving belt suriace, which is usually set so as to give a inal membrane thickness o-f 1.0-1.5 mils.
The freshly cast and wet film is allowed to air dry at tempera-tures ranging from about 20 to about 30C for periods ranging from about 1.0 to about 5.0 tninutes to partially evaporate the solvent therefrom, the drying time being determined by both the belt speed and the drying distance.

~r The pa ally driedfilm i~ gelled to producc the final membr~ne by immcrslo while still adhering to the moving belt, in a water bath. The gelation bath temperature may be varied between about zero to about 40 C, the preferred range being 20 to 30~C. After gelation, the membrane is peeled from the moving belt and rolled up separately from the belt onto a cylindrical core.
The membrane is finally washed thoroughly with deionized water to remove the last traces of solvent and swelling agent and stored in a sealed plastic bag or other container containing water and a sterilant such as formaldehyde.
The final thickness of the membrane generally varied from about 1. 0 to 1. S
mils, depending upon the knife gap setting, casting solution visco~ity and belt æpeed.
¦ The following examples are given for the purpose of illustrating the present invention.

l Example l 15 1 A mixture of 491 gm of the polyether-polycarbonate block copolymer ¦ obtained by reacting phosgene with a comonomer mixture of bisphenol A
(75 wt %) and Carbowax 6000 (25 wt %), and having an intrinsic viscosity of 1. 7 (in chloroform at 25C) corresponding to a molecular weight of 377, 000 1 3146 gm of 1, 3-dioxolane and 98. 2 gm of dimethyl suifoxide, was slowly 20 l agitated until solution was effected (approximately 8 hours). The crude solution was filtered in a pressure filter at 30 to S0 psig through a polypropylene felt or 25~ m porosity asbestos sheet medium to remove a small residue of fine insoluble matter. The resulting casting solution has a viscosity of 16, 000 cps at 25 C.

.. . . . .

~ 7467 . . .
. , , .
~pproximately one-half gallon of the above 901ution was then filtered through a lOf~ m inline filter and cast via a doctor blade onto the surface of a 16~incll wide bclt moving at a 6peed of 2. 36 feet per minute. The hopper end guides were set to provide a cast film width of 15-1/2 inches and the gap between the doctor knife and the moving belt ~urface was set at 7. 0 mils.
These dimensions provide samples suitable for use in the Kiil dialyzer. A
total of 2. 54 minutes drying time was allowed before gelation of the cast film in a water bath. The ambient air temperature was maintained at 24. 7 +
0. 4C and the gelation water bath temperature at 25 ~ 0. 5 C. After gelation, the resulting mernbrane was peeled from the moving belt and rolled up separately from the belt o~lto a cylindrical core, A total of 177 feet of membrane was thus produced during a period of 75 minutes. The membrane was washed in a flowing streamof deionizedwater and stored in a 6ealed polyethylene bag containing 2% aqueous formalydehyde.
The polycarbonate membrane fabricated as above was found to have physical and permeability properties as set forth in Table 1, below. For purposes of comparison, corresponding values are given for a typical ~ample of Cuprophan PT150 membrane. The permeability properties were ¦ determined in a dialysis test cell of the type designed by the National Bureau of Standards.

, ~;)7467 ¦ Tablc 1 .
¦ Polycarbonate McmL>rane of Cupropllan PT
xample 1 150 Membranc IWet Thickness, mils 1. 3 0. 9 ¦Relative Burst Strength, Cm Hg. 30 20 ¦Ultrafiltration Rate at 37 C, .
¦ 200 mm Hg ,~ P, ml/m2 -hr-mm Hg3. 6 3. 9 . .
~ irîusive permeability, at ¦ 37C, cm/min (x 104) (Solute ¦ molecular weight in parenthesis) Sodium chloride (58.4) 709 707 Vitamin B12 (1355) lOl 46 ¦ Human Serum .
Albumin (60, 000) 0 0 ~ .,~
¦ It can be seen from the data in Table 1 that the polycarbonate membrane fabricated in accordance with the present invention, with approximately 40%
greater thickness than the Cuprophan membrane, and approximately the same ultra~iltration rate and permeability towards sodium chloride, a representa-tive low molecular weight solute in blood, exhibits a 50% higher burst I
I . .

! ~

.

~7467 strength and a 120% higher permeability toward Vitamin B12, a model medium molecular weight solute, while being completely impermeable to serum albumin, a high molecular weight component of blood whose removal from the blood during hemodialysis is not desirable.
It has further been found that the polycarbonate membrane prepared in accordance with the present invention is considerably stiffer in its wet state than Cuprophan membranes. This is of importance in hemodialysis in maintaining a thin blood film, a greater area of blood ; for dialysis, and a low blood priming volume. Also, the polycarbonate membrane of the present invention is heat sealable, making possible greater latitude in hemodialyzer design. Furthermore, the polycarbonate membrane of the present invention has proven to be non-toxic in a battery of in vitro and animal tests, is blood compatible, and its thrombogenicity is approximately the same as Cuprophan membranes in vitro.
Examination of the polycarbonate membrane prepared in accordance with Example 1, employing water as a gelation medium, by scanning electron photomicrography showed the side of the membrane which was facing the air during drying to be smoother and more regular than the side of the mem-brane which was ln contact with the casting surface, indicating that the membrane was formed with its barrier or active layer on the side of the membrane facing the air during drying rather than on the side of the membrane in contact with the casting surface as was the case with methanol-gelled polycarbonate membranes. Hence, the continuous peeling of the membrane from the moving belt surface has no deleterious effect on the delicate barrier layer of the membrane, making large scale machine - ., - . ~
- .

- : .. .:
:: : :' . ' :
.. . . .
. ' ' ' :

,.
production of the membrane feasible. The water-gelled polycarbonate membrane prepared in accordance with Example 1 alsc appeared to have a much finer and more uniform ultragel structure than a simi]ar mem-brane prepared by methanol gelation. This is reflected in the consider-ably higher strength of the water-gelled polycarbonate membranes, which were found to have burst strength 50 to 70% greater than the correspond-ing methanol-gelled polycarbonate membrane.
Example 2 This example shows the efficacy of swellLng agent added to the casting solution formulation in enhancing the water and solute permeability of polycarbonate membranes prepared according to the present invention.
Gelled membranes were cast under identical conditions from casting formulations containing a polyether-polycarbonate block copolymer obtained by reacting phosgene with a con-onomer mixture of bisphenol A (75 wt %) and Carbowax 6000 (25 wt %) and having an intrinsic viscosity of 1.3 (in chloroform at 25 C), correspondirlg to a mol. wt of 190,000. The casting solution formulations contained varying amounts of the swelling agent dimethyl sulfoxide (DMS0). The properties of the resultant polycarbonate membranes as a function of the amount of DMSO swelling agent in the casting formulation are summarized in Table 2. Corresponding values for a typical sample of Cuprophan PT-150 are given for comparison.

, :

- ~Lq3 7~s67 a~ ~ ~ oaoc~ o o o ~. ~

C~

P~ r. ~ . ~ ~ ~ ~r ~ ~ ~ o ~ . ~ o J ~ ~ U~ o~ ~ ~
U~
al ~
a~0 ,1 o 0~ ~ ~1 1` ~ u~
~o o S o o ~ ~ O
o o ,~ J ,~
.~ .

,, o t, ~
o O OC~ r U~ ~
o 3 o~) ~ o s~ ~ J o~ ^ ~ ô
~1 ~ 1~1 ~ V~ ~ ~ J
~ n ~ 13 ~ .
Y; -1 0 c ~d I ~rl ~ 00 R
!~! s~ .,~ ~ ~ o ~ ~ ~ e ~ ~ R :~: i~! ~ t~ 1~ ~ ~ `J R
X ~ o R ~! ~a) ~ . ~
o 'JJ ~n ~ rl I f~ ~ 0 _~¢
u~ :~ to R ~ O ~ ~ ~ ~ O
,~ ~J cJ ~a~ P~ ~1 0 o ~ ~ o Ei o ~ R ~ ~ o V~ 0'- R ~ Z
V O ~ ~ ~ ~ O
O bO ~ ~ ~ rl ~E~ ¢ O R u~
u~ rl ~ h [~ ~ ~ 5 ~ ~ rl R
e~ ~ ~ ~0 5, ~ u 0~ ~q ~ ~1 0~ ~ X ~) ' a) ~ ~d ~ t3 c~ ~ ~z 5 ~ :~ ~ u~ Pl ,,7'' , ,:
;
: : : .
::

~ 7467 The data of Table 2 clearly show the marked effect of adding DMS0 to the casting solution on the degree of membrane swelling, as measured by membrane wet thickness and water content, with resultant enhancement of membrane permeability to water and a variety of solutes. The polycarbonate membrane prepared using the casting for~ulation containing no swelling agent exhibited permeability properties comparable to those of a typical Cuprophan PT150 membrane. Addition of the first increment of DMS0 swelling agent (2 grams per 15 grams of polymer) to the casting formula-tion is seen to have nearly doubled the water content and tripledthe hydraulic permeability (as measured by ultrafiltration rate) of the membrane, and increased the permeability to all the solutes tested. The degree of permeability enhancement increased with solute molecular size, with 24-37% higher values observed with the smaller solutes, such as urea and creatinine, and a very marked increase of 160% found for inulin, a model solute representative of the upper "middle molecule" range. Further increase in the level of swelling agent in the casting formulation (to 4 grams per 15 grams of polymer) is seen to have still further increased the polycarbonate membrane water content and water permeability, only slightly (2-7%) increased smaller solute permeability (i.e. sodium chloride, urea, creatinine and uric acid while still resulting in a substantial increase in "middle molecu]e" permeability (22, 24 and 69% increase for phosphate, raffinose and inulin respectively). Significantly, the poly-carbonate membranes completely reject albumin even when substantial amounts of swelling agent are added to the casting formulation.

~' , . . ..

.
.
. ' . ' ' ' ' :
, ' : .

11~7467 Example 3 This example serves to illustrate the effectiveness of several cosolvents - swelling agents for enhancing polycarbonate membrane permeability when added to the membrane casting solution formulation.
Casting solutions were prepared from the following formulation, using a polyether-polycarbonate block copolymer obtained by reacting phosgene with a comonomer mixture of bis-phenol A (75 wt %) and Carbowax 6000 (25 wt %) and having an intrinsic viscosity (in chloroform at 25C) of 1.52 corresponding to a molecular weight of 301,000.
COMPONENT WEIGHT - GRAMS

Polyether-Polycarbonate 40.0 Block Copolymer 1,3-Dioxolane 256.2 Swelling Agent 8.0 Membranes were prepared from each formulation by hand casting under identical conditions on glass plates at room temperature and gelling in water at 25C after varying drying periods. The physical and permeability properties found for these membranes are shown in Table 3.

.
.

::

, ':, ' ~7~67 a) ~1 ~o g ~ __ U~ ~`J
,,~ ~ o U~ ~ X
a~ ,i ~~ a~~ ~ ~
V~ ~ ~ ~ ~

~0~
~ ~1 ~ o~

'J~ 0~ 0~ ~
, . . .. ~ a) ,~~ ~a ~
t~) h ~1 [~ ~ ~ :~Z
E~
o it 0~ ~D
o~) . . . . . .
~1 0~ ~ ' P~
o~
U
~ V) ^
o ta ~1 ,~
g ¢ ~ U~ o U~
~ O E~ '' O
U~ P ~

- O --~7~67 The data outlined in Table 3 indicates that, after appropriate adjustment of drying time before gelation, polycarbonate membranes of equi-valent strength and permeability characteristics can be prepared through formulations with any one of the three swelling agents, pyridine, dimethyl formamide and dimethyl sulfoxide.
Example 4 Several batches of polyether-polycarbonate block copolymer were prepared by reacting phosgene with a comonomer mixture of bisphenol A (75 wt%) and Carbowax 6,000 (25wt%). These batches were then blended into a single - 10 master batch. The resultant blend had an intrinsic vlscosity of 1.7 (in chloroform at 25C) which corresponds to a molecular weight of 377,000. Each batch of polymer was formulated and cast into membrane in the same manner as Example 1, producing several samples of membrane each approximately 300 to 1,000 ft. long. The thickness, strength and permeability properties of these membranes are listed in Table 4 below:
Table 4 Sample Thickness Burst U.F. Rat~ P NaCl P Urea P B 2 No. (0.001") (cm Hg) (ml/hr/M / (cm/min (cm/2in (cm~min mm/Hg) x104) xlO ) x104) M-14-38-1 1.24 + 0.05 35.3 + 2.2 4.03 ~ 0.04 713 - 1 796 98.9 + 0.3 M-14~38-2 1.25 - 0.05 34.8 + 1.6 4.15 - 0.10 698 + 15 713 98.4 - 0.2 M-14-46-1 1.28 + 0.04 35.4 + 1.7 4.18 + 0.54 687 + 8 740 - 15 92.5 + 0.9 M-14-46-2 1.31 ~ 0.03 34.8 + 1.4 4.47 + 0.25 673 + 7 731 - 6 92.7 - 0.7 M-14-54-1 1.38 +- 0.07 32.1 +- 1.5 5.23 + 0.29 648 +- 19 735 +- 16 93.4 +- 2.2 M-14-54-2 1.39 ~ 0.06 32.8 - 1.6 5.14 - 0.39 656 - 27 724 - 5 93.1 - 1.9 M-14-54-3 1.40 - 0.05 34.0 - 1.5 4.74 - 0.01 664 + 20 716 - 13 93.2 - 1.7 M-14-65-1 1.18 - 0.05 33.5 + 1.7 4.68 + 0.43 718 + 0 754 + 35 100 + 1 M-14-65-2 1.19 - 0.05 33.2 - 1.5 5.27 - 0.17 714 - 4 742 - 23 106 - 5 M-14-73-1 1.35 - 0.06 36.5 + 2.4 4.54 + 0.16 637 - 3 -- 91.4 - 0.02 M-14-73-2 1.36 - 0.06 36.4 - 2.1 4.51 - O.L9 655 - 21 -- 91.4 - 0.02 M-14-80-B 1.17 - 0.06 37.0 - 1.9 3.43 742 813 99.3 M-14-80-E 1.17 - 0.05 35.7 - 0.9 3.68 746 795 99.3 M-14-86-B 1.28 + 0.03 33.6 + 1.4 4.22 723 -- 96.1 M-14-86-E 1.27 - 0.07 33.5 - 1.3 4.66 684 -- 94.2 ~ - 21 -- . ,' ~ ' ' . -' ' " ' ': . ' ' : : , , .

~74~7 Toxicological evaluation of these samples revealed that the membrane was non toxic in all implantations, extractions, and animal tests, non toxic in tissue culture, non toxic in all blood tests and showed no absorption of protein. Subsequent evaluation on patients failed to reveal any toxicity.
Table 5 compares the clotting times of polycarbonate membranes of the invention with that of Cuprophan membranes.

Table 5 Clotting Time Lot No. % of Cuprophan Membrane M-14-16 120.3 M-21-21 118.0 M-14-54 107.8 M-14-46 136.6 M-14-65 109.5 __ This comparison, using the Lindholm test on the above or similarly prepared membranes demonstrated that the polycarbonate membrane is up to 36.6% more compatible with blood than is Cuprophan.
Example 5 Polycarbonate membranes were prepared in the same manner as in Example 4.
Study 1 - The permeability of these polycarbonate membranes toward several midd]e molecules was determined at three different testing facilities.
These values are compared in Table 6 with those for Cuprophan membrane, obtained by those facilities using the same procedures an~ equipment.
The data obtained at Facility 1 is also plotted in Figure :L. The results of these comparisons, show that the polycarbonate membranes of the invention have a consistently superior permeability for middle molecules than does Cuprophan membrane while maintaining permeabilities comparable to Cuprophan membranes for low molecular weight molecules.

,~"

311~7467 Table 6 . __ MEMBRANE PERMEABILITIES - TEST CELLS cm/min (X 104) Facility I Facility II Facility III
PCM* CM** PCM CM PCM CM
UREA (60) 665 654 667 629 696 518 CREATININE (113) 389 370 423 351 422 319 PHOSPHATE (140) 210 184 206 167 URIC ACID (168) 355 264 338 188 SUCROSE (342)201 129 182 135 185 103 ~AFFINOSE (504) 156 97 141 76 VITAMIN B12 (1355) 92 30 95 42 108 28 BACITRACIN (1410) 50 17 INULIN (5200)21 5 23 7 23 6.6 BSP (838) 230 47 (at 27 c) UF RATE 2 4'3 3'5 4.4 3.6 2.9 2.5 (ml/hr/ /mm Hg_ * POLYCARBONATE MEMBRANE
~* CUPROPHAN MEMBRANE

Study 2 - Clearances obtained with these polycarbonate membranes, using D-4 Kiil dialyzers, were obtained at four testing facilities. Table 7 lists the results of these evaluations and compares the clearance of the polycarbonate membrane with Cuprophan membrane clearance data obtained at the same facilities using the same equipment.

.

Table 7 MEMBRANE CLEARANCE - D-4, ml/min Facility 1 Facility 2 Facility 3 Facility 4 PCM CM PCM CM PCM CM PCM CM

UREA (60) 120.0124.0121.0 122.0 117.8* 120.8* 100.00*

CREATININE 91.9 86.0 102.0 103.0 96.8* 106.2* 99.0* 94 PHENOBARBITOL 94.0 77 SUCROSE (342) 67.0 60.0 RAFFINOSE 43.2* 44.8** 46.1 27.7 (504) BSP (838) 117.0 43.0 (1335) 1 43.022.0 39.0 21.0 46.5** 26.0** 41.5 20.6 (1410) 30.3 15.6 INULIN 7.0 4.4 10.1 . . . .. ...
CM = Cuprophan Membrane * Ln Vivo PCM = Polycarbonate Membrane ** At QF = 5 ml/min . .. ___ .
Except for the unexplained difference on raffinose clearance obtained at Facility 3, the evaluations indica~e that the polycarbonate membrane exhibits a consistently superior clearance for middle molecules while maintaining the clearance of low molecular weight species, such as urea and creatinine, at approximately the same level exhibited by Cuprophan membranes. The unexpected high clearance as well as permeability of bromo sulfo phthalein (BSP) is explained by the rapid absorption of BSP
by the polycarbonate membrane.
Example 6 Polycarbonate Membranes were prepared in the same manner as Example 4 and a clinical testing program instituted.
Study 1 - Using D-4 Kiil dialyzers twenty-five hemodialysis ~reat-ments in 10 patients were performed, without any patient complications necessitating special treatment or hospitalization. The patients were unable to describe any difference in their symptoms during therapy from those experienced during therapy with other dialyzers, using Cuprophan membrane or hollow fiber cellulose acetate membrane. Blood flow during dialysis varied between 102 and 250 ml/min. Clearance of BUN, creatinine, uric acid, and phosphorus lncreased as blood flow increased, within the limits of flow observed in this study. Arterial pre-dialysis and post-dialysis blood samples showed that hematocrit increased an average of 1.2%
and white blood counts decreased an average of 950 cells/cm (p ~0.001).
Platelets did not change significantly. There were no pyrogenic reactions during the study. Ultrafiltration rates varied between 1.5 and 6.7 ml/hr/
mm Hg pressure and averaged 4.23 - 0.14 ml/hr/mm/Hg.
Study 2 - A group of six patients who had been on maintenance hemo-dialysis therapy between nine and sixty months was selected to undergo a double-blind evaluation of polycarbonate membrane. Each patient was clini-cally stable and had been treated with a variety of hemodialyzers prior to ; 20 entering the study. There were three adult females and three adult males, aged twenty-two to fifty-two years. Each patient was treated for five hours three times weekly with a D4 Kiil dialyzer. The patients, nurses and physicians were unaware of the type of membrane being used during therapy.
Three patients were randomly assigned to begin therapy with Cuprophan membrane; the other three were started with the polycarbonate membrane.
Each patient was treated for three months with either Cuprophan or poly-carbonate membrane and then switched to the other membrane. During the six months of therapy, the only episode requiring hospitalization was one ~ 25 -, .` ~. . ':

.' :.

~ 7467 :
patient hospitalized for three days with bronchitis. She was undergoing therapy with polycarbonate membrane when hospitalized. The patient's general sense of well-being did not change when the two three-month periods were compared.
Pre-dialysis supine arterial pressure was 122/78 during poly-carbonate period but 150/96 during Cuprophan membrane therapy (p ~.001) in one patient: the others showed no significant change in arterial pressure during each three month period. Pre-dialysis weights were higher during Cuprophan membrane therapy in four patients; two were lower. Two patients showed a small but signficant increase in pre-dialysis hematocrit while on polycarbonate membrane, as shown in Figure 2. Two others had a significantly higher platelet count while on polycarbonate membrane therapy, while two had significantly lower platelet count on polycarbonate therapy. Three patients had higher WBC counts on polycarbonate therapy while two had lower WBC
count on polycarbonate therapy. Predialysis serum creatinine was signifi-- cantly lower in one patient while receiving therapy using Cuprophan membrane.
(Mean creatinine 13 PCM, 11 mg/dl Cuprophan). Serum uric acid was higher in two patients while on polycarbonate membrane therapy and in one patient while on Cuprophan membrane. BUN was lower on Cuprophan membrane in one patient, the same one with the lower creatinine. Phosphate was lower on Cuprophan membrane in three patients and lower on polycarbonate membrane in one. Residual renal function remained unchanged in these patients during the six month period of observation.
The clinical condition and laboratory studies did not indicate any deleterious change when patients were treated with polycarbonate membrane as compared to Cuprophan membrane therapy in this six month study. Although - individual changes were seen within a given patient, the group of six patients ~7467 did not consistently show changes either better or worse, in any direction.
However, the increased hematocrit observed in two patients and decrease of blood pressure in one patient was highly significant and demonstrates important advantages obtainable using polycarbonate membranes.
Study 3 - A clinical evaluation was instituted with the major purposes being to try to reduce patient time on dialysis from the present average of 24 hrs./wk., and still maintain, by present standards, adequate dialysis for the individual patient, and, by more adequate removal of middle molecular weight toxins, to try to reduce some of the remaining complications of chronic dialysis.
The DI (MM) or Dialysis Index for middle molecular weight solutes for each individual was determined. This DI (MM) takes into consideration the G.F.R. or the remaining kidney function of the individual patient and the mass area of the individual patient based on height and weight as com-pared to the minimum weekly volume of middle molecules (with Vitamin B12 as a marker) which must be removed from an average sized patient (1.73M2) to prevent uremic symptoms.
The minimum weekly volume of middle molecules which must be removed DI (MM) was determined by a retrospective study of four years of accumulated clinical cases where all the known variables were measured relative to the minimum DI (MM) sufficient to prevent the development of motor nerve conduction velocity (MNCV) reduction for the individual pati~nt:

Calculated weekly amount of MM removed DI (MM) Minimum weekly amount of MM to be removed , - 27 -.
'.

11~7467 A DI (MM) of 1 is adequate dialysis,~ is overdialysis, and C 1 is underdialysis.
One of the most sensitive indicators of inadquate dialysis is peripheral neuropathy since it develops and progresses in a seemingly well-dialyzed patient. The first indicator of peripheral neuropathy is "Motor nerveconduction velocity" (MNCV) reduction. Having Cuprophan membrane, two patients entered a control phase in which the dialysis schedule for the individual patient produced a DI (MM) equal to or greater than 1. This was followed by an induction phase in which the DI (MM) of both patients was lowered to below 0.7 by shortening time on dialysis with the conventional Cuprophan membrane. During this phase, the small molecule concentrations were elevated by about 20%, while middle molecules were elevated as much as 100%. This phase was continued until under dialysis was evident by the appearance of peripheral neuropathy.
The _ecovery ~hase was then instituted wherein Cuprophan membrane wasreplaced by polycarbonate membrane with the same reduced time schedule.
Both patients stabilized and then showed improved MNCV after several weeks on polycarbonate membrane treatment. In one patient there was a direct measured confirmation of a lowering of middle molecule levels.
Based on MNCV and EEG and using the Babb-Scribner Charts for Estimating Minimum Adequate Dialysis Times for Patients in Terms of Body Size, GFR and Various Membrane/Dialyzer Combinations, projections were made of the minimum adequate dialysis time for an average body mass man (surface area 1.7M ) with no kidney function or with no Residual Glomerular Filtration Rate (G.F.R. = O) and partial kidney function (G.F.R. = 1).
The minimum required time is less than 2/3 of the time required using Cuprophan membrane (Table 8).

~- ~ - 28 -~lQ7467 ., .

. . . ~ ~
Table 8 Projections of Minimum Adequate Dialysis Time (Based on Creatinine and B12 Clearances) (For average Body Mass patient of 5'7", 143 lbs.) DIALYZER G.F.R. = 0 G.F.R = l D-4 Kiil Polycarbonate - 1* 12.4 hrsiwk. 10.5 hrs/wk Gambro Cuprophan (1 ~ ) - lM2 27.3 " " 18.0 " "

Gambro Cuprophan (13.5~u) - lM2 18.6 " " 12.3 " 1~

Travenol Cuprophan - lM2 18.6 " " 12.3 " "

Dow 4 - Hollow Fibre2Cellulose 20 4 " " 13 5 " "

QB = 200 QD = 500 QV = 5 . ~
; Study 4 - Two patients were dialyzed for a six month period using Kiil Dialyzers and Cuprophan membrane and baseline data for these patients as listed in Table 9, were obtained. Treatment was then continued with the Cuprophan membrane replaced by polycarbonate membrane. The results obtained after one month's dialysis using polycarbonate membrane, listed in Table 9 show an improvement in neurobehavorial functions with decreases in both urea nitrogen and serum creatinine as well as an increase in hematocrit.
'~, All these changes, indicative of an improved medical status of the patient undergoing dialysis, is evidence of more adquate removal of middle molecular weight toxins.

.~ , .

., .

~ ~, - 29 -.
.: - -, . ' :
. ' ': ~ .
', ~ ' .

. ~ .
' ~
,: ~, . : .
, ` ~L3L~3174~7 .. __ ..
Table 9 Patient 1 ¦ Patient 2 CM* PCM*** ¦ CM* PCM***

Urea Nitrogen, mg% 85.6 80 95.15 90 Creatinine, mg~ 9.738.7 18.7 16.5 Hematocrit, % 18 19.5 17-18 20-21 Neurophysiological Conditions EEG* % 30 26 13 12 CMT 2-4 2 14.3 15 CPT (Continuous Performance Test) 52.3 55 53.3 54 CRT .499.443 .492 .447 *EEG is expressed as percent of 3 to 7 Hz divided by 3 to 13 Hz.
¦**Cuprophane Membrane; after 6 months treatment ¦***Polycarbonate Membrane; results are after one month's tr~atment. i ~ = =
Use of polycarbonate membranes of the invention, enable improved transport of middle molecules without substantially varying the levels of ultrafiltration rate and transport of low molecular weight molecules from the desired range, result in improved hematocrit and neurobehavorial functions in patients and allow reduced dialysis times all without toxic reactions or other detrimental effects on the patient. In addition, these polycarbonate membranes are more blood compatible and significantly stronger than Cuprophan membranes.
For further reference, compare the six (6) following reports to the National Institutes of Health:

'C

~7~7 (1) Modified Polycarbonate Membranes for Hemodialysis. National Institute of Scienti~ic Research, Rancho Santa Fe, California.
Ann. Rept. 1 July 70-31 Dec. 71. PB-213 160/6. This document wa8 received in NTIS (National Technical Information Service) in January, 1973, and was announced in the bi-weekly journal, GRA, Number 2, dated January 25, 1973.
(2) Modified Polycarbonate Membranes for Hemodialysis. National Institute of Scientific Research, Rancho Santa Fe, California.
Ann. Rept. 1 Jan.-31 Dec. 72. PB-225 043/9. This document was received in NTIS in January, 1974, and was announced in the bi-weekly journal, GRA, Number 3, dated February ~, 1974.
(3) Modified Polycarbonate Membranes for Hemodialysis. National Institute of Scientific Research, Rancho Santa Fe, California.
Rept. 15 Jun.-20 Sept. 69. PB-225 135/3. This document was received in NTIS in December, 1973, and was announced in the bi-weekly journal, GRA, Nu~er 2, dated January 25, 1974.
(4) Modified Polycarbonate Membranes for Hemodialysis. National Institute of Scientific Research, Rancho Santa Fe, California.
Ann. Rept. 1 Aug. 73-31 Mar. 74. PB-233 669/1. This document was received in NTIS in August, 1974, and was announced in the bi-weekly journal, GRA, Number 18, dated September 6, 1974.
(5) Modified Polycarbonate Membranes for Hemodialysis. National Institute of Scientific Research, Rancho Santa Fe, California.
National Institute of Arthritis and Metabolic Diseases, Bethesda, Maryland. Ann. Rept. 1 Jan.-31 Jul. 73. PB-235 792/9SL. This document was received iTI NTIS in October, ]974, and was announced in the bi-weekly journal, GRA, number 24, dated November 29, 1974.

(6) Modified Polycarbonate Membrane for Hemodialysis. National.
Institute of Scientific Research, Rancho Santa Fe, California.
Final Report March 31, 1974 - June 30, 1975. Submitted to the National Institute of Arthritis, Metabolism and Digestive Diseases, National Institutes of Health September 1975.

.

Claims (11)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY
OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A membrane of hydrophilic polycarbonate polymer adaptable for use in a hemodialysis apparatus as a means for preferential removing of middle molecular weight molecules from blood, said membrane has a diffusive permeability measured at 37°C to sodium chloride of about 630 to 825 cm/min x 10-4, a permeability to urea of about 650 to 850 cm/min x 10-4, a permeability to vitamin B12 of greater than 90 cm/min x 10-4 and an ultrafiltration rate of less than 5.5 ml/hr/M2 /mm Hg.

2. A membrane as in claim 1 wherein the polycarbonate polymer consists of a polyether-polycarbonate block copolymer containing 5 to 35%
by weight of repeating alkylene ether carbonate units and from 95 to about 65% by weight of repeating bisphenol A - carbonate units.

3. A membrane as in claim 2 wherein the polymer has a molecular weight within the range of from 50,000 to about 750,000 as determined by intrinsic viscosity measurement.

4. A membrane as in claim 2 wherein the polymer has a molecular weight within the range of from about 200,000 to 500,000 as determined by intrinsic viscosity measurement.

5. A membrane as in claim 1 wherein the membrane has a diffusive permeability measured at 37°C to sodium chloride of about 630 to 750 cm/min x 10-4, a permeability to urea of about 665 to 815 cm/min x 10-4, a permeability to vitamin B12 of about 90 to 110 cm/min x 10-4 and an ultra-filtration rate of about 2.9 to 5.5 ml/hr/M2/mm Hg.

6. A membrane as in claim 5 wherein the membrane has a thickness of about 0.00098 to 0.00145 inches.

7. A membrane of hydrophilic polycarbonate polymer as in claim 1, consisting essentially of a polymer having recurring units of the formula wherein A is selected from the group consisting of -CH2 CH20- and a combination of , and , wherein x is from about 10 to 155 and a and b are such that the bisphenol A
carbonate unit is about 95 to 65% of the weight of this recurring unit and the alkylene ether carbonate unit is about 5 to 35% of the weight of the recurring unit.

8. The membrane as in claim 7 wherein A is -CH2-CH2-0-.

9. The membrane of claim 8 wherein the polymer has a molecular weight of 50,000 to 750,000 as determined by intrinsic viscosity measure-ment.

10. The membrane of claim 8 wherein the polymer has a molecular weight of 200,000 to 500,000 as determined by intrinsic viscosity measure-ment.

11. The membrane of claim 8 wherein the value of x is about 152, b is 1 and a is about 80.