Comparison of Two Filter Combinations for Low-DensityLipoprotein Apheresis by Membrane Differential Filtration:

A Prospective Crossover Controlled Clinical Study

*Goran Matic, †Peter Kohlschein, ‡Uwe Wallstab, §Michael Tiess, §Roland Winkler,§Heinrich Prophet, §Wolfgang Ramlow, and †Peter Schuff-Werner

*Labor Müller; †Institute for Clinical Chemistry and Pathobiochemistry, University of Rostock, Rostock; ‡IngenieurbueroUwe H. Wallstab, Overath; and §Dialysegemeinschaft Nord e.V., Rostock, Germany

Abstract: Membrane differential filtration is an acceptedprocedure for the extracorporeal removal of low-densitylipoprotein (LDL). Reduction rates largely depend on thenature of the membranes and are ideally evaluated in acrossover study design. Four patients who had beentreated by LDL apheresis for at least 6 months were in-cluded. Six consecutive weekly sessions (40 ml plasma/kgbody weight) were scheduled per system (PlasmacurePS06/Evaflux Eval 5A [Kuraray] versus PlasmafloOP05W/Cascadeflo AC1770 [Asahi]). Laboratory mea-surements indicated reductions of plasma concentrationsfor fibrinogen (37% [Kuraray] versus 44% [Asahi]), IgG

(15% versus 20%), IgA (24% versus 28%), IgM (63%versus 53%), and total protein (11% versus 16%). Totalcholesterol was eliminated by 52% versus 49%, LDL by67% versus 66%, triglycerides by 56% versus 41%, andhigh-density lipoprotein by 10% versus 20%. Three thera-pies employing the Asahi filter combination were termi-nated prematurely due to saturation of the plasma frac-tionator. In conclusion, despite similar physical properties,the membranes differ significantly concerning selectivityand sensitivity to saturation. Key Words: Hypercholes-terolemia—Cascade filtration—Double filtration—Low-density lipoprotein—Elimination plasmapheresis.

Hypercholesterolemia and in particular elevationsof low-density lipoprotein (LDL) cholesterol are ac-knowledged as major risk factors for atheroscleroticvessel disease (1). Large clinical intervention studiesestablished primary as well as secondary preventionof coronary artery disease and/or myocardial infarc-tion by powerful drugs (statins) inhibiting liver cho-lesterol synthesis and thus lowering serum choles-terol levels (2,3).

Cholesterol lowering drugs represent a sufficienttreatment modality in the majority of patients. Somepatients, however, either do not tolerate the sub-stances or, because they suffering from, e.g., familialforms of hypercholesterolemia, do not respond ad-equately to medication. First clinical trials in this re-gard showed the possibility of LDL reduction by

plasma exchange (4), and ongoing innovation led tothe introduction of a number of medical devices al-lowing selective extracorporeal LDL elimination.

Today familial hypercholesterolemia poses aclearcut indication for LDL apheresis, and plasmaexchange has been replaced widely by plasma-sparing, selective procedures (3,5,6). Since the em-ployed equipment is easy to handle and the treat-ments are well tolerated, the list of indications forLDL apheresis has been expanded meanwhile andalso includes secondary prevention in hypercholes-terolemic patients suffering from arteriosclerosiswho do not respond adequately to diet and drugtherapy (7–9). Although well documented eligibilitycriteria exist, the total number of patients on extra-corporeal LDL lowering therapy fluctuates fromcountry to country depending more on the reim-bursement situation than on feasability (10,11).

In membrane double filtration (MDF, also calledcascade filtration or double filtration), plasma de-rived from a plasma filter with large pore sizes is

Received April 2001; revised July 2001.Address correspondence and reprint requests to Dr. Wolfgang

Ramlow, Dialysegemeinschaft Nord, Nobelstr. 53, 18059 Rostock,Germany. E-mail: ramlow@dialysis-north.de

Artificial Organs26(4):371–377, Blackwell Publishing, Inc.© 2002 International Society for Artificial Organs

371

passed through a secondary filter holding back largerproteins, protein complexes, and larger lipoproteins(12–18). Usually, no specific interaction takes place,and the differences in molecular weights betweenLDL and the majority of plasma proteins leads to acertain degree of selectivitiy (12,16–18). The suitabil-ity of the heterogeneous systems depends on the na-ture of the membranes which often is underesti-mated in discussions on clinical effectiveness.

This prospective cross over controlled clinicalstudy was designed to compare quantitative LDLelimination, selectivity, and the unfavorable effectsof 2 widely available membrane combinations.

PATIENTS AND METHODS

Study protocolThe study was planned and conducted as a pro-

spective, cross over controlled clinical study (Fig. 1)according to the guidelines for good clinical practiceand the declaration of Helsinki including approvalby an independent institutional review board. It wascarried out at the Rostock apheresis center.

Eligibility criteria included hyperlipoproteinemia,indication for LDL apheresis, and informed consent.Exclusion criteria included decompensated heart in-sufficiency, myocardial infarction during a period of3 months before beginning the study, unstable an-gina, and coagulopathy.

The final analysis contained 24 Kuraray sessionsversus 20 Asahi sessions because of shunt thrombo-sis in 1 patient.

PatientsFour patients were included for 12 weekly treat-

ments, respectively (Table 1 and Fig. 2). Measureswere taken to ensure steady state of serum choles-terol concentrations.First, patients had been onweekly LDL apheresis for at least 6 months beforethe beginning of the study. Second, before andthroughout the study they received an individual,constant daily dose of cholesterol synthesis enzymeinhibitors (atorvastatin or simvastatin). Third, pa-tients were advised to keep to their low cholesteroldiet.

By including only patients on long-term apheresistreatment, the typical side effects resulting fromanxiety in inexperienced patients were avoided.

Filter propertiesFilter combinations from 2 manufacturers were

employed: plasma filters Plasmaflo OP-05 and sec-ondary filters Cascadeflo AC1770L (Asahi MedicalCo., Ltd., Tokyo, Japan) and plasma filters Plas-macure PS-06 and secondary filters Evaflux EVAL5A (Kuraray Co., Ltd., Osaka, Japan). Filter prop-erties including protein rejection curves are de-scribed in detail elsewhere (12). Secondary filter hol-low fibers were manufactured from ethylene vinylalcohol copolymer (Evaflux EVAL 5A) versus cel-lulose diacetate (Cascadeflo AC1770L) with an in-ner diameter of 175 versus 220 mm, wall thickness of40 versus 80 mm, pore size of 0.03 versus 0.037 mm,and a total surface area of 2 versus 1.7 m2, respec-tively.

Apheresis treatmentsMembrane filters were filled and rinsed according

to the manufacturers’ instructions. The plasmapher-esis monitor Haemomat-Plasmomat (Haemo-Scan,Medizinische Messtechnik GmbH, Goettingen, Ger-many) was composed of 1 blood and 1 plasma pump,appropriate pressure detectors, and an air detector(Fig. 1). The fractionating filters were operated inthe dead end mode, and plasma was warmed in aheater to 36.2 to 36.8°C before entering the second-ary filter. This procedure has been called thermofil-tration and ensures constantly warm temperatures inthe secondary filter unit leading to decreased cryogelformation and increased selectivity. Further detailscan be found elsewhere (17).

The blood flow was kept at a maximum of 60 ml/min, and the plasma flow rate at approximately one-third of the blood flow. Anticoagulation included abolus dose of 5,000 to 10,000 IU unfractionated hep-arin combined with continuous intravenous infusionof 1,000 to 2,000 IU/h (Table 1).

Transmembrane pressure (TMP) of the plasma

FIG. 1. The flow chart shows the plasmapheresis monitor. It il-lustrates the principle of the dead-end mode with plasma filteredcompletely through the fractionating membrane and thermofiltra-tion, the warming of plasma to >36°C improving selectivity of thesecond filtration step (17).

G. MATIC ET AL.372

Artif Organs, Vol. 26, No. 4, 2002

fractionator was monitored, and therapy was termi-nated at a TMP >400 mm Hg.

Clinical chemistryTotal cholesterol, LDL cholesterol, and high-

density lipoprotein (HDL) cholesterol were mea-sured on a Hitachi 911 clinical chemistry analyzer(HiCo Cholesterin, enzymatic colorimetric test;LDL-C homogeneous, homogeneous turbidimetrictest; HDL-C, enzymatic colorimetric test, Roche Di-agnostics, Mannheim, Germany). IgG, IgM, IgA,and lipoprotein a (Lp[a]) were measured by nephe-lometry on the BNII (Dade Behring, Marburg, Ger-many, N Lp[a] reagent, N antiserum to IgG, IgA,IgM). Fibrinogen was estimated by a modification ofthe Clauss method (BCS, Dade Behring, Marburg,Germany, Multifibren U). Protein and albumin weremeasured on a Vitros 950 clinical chemistry analyzer(Johnson and Johnson, Neckargmuend, Germany;TP Slides and ALB Slides).

Statistical data analysisData analysis was carried out on a personal com-

puter system employing a spreadsheet application(Microsoft Excel). Data in diagrams are displayed asmean ±95% confidence interval of the mean. Statis-tical analysis was performed with the Student’s t-test;p Values are indicated in the figures and the text.

RESULTS

Handling and performance of the membrane filtersThe primary and secondary (cascade) filter com-

binations from Asahi Medical and from Kuraraywere almost identical concerning priming and wash-ing procedures and preparation. No difficulties werenoted with trapped air, quality including leakage, oradverse effects attributable to sterilization.

Processed plasma volume was estimated directlybefore extracorporeal therapy proportionally to thepatient weight (40 ml/kg body weight) (Table 1).Three out of all 24 sessions with the Asahi filter sethad to be terminated prematurely. During those ses-sions, TMPs of the secondary filter AC1770 reachedthe endpoint of 400 mm Hg defined by the studyprotocol at processed plasma volumes of 3,300, 3,400(Patient HF), and 3,500 ml (Patient FW), respec-tively. Therefore, 3 out of 12 therapies aimed at avolume approaching 4,000 ml had to be terminateddue to saturation of the secondary filter AC1770.

Average maximum TMP of the secondary filterswas 125 ± 64 mm Hg for the AC1770 versus 73 ± 41mm Hg for the Eval 5A (mean X ± standard devia-tion, all therapies), indicating a somewhat higherTMP with the AC1770 secondary filter. The rise ofTMP to 400 mm Hg was unpredictable, and neitherrelated to fibrinogen levels nor another measuredlaboratory parameter.

Untoward effectsNo serious adverse events were noted. Dizziness

was noted twice (10%) (Patients HO and FW) andheadache once (5%) (Patient HF) during therapywith the Kuraray filter combination. They may beinterpreted as side effects of extracorporeal circula-tion and were manageable by physical measures.Nausea and dizziness were noted during 1 sessionwith the Asahi filter combination (4.2%). However,the afflicted Patient EJ experienced similar symp-toms in the morning before commencement oftherapy, and they may not be attributable to apher-esis.

FIG. 2. The scheme is of the study protocol illustrating the cross-over design. It also shows the drop out of Patient HF.

TABLE 1. Patients and details of extracorporeal treatment

PatientAge

(years) SexWeight

(kg)Height(cm)

Plasmaa

(ml)Heparin bolus

(IU)Heparin/h

(IU) Diagnosis

HO 70 Male 70.2–74.4 178 2,897 5,000 1,000 PVDHF 60 Male 93.4–95.9 174 3,885b 5,000 1,000 CADFW 63 Male 101.7–101.9 178 4,040b 5,000 1,000 CADEJ 48 Male 69.3–70.8 172 2,913 10,000 2,000 CAD

a Mean treated plasma volume.b Not including prematurely terminated sessions.CAD: coronary artery disease, PVD: peripheral vascular disease.

LDL APHERESIS BY MEMBRANE DIFFERENTIAL FILTRATION 373

Artif Organs, Vol. 26, No. 4, 2002

Drop outsPatient HF did not complete the study. After 6

apheresis sessions on the Asahi Medical filter com-bination and 2 sessions on Kuraray he developed ashunt thrombosis. Shunt complications were knownfrom the patient’s history, and a relationship to theextracorpoeal therapy or a specific medical deviceseems unlikely.

Elimination of lipoproteinsOne major aim of the cross over protocol was to

avoid bias attributable to interindividual variationallowing a correct comparison of in vivo perfor-mance of the 2 tested filter combinations. Patientswere in a steady state regarding their lipid-loweringmedication. They also had been treated by weeklyLDL apheresis for at least 6 months before theywere included in the study.

No significant difference could be found concern-ing the extracorporeal elimination of LDL, total cho-lesterol, and Lp(a). Since only 1 of the tested pa-tients expressed Lp(a) values above the referencevalue of 0.3 g/L, only data from this patient wereanalyzed (Fig. 3).

Differences could be shown regarding the elimi-nation of HDL and triglycerides. The Kuraray filtercombination eliminated significantly less HDL thanthe Asahi filter combination (p < 0.001) (Fig.3).

Elimination of plasma proteinsSecondary filter characteristics depend not only

on pore size but also on porosity, surface area, andthe biomaterial. Sieving coefficients represent the

proportion of proteins that pass through the second-ary filter (Table 2). One would expect a direct rela-tion between diameter of the protein or its molecularweight and pore size of the membrane. With fibrino-gen, the situation is different. In addition to the non-globular shape of the fibrinogen molecule, it forms agel together with heparin and fibronectin that leadsto trapping within the cascade filter and higher re-duction rates than one would expect from its mo-lecular weight (12,17).

No significant differences could be shown in elimi-nation rates for albumin and fibrinogen (Fig. 4). To-tal protein, IgG, and IgA were more readily elimi-nated by the Asahi whereas the Kuraray filtercombination showed higher elimination rates forIgM (Fig. 4). Also, mean values for HDL and IgMbefore the first apheresis were not significantly dif-ferent from the values before the sixth apheresis,respectively, for both filter combinations (data notshown).

DISCUSSION

In MDF, after separation from whole blood,plasma is fractionated by a secondary membranewith pore sizes between 0.013 and 0.037 mm (12).Depending on the properties of the plasma fraction-ator, MDF can be applied for more or less selectiveextracorporeal elimination of IgG (approximately150 kilodaltons) up to the elimination of the largeLDL complex (approximately 2,000 kilodaltons)(12). Since a number of primary (plasma) and sec-ondary filters are available on the market, numerouscombinations and configurations are possible (12).

Adding to the heterogenity of MDF are proce-dural modifications suggested by some authors. Theyinclude pulsatile flow or recirculation within the hol-low fibers of the secondary filter (13,19). All sugges-tions aim at reduction of membrane plugging thuskeeping sieving factors constant throughout thetherapy and allowing larger plasma volumes to be

FIG. 3. Shown are elimination rates for total cholesterol, triglyc-erides, and lipid-containing plasma constituents. Data were de-rived from measurements directly before and after LDL apher-esis. Only 1 of the patients expressed Lp(a) levels >0.3g/L, andonly his data were included into data analysis (p values fromStudent’s t-test; not significant, p $ 0.05) (LDL: low-density lipo-protein, Lp[a]: lipoprotein a, HDL: high-density lipoprotein).

TABLE 2. Sieving coefficients of the plasmafractionators (data from the manufacturer)

Sieving coefficients

CascadefloAC1770

EvafluxEVAL 5A

Albumin 0.9 0.9IgG 0.7 0.83IgM 0.3 0.1Fibrinogen NA 0.25LDL cholesterol NA 0.02HDL cholesterol NA 0.82

Ig: immunoglobulin, LDL: low-density lipoprotein, HDL: high-density lipoprotein, NA: not available.

G. MATIC ET AL.374

Artif Organs, Vol. 26, No. 4, 2002

treated. It appears, however, that the procedure ap-plied in the present study and schematically demon-strated in Fig. 1 is the most common one. Plasma iswarmed to body temperature before entering thesecondary filter (thermofiltration), being subse-quently passed completely through the fractionatormembrane in the dead-end mode (Fig. 1) (16,17,

20). Only 2 pumps are required for this procedure,keeping the system simple and easy to operate.

The study results fit well the membrane propertiesin terms of sieving coefficients, pore sizes, and pro-tein rejection curves. However, one might have ex-pected less marked differences in reduction rates.Sieving coefficients derived from the manufacturerand listed in Table 2 are similar for both membranesbut show a tendency toward higher permeability ofthe Kuraray secondary filter in the lower molecularweight range up to 300 to 400 kilodaltons. Fibrino-gen gel formation as mentioned above leads to dif-ferent behavior of this protein on and inside the fil-ter capillaries and higher reduction rates. Theprotein rejection curve shows a steeper increase withthe Kuraray Evaflux EVAL-5A membrane explain-ing the significantly higher reduction of serum IgMafter therapy with this membrane and, on the otherhand, sparing of IgG and IgA (12).

Premature rises of TMP in the Asahi AC1770Lsecondary filter cannot be easily explained. Al-though the surface area of 1.7 m2 is somewhatsmaller compared to 2 m2 in the Kuraray model,which may explain the differences in average maxi-mal TMP, the unpredictable saturation of the Asahifilter was related neither to a laboratory parameter(e.g., fibrinogen concentration) nor to peculiarities

FIG. 4. Shown are elimination rates for total protein and variousplasma proteins sorted according to increasing order of molecu-lar weight. A switch in magnitude of elimination rates between the2 fractionating filters can be noted between fibrinogen and IgM. Itcan be explained by differing protein rejection curves (p valuesfrom Student’s t-test, not significatn, p $ 0.05) (Ig: immunoglob-ulin).

TABLE 3. Laboratory values before and after MDF

PatientLDL-C

(mmol/L) ±SDHDL-C

(mmol/L) ±SDAlbumin

(g/L) ±SDIgG

(g/L) ±SDIgM(g/L) ±SD

AsahiHO

Before 3.71 0.45 1.03 0.03 35.70 1.17 10.73 0.59 0.19 0.01After 1.23 0.06 0.81 0.03 28.87 0.95 8.37 0.14 0.09 0.01

HFBefore 4.96 0.54 1.07 0.20 40.15 2.03 6.16 0.42 1.08 0.13After 1.83 0.27 0.79 0.10 31.80 1.51 4.81 0.29 0.53 0.08

FWBefore 4.79 0.47 1.14 0.06 35.88 0.75 12.70 0.41 1.15 0.10After 1.63 0.30 0.94 0.05 30.65 0.93 10.60 0.63 0.59 0.06

EJBefore 4.08 0.26 0.85 0.04 39.38 1.40 7.41 0.22 0.74 0.22After 1.23 0.12 0.72 0.04 34.22 1.20 6.02 0.25 0.29 0.07

KurarayHO

Before 3.56 0.41 0.92 0.05 33.53 1.64 10.99 0.87 0.21 0.02After 1.46 0.40 0.88 0.11 29.23 1.25 6.79 0.37 0.07 0.01

HFa

Before 3.86 1.05 40.60 6.12 0.95After 1.34 0.81 28.50 4.88 0.38

FWBefore 5.79 0.44 1.11 0.07 35.92 1.27 14.67 1.29 1.16 0.07After 1.60 0.16 0.99 0.04 33.13 2.47 12.45 0.99 0.42 0.03

EJBefore 4.37 0.42 0.80 0.01 39.00 1.42 7.83 0.47 0.59 0.03After 1.26 0.17 0.72 0.08 2.03 0.08 6.79 0.37 0.23 0.02

a Mean of 2 treatments, no standard deviation (SD).MDF: membrane differential filtration, LDL-C: low-density lipoprotein cholesterol, HDL-C: high-density lipoprotein cholesterol; Ig:

immunoglobulin.

LDL APHERESIS BY MEMBRANE DIFFERENTIAL FILTRATION 375

Artif Organs, Vol. 26, No. 4, 2002

of the extracorporeal therapy. Clogging of the frac-tionating filters belongs to the well known problemsof MDF and relates to the properties of the fraction-ating filter.

Comparing the differences between the tested fil-ters, the highest significance values (or lowest p val-ues) are achieved by HDL and IgM. In theory, thereduction of HDL by extracorporeal lipid-loweringsystems is undesirable since HDL is considered pro-tective in the pathogenesis of vessel wall damage.However, it has been shown that extracorporealLDL elimination over a longer period of time leadsto an increase of serum HDL levels (21). Thus thequestion whether the difference in HDL reduction isof clinical relevance cannot be conclusively an-swered. Comparing HDL values before the first andbefore the sixth treatment with each of the filtercombinations, respectively, did not show a signifi-cant decrease or increase of this lipoprotein withinthe time frame of the study.

As a theoretical possibility, heparin might havehad an impact on HDL levels through enhancinglipolysis and very low density lipoprotein catabolism,but applied heparin doses were identical in bothstudy groups. In all patients, they also exceeded thedoses that lead to maximal lipoprotein lipase activa-tion. Therefore, a heparin effect as a cause for thedifference in HDL reduction rates can be ruled out.

IgM as acute phase immunoglobulin is distributedmainly within the intravascular bed, and reduction ofplasma levels leads to a significant depletion of thetotal body content (22). Again, no difference couldbe found in IgM values before the first and sixththerapies. The clinical significance of IgM depletionhas not been analyzed systematically in the litera-ture, and it seems appropriate to control serum IgMlevels regularly in patients under MDF treatment. Atthis time, to our knowlegde, no fractionating filter isavailable that would spare IgM (12).

Elimination of lipoproteins in the study is largelycomparable to already published investigations witha tendency toward higher LDL elimination rates ac-companied by less removal of HDL (16,18,20,21,23–25). In the light of the present data, MDF configu-rations with HDL elimination rates up to 35% to40% are not suitable for the purpose of LDL elimi-nation (26). The unusually high HDL depletion rateswere related mostly to the employed secondary fil-ters with unfavorable protein rejection characteris-tics (26).

In this context one has to keep in mind that MDFis a heterogeneous procedure concerning employedfilters, methods of secondary filter perfusion (e.g.,thermofiltration) (17), treated plasma volume, pos-

sibly necessary substitution fluids (e.g., albumin orsodium chloride solution), the equipment, the in-house procedure (extracorporeal circulation maylead to either plasma dilution or plasma loss influ-encing measurements), the time point of blood with-drawal after apheresis (before or after reinfusion ofthe components from the extracorporeal circuit),and presentation of important technical details.Thus, data are influenced by many uncontrolled fac-tors and are not compared easily. Attempts to nor-malize elimination rates may add to comparabilitybut do not solve the entire problem (23,27).

Presumably, the lipoprotein elimination ratesdemonstrated in this study are close to the optimumthat can be achieved with the presently availablemembranes. This assumption is underlined by theselection of the secondary membranes. They havethe largest pore sizes available from the respectivemanufacturers. Consistent thermofiltration seems toimprove the results too.

CONCLUSIONS

The data show that MDF with the employed filtercombinations is suitable for LDL elimination, andsignificant differences in reduction rates for variousplasma proteins were found despite similar sievingcoefficients of the membranes. As a consequence,not MDF as such, but MDF employing a specificfilter combination in a specific setting (e.g., thermo-filtration) would be comparable to possibly more se-lective LDL elimination procedures.

REFERENCES

1. Martin MJ, Hulley SB, Browner WS, Kuller LH, WentworthD. Serum cholesterol, blood pressure, and mortality: Implica-tions from a cohort of 361,662 men. Lancet 1986;2:933–6.

2. Randomised trial of cholesterol lowering in 4444 patients withcoronary heart disease: The Scandinavian Simvastatin Sur-vival Study (4S). Lancet 1994;344:1383–9.

3. Mancini GB. Angiographic trials of lipid-lowering therapy:An update. Curr Opin Lipidol 1995;6:379–85.

4. Thompson GR, Myant NB, Kilpatrick D, Oakley CM,Raphael MJ, Steiner RE. Assessment of long-term plasmaexchange for familial hypercholesterolaemia. Br Heart J 1980;43:680–8.

5. Gordon BR, Saal SD. Current status of low density lipopro-tein-apheresis for the therapy of severe hyperlipidemia. CurrOpin Lipidol 1996;7:381–4.

6. Mabuchi H, Koizumi J, Shimizu M, Kajinami K, Miyamoto S,Ueda K, Takegoshi T. Long-term efficacy of low-density li-poprotein apheresis on coronary heart disease in familial hy-percholesterolemia. Hokuriku-FH-LDL- Apheresis StudyGroup. Am J Cardiol 1998;82:1489–95.

7. Aengevaeren WR, Kroon AA, Stalenhoef AF, Uijen GJ. vander WT. Low density lipoprotein apheresis improves regionalmyocardial perfusion in patients with hypercholesterolemiaand extensive coronary artery disease. LDL-Apheresis Ath-erosclerosis Regression Study (LAARS). J Am Coll Cardiol1996;28:1696–704.

G. MATIC ET AL.376

Artif Organs, Vol. 26, No. 4, 2002

8. Kroon AA, Aengevaeren WR. van der WT, Uijen GJ, ReiberJH, Bruschke AV, Stalenhoef AF. LDL-Apheresis Athero-sclerosis Regression Study (LAARS). Effect of aggressiveversus conventional lipid lowering treatment on coronary ath-erosclerosis. Circulation 1996;93:1826–35.

9. Thiery J, Seidel D. Safety and effectiveness of long-termLDL-apheresis in patients at high risk. Curr Opin Lipidol1998;9:521–6.

10. Lees RS, Cashin-Hemphill L, Lees AM. Non-pharmacologi-cal lowering of low-density lipoprotein by apheresis and sur-gical techniques. Curr Opin Lipidol 1999;10:575–9.

11. Gordon BR, Stein E, Jones P, Illingworth DR. Indications forlow-density lipoprotein apheresis. Am J Cardiol 1994;74:1109–12.

12. Sueoka A. Present status of apheresis technologies: Part 2.Membrane plasma fractionator. Ther Apher 1997;1:135–46.

13. Busnach G, Dal Col A, Perrino ML, Brando B, Brunati C,Minetti L. Performance evaluation of cascade filtration withhigh flow rate recirculating plasma on the secondary filter. IntJ Artif Organs 1987;10:121–8.

14. Valbonesi M, Garelli S, Montani F, Cefis M, Rossi U. Man-agement of immune-mediated and paraproteinemic diseasesby membrane plasma separation and cascade filtration. VoxSang 1982;43:91–101.

15. Homma Y, Mikami Y, Tamachi H, Nakaya N, Nakamura H,Araki G, Goto Y. Comparison of selectivity of LDL removalby double filtration and dextran-sulfate cellulose column plas-mapheresis. Atherosclerosis 1986;60:23–7.

16. Matsuda Y, Malchesky PS, Nose Y. Assessment of currentlyavailable low-density lipoprotein apheresis systems. Artif Or-gans 1994;18:93–9.

17. Nose Y, Usami M, Malchesky PS, Emura M, Horiuchi T,Sakamoto H, Smith JW. Clinical thermofiltration: Initial ap-plication. Artif Organs 1985;9:425–7.

18. Suzuki M, Yamane S, Matsugane T, Nobuto T, Azuma N,Nishide T, Shinomiya M, Saito K, Sasaki N, Nose Y. Evalu-ation of double filtration plasmapheresis, thermofiltration,

and low-density lipoprotein adsorptive methods by crossovertest in the treatment of familial hypercholesterolemia pa-tients. Artif Organs 1996;20:296–302.

19. Legallais C, Jaffrin MY. A process for enhancing selectivityand limiting plugging in plasmafractionation for APOB re-moval. Int J Artif Organs 1993;16:108–13.

20. Matsuda Y, Malchesky PS, Nose Y. Low-density lipoproteinremoval methods by membranes and future perspectives. Ar-tif Organs 1996;20:346–54.

21. Geiss HC, Parhofer KG, Donner MG, Schwandt P. Low den-sity lipoprotein apheresis by membrane differential filtration(cascade filtration). Ther Apher 1999;3:199–202.

22. Matic G, Hofmann D, Winkler R, Tiess M, Michelsen A,Schneidewind JM, Hebestreit G, Keysser M, Muller W, KinzeEM, Ramlow W. Removal of immunoglobulins by a proteinA versus an antihuman immunoglobulin G-based system:Evaluation of 602 sessions of extracorporeal immunoadsorp-tion. Artif Organs 2000;24:103–7.

23. Schwartzkopff W, Nold A, Grafenhahn H, von Baeyer H,Schwerdtfeger R. Efficacy of plasmapheresis, cascade-filtration and dextran-sulfate adsorption apheresis with calcu-lation of the recovery rate of lipids and lipoproteins. BeitrInfusionsther 1988;23:103–11.

24. Legallais C, Moriniere P, Wojcicki JM, Lalau JD, Fournier A,Jaffrin MY. A high selectivity cascade filtration technique forLDL-cholesterol and Lp(a) removal. Artif Organs 1995;19:887–95.

25. Valbonesi M, Bigi L, Ferrari M, Florio G. Cascade filtrationfor familial hypercholesterolemia. Int J Artif Organs 1991;14:751–3.

26. Busnach G, Cappelleri A, Vaccarino V, Franceschini G, DalCol A, Perrino ML, Brando B, Sirtori C, Minetti L. Selectiveand semiselective low-density lipoprotein apheresis in familialhypercholesterolemia. Blood Purif 1988;6:156–61.

27. Busnach G, Mandelli E, Brando B, Brunati C. Clinical rel-evance of fractionation characteristics in cascade filtration. IntJ Artif Organs 1993;16(Suppl 5):155–9.

LDL APHERESIS BY MEMBRANE DIFFERENTIAL FILTRATION 377

Artif Organs, Vol. 26, No. 4, 2002