Improved protamine-sensitive membrane electrode for monitoring heparin concentrations in whole blood via protamine titration^§

¹ Departments of Chemistry and
² Anesthesiology, and
³ College of Pharmacy, The University of Michigan, Ann Arbor, MI 48109.
^a Address correspondence to this author at: Department of Chemistry, The University of Michigan, 930 N. University Ave., Ann Arbor, MI 48109-1055. Fax 313-647-4865; e-mail mmeyerho{at}umich.edu.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
An improved protamine-sensitive electrode based on a polymeric membrane doped with the charged ion exchanger dinonylnaphthalenesulfonate (DNNS) is used for monitoring heparin concentrations in whole blood. The electrode exhibits significant nonequilibrium potentiometric response to polycationic protamine over the concentration range of 0.5–20 mg/L in undiluted whole-blood samples. The sensor can serve as a simple end point detector for the determination of heparin via potentiometric titrations with protamine. Whole-blood heparin concentrations determined by the electrode method (n 157) correlate well with other protamine titration-based methods, including the commercial Hepcon HMS assay (r = 0.934) and a previously reported potentiometric heparin sensor-based method (r = 0.973). Reasonable correlation was also found with a commercial chromogenic anti-Xa heparin assay (r = 0.891) with corresponding plasma samples and appropriate correction for whole-blood hematocrit levels. Whereas a significant positive bias (0.62 kU/L; P <0.001) is observed between the anti-Xa assay and the protamine sensor methods, insignificant bias is observed between the protamine sensor and the Hepcon HMS tests (0.08 kU/L; P = 0.02). The possibility of fully automating these titrations offers a potentially simple, inexpensive, and accurate method for monitoring heparin concentrations in whole blood.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Heparin, a highly sulfated polysaccharide, is the anticoagulant of choice in extracorporeal procedures such as cardiopulmonary bypass surgery (CPB).² Accurate measurement of heparin concentrations during such a procedure is important because high heparin doses can lead to bleeding complications in patients. Accurate monitoring of blood heparin concentrations can also aid in heparin therapy by providing a means to determine the minimal protamine dose required to neutralize the anticoagulant activity of heparin at the conclusion of CPB. This is of clinical significance since protamine overdose is known to cause complement activation and other toxic effects in patients. Furthermore, Despotis et al. recently demonstrated that closely regulated heparin concentrations at somewhat higher doses than traditionally used in CPB could significantly decrease postoperative complications and the use of blood products (1).

At present, the use of heparin during CPB is typically monitored by clotting-time measurements, the most common and widely used one being the activated clotting time (ACT). This method, however, is indirect and its results can be misleading because ACT values are affected by many variables such as hemodilution and hypothermia that do not correlate directly with plasma heparin concentrations (2)(3). Chromogenic heparin assays based on factor Xa inhibition, although widely used in clinical laboratories, cannot be performed with whole-blood samples (4). The only method currently available to determine heparin concentrations in whole-blood samples is the Hepcon HMS assay system (Medtronic Blood Management), which is based on a protamine titration and the use of clot formation for end point detection. The Hepcon assay is performed on a cartridge consisting of four to six channels that contain different amounts of protamine as well as dilute thromboplastin. The end point of the test is the detection of clot formation, which is achieved by measuring the movement of a plunger mechanism in each cartridge. The channel containing the smallest quantity of protamine that completely neutralizes the heparin in the sample exhibits the shortest clotting time. The heparin concentration is determined from the quantity of protamine in that channel (on the basis of stoichiometry of the protamine–heparin complex). Each Hepcon cartridge thus tests a limited range of blood heparin concentrations. Often, multiple cartridges must be run, resulting in an expensive and time-consuming procedure. Moreover, the Hepcon determines heparin in a "discontinuous" manner (i.e., results are displayed in intervals of 0.7 kU/L), and its accuracy is typically within ± 0.34 kU/L (Hepcon HMS insert). Nevertheless, the Hepcon assay is the most reliable and rapid point-of-care test method available. Whole-blood heparin concentrations determined by the Hepcon assay system have been shown to correlate well with laboratory-determined plasma anti-Xa concentrations after correcting for blood hematocrit levels (5).

Recent research in our laboratories has demonstrated that appropriately formulated polyvinyl chloride (PVC) films doped with tridodecylmethylammonium chloride (TDMAC) or potassium tetrakis(4-chlorophenyl)borate (KTpClPB) exhibit large and reproducible potentiometric responses toward low concentrations of heparin (6)(7) and protamine (8), respectively. Response of these polyion-sensitive electrodes has been ascribed to the favorable extraction of the polyions into the membrane phase via cooperative ion-pairing interactions with lipophilic ion exchanger sites doped within the polymeric membrane phase. This extraction process yields a nonequilibrium steady-state change in the phase boundary potential (EMF) at the membrane/sample interface (9).

Herein we describe the clinical utility of an improved polycation-sensitive electrode based on a polymeric membrane doped with the cation exchanger dinonylnaphthalenesulfonate (DNNS) for determining heparin concentrations in whole blood. Such a DNNS-based polymeric membrane electrode has been reported previously by our group (10) for use in developing assays for specific proteases involving polycationic substrates. Although we have shown earlier that whole-blood heparin concentrations can be monitored with the TDMAC-based heparin-sensitive electrode (11), pseudotitrations with this electrode are very labor intensive, owing to the irreversibility in the sensor's heparin response. Consequently such pseudotitrations could only be carried out by using multiple tubes containing equal aliquots of whole blood but different amounts of protamine in each tube. In contrast, titrations with the DNNS-based protamine-sensitive membrane electrode are direct and can be performed more easily by monitoring the EMF change after the addition of small aliquots of protamine to a single heparinized sample of whole blood. Because the sensor does not respond to the heparin–protamine complex, it can be used to monitor the titration end point (the presence of excess protamine). The heparin concentration is then determined by using a predetermined binding stoichiometry between heparin and protamine.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
reagents
High-molecular-weight PVC, TDMAC, KTpClPB, calcium bis[4-(1,1,2,3-tetramethylbutyl)phenyl] phosphate (CaTMBPP), bis(2-ethylhexyl) sebacate (DOS), 2-nitrophenyloctyl ether (NPOE), tris(2-ethylhexyl) phosphate (TOP), dioctyl phthalate (DOPth), dioctyl adipate (DOA), and tetrahydrofuran (THF) were obtained from Fluka Chemika Biochemika. Protamine sulfate (from herring), beef lung heparin, and tris[(hydroxy methyl)amino methane] (Tris) were from Sigma Chemical Co. DNNS was a kind gift from King Industries. Polyurethane (M48) was kindly supplied by Medtronic Inc. Tecoflex SG-80A was a gift from Thermedics Inc. and Pellethane 2363–80AE was from Dow. Polyurethane PU-2060 was a gift from Prof. G. S. Cha (Kwangwoon University, Seoul, S. Korea). Injectable heparin (from porcine intestinal mucosa, 1000 USP kilounits/L) was from Elkins-Sinn Inc. Fresh frozen human plasma was obtained from the American Red Cross. All other reagents were of analytical grade. All solutions were prepared with distilled deionized water. Unless otherwise stated, the primary buffer solution used in all experiments was 50 mmol/L Tris-HCl, pH 7.4, containing 120 mmol/L NaCl.

preparation of protamine-sensitive cylindrical membrane electrodes
Protamine-sensitive cylindrical membrane electrodes (Fig. 1 ) were prepared as described previously (10). The membrane casting solution was formulated by dissolving 200 mg of the components (DNNS:polymer:plasticizer = 1.0:49.5:49.5 by weight) in 2 mL of THF. This casting solution was then dip coated (12 times at 30-min intervals) over rounded glass rods protruding from a narrow-bore Tygon tube (i.d. 1.3–1.5 mm) and dried overnight. After soaking in 15 mmol/L NaCl for about 6 h, the glass rods were carefully removed, and the tube was then internally filled with 50 mmol/L Tris-HCl, pH 7.4, containing 120 mmol/L NaCl. An Ag/AgCl wire was then inserted into the inner bore of the tube to serve as the internal reference electrode. Before use, all cylindrical electrodes were presoaked in a 15 mmol/L NaCl solution for at least 6 h. Unless otherwise specified, these cylindrical electrodes were used in all experiments and disposed of after each use.

View larger version (30K): [in this window] [in a new window]   Figure 1. Schematic diagram of a protamine-sensitive membrane electrode used in this study and an expanded view of the extraction and ion-pairing of the polycation with DNNS within the membrane phase.

measurement of emf response of dnns-doped cylindrical electrodes
The EMF responses of the cylindrical DNNS-based electrodes were measured vs a miniature external Ag/AgCl reference electrode (Bioanalytical Systems Inc.), via a Macintosh IIcx computer coupled with an NB-MIO analog/digital input/output board (National Instruments) and a VF-4 electrode interface module (World Precision Instruments) controlled by LabView 2 software (National Instruments) as described in our previous work (12).

potentiometric titrations with protamine-sensitive membrane electrodes
Potentiometric titrations of various heparin preparations were performed by adding small aliquots of a protamine solution (1 g/L) to 10 mL of a well-stirred buffer solution containing a fixed concentration of either porcine or beef heparin. The resulting potential change was recorded 3 min after each addition. Titration curves were constructed by plotting the potential change (from the initial baseline value) vs the concentration of added protamine. Titration end points were computed by the Kolthoff method (13), followed by applying a subtractive correction factor. The latter was done to offset the effect of the small protamine concentration required to elicit sensor response. For blood measurements, the correction factor was 4 mg/L protamine (which corresponds to 0.4 kU/L heparin). The heparin/protamine binding ratios determined from the corrected end points were then used to determine heparin concentrations in unknown samples.

heparin determination in whole-blood samples
Twenty-two cardiac surgery patients were enrolled in this study after providing informed consent. The protocol used in these studies was preapproved by the Human Subjects Internal Review Board at the University of Michigan. All patients received a bolus dose of 300 units/kg of porcine heparin to achieve systemic anticoagulation. Additional doses of heparin were administered as needed to maintain a kaolin ACT of 480 s or above. Whole-blood samples were drawn from patients before the initial heparin infusion, every 30 min after heparin infusion, and after neutralization with protamine. Blood samples were drawn into EDTA-coated Vacutainer Tubes (Becton Dickinson) and stored at 4 °C until analysis. Hepcon assays were performed immediately on each blood sample (non-EDTA) in the operating room. A kaolin-activated ACT (Hemochron; International Technidyne Corp.) value was obtained at each of the time points as well. Plasma samples for the anti-Xa (Coatest^®, Pharmacia Hepar) heparin assay were obtained by centrifuging a portion of the respective blood samples at 670g for 20 min. The plasma samples were frozen until subsequent analysis. Plasma heparin concentrations thus obtained by the Xa assay were corrected for blood hematocrit with the following formula:

Heparin determinations with the heparin sensor were carried out via pseudotitrations with protamine as previously reported (11)(14). In brief, whole blood (250 µL) was added to each of a series of 11 tubes, each coated with a known but increasing quantity of dried protamine sulfate. The EMF response of a cylindrical heparin sensor in each tube vs an external Ag/AgCl reference electrode was recorded sequentially with the computerized LabView setup mentioned earlier. A protamine titration curve was constructed by plotting the EMF readings against the amount of protamine in each tube. The regression lines for the two linear portions of the titration curve were extrapolated to yield the break point, from which the heparin concentration was determined. Potentiometric titrations with DNNS-based electrodes in whole blood were performed by using the procedure described in the previous section, except that a smaller volume (4 mL) of blood was used for each titration. Heparin concentrations were determined from the corrected end points by using the heparin/protamine neutralization stoichiometry determined with the sensor (1 unit of heparin binds to 10 µg of protamine).

statistical analysis
Bias among the various methods was calculated as recommended by Bland and Altman (15). Student's paired t-test was used to compare the mean whole-blood heparin measurements between the various methods. Least-squares linear regression was used to estimate a linear relation and generate correlation coefficients for heparin determinations from the methods, with P values <0.05 considered statistically significant. Confidence intervals were computed at the 95% level.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
optimization of protamine-sensitive membrane electrodes
The development and analytical utility of polyion-sensitive membrane electrodes have been demonstrated only recently (7). Development of these polymer membrane electrodes requires the identification of appropriate ion-complexing agents and membrane chemistries that yield significant potentiometric response to polyionic macromolecules. The choice of the ion exchanger used within the membrane is thus very crucial for their optimal performance. Plasticized PVC membranes doped with lipophilic borate (8) or organophosphate salts were shown previously to elicit significant response towards protamine in physiological saline solution. However, their poor response (<5 mV) in complex media such as human plasma restricted their analytical utility, thereby prompting the need to identify an improved ion exchanger that would provide a more favorable and selective extraction of protamine into the membrane phase. Toward this end, DNNS was investigated as the ion exchanger for use in the development of improved polycation-sensitive electrodes. The use of DNNS as an ion exchanger in polymeric membranes has been reported previously in ion-selective electrodes for lanthanides (16) and organic cations (17). Only recently was DNNS suggested as a useful ion exchanger in the design of polycation-sensitive electrodes for use in monitoring synthetic polycationic peptides (10). From Fig. 2 it is obvious that the membrane formulation with DNNS offers a significant improvement in the maximum EMF response as well as detection limits towards protamine over other ion-complexing agents tested. This is very likely due to the high ion-pairing affinity of DNNS towards protamine within the organic membrane phase of the electrode. It should be noted that detection limits towards protamine improve when the DNNS concentration in the membrane phase is reduced below 1 wt% (in accordance with the mechanism of polyion response (9)); however, such reduction in DNNS concentrations yield a much lower total EMF response at higher concentrations of protamine (e.g., 10 mg/L). Hence, 1 wt% was chosen as a compromise between detection limits and the magnitude of observed EMF signal.

View larger version (15K): [in this window] [in a new window]   Figure 2. Effect of the specific cation exchanger on the potentiometric protamine response of cylindrical polycation-sensitive electrodes in 50 mmol/L Tris-HCl, pH 7.4, containing 120 mmol/L NaCl. Electrode membranes were composed of DNNS:NPOE:PVC, 1:49.5:49.5 by weight (); KTpCIPB:NPOE:PVC, 1:33:66 by weight (); and CaTMBPP:TOP:PVC, 1:49.5:49.5 by weight (). The average response of four electrodes ± SD is shown.

Plasticizers play an important role in polymeric membrane electrodes by solvating the charged components within the organic phase. In polyion-sensitive electrodes, they also influence the diffusion coefficients of the polyions in the membrane. A host of plasticizers (DOS, DOA, NPOE, TOP, and DOPth) were examined in films containing DNNS (Fig. 3 ). It is noted that although significant protamine response was observed with most plasticizers, membranes containing the polar plasticizer NPOE yielded the best detection limits and total EMF response. It appears that the relatively high polarity of this plasticizer (dielectric constant, = 24 vs = 3–5 for other plasticizers) enables a stronger cooperative ion pairing of DNNS with protamine in the membrane phase, thereby enhancing the extraction of protamine into the polymer film. The precise mechanism by which NPOE enhances the ion pair formation in the membrane phase is currently under investigation.

View larger version (19K): [in this window] [in a new window]   Figure 3. Effect of the plasticizer component on the protamine EMF response of DNNS-based cylindrical PVC membrane electrodes in 50 mmol/L Tris-HCl, pH 7.4, containing 120 mmol/L NaCl. Membranes were composed of DNNS:PVC:plasticizer (NPOE, ; DOA, ; TOP, ; DOPth, ; DOS, ), 1:49.5:49.5 by weight. The average response of at least six electrodes ± SD is shown.

The nature of the polymer also plays a significant role in a membrane's polyion response, especially for applications in undiluted plasma/whole blood. The EMF response in undiluted plasma is usually much lower compared with that in a physiological buffer solution, presumably because of nonspecific adsorption of various plasma proteins on the surface of the membrane. Because polyurethanes have been preferred over PVC for use as biomaterials (18), they were investigated as the polymeric matrix component in the development of an improved protamine-sensitive membrane electrode. The effect of varying the polymer matrix on the protamine response of DNNS-doped membranes in undiluted human plasma is shown in Fig. 4 . The use of polyurethanes as the polymeric membrane component appears to offer very significant advantages for preparing protamine-sensitive electrodes. Notably, membranes containing Tecoflex as the polymeric membrane component exhibited the greatest total EMF response; however, this was at the expense of detection limits. This is probably due to the lower rigidity of Tecoflex compared with PVC, which results in a greater diffusion coefficient in the membrane phase that affects the accumulation of polyions at the membrane/sample interface (9) and hence the magnitude of the EMF response at low protamine concentrations. Alternatively, membranes prepared with the polyurethanes M48 and PU-2060, which contain a higher proportion of hard segments, showed improved detection limits. The lower detection limits and high potentiometric response of these electrodes render them the optimal choice for titrations of blood heparin concentrations with protamine. Thus, optimized membranes containing 1.0 wt% DNNS, 49.5 wt% o-NPOE, and 49.5 wt% M48 were used in all subsequent studies.

View larger version (19K): [in this window] [in a new window]   Figure 4. Effect of the polymer component on the protamine EMF response of DNNS doped, NPOE-plasticized cylindrical membrane electrodes in undiluted human plasma. The membranes contained DNNS:NPOE:polymer (PVC, ; M48, ; PU 2060, ; Tecoflex, ; Pellethane, ), 1:49.5:49.5 by weight. The average response of four electrodes ± SD is shown.

heparin determinations via potentiometric titrations
The DNNS-based membrane electrodes can be used to detect the end point of the titrations of heparin with protamine. Indeed, as shown in Fig. 5 , heparin concentrations as low as 0.5 kU/L can be readily determined. Mass neutralization ratios of protamine with bovine and porcine heparin (powdered forms) as measured with the protamine-sensitive membrane electrode agree well with values determined with a heparin-responsive membrane electrode (19). Titrations in whole-blood samples obtained at different time intervals during the course of a typical bypass surgery are shown in Fig. 6 . Heparin concentrations determined with DNNS-based electrodes correlate well with those determined with the previously reported heparin sensor-based method (11)(14), the Hepcon assay, and the chromogenic anti-Xa assay (Table 1 ). The correlations for heparin measurements by both sensors and the Hepcon assay with the ACT, however, were poor. This finding is expected, given that ACT values are affected by various factors such as hemodilution and temperature, which yield no effect on whole-blood heparin concentrations.

View larger version (19K): [in this window] [in a new window]   Figure 5. Typical potentiometric titration of 0.0 (), 0.5 (), 1.0 (), 2.0 (), and 4.0 () kU/L, respectively, of porcine heparin with protamine as monitored by polycation-sensitive membrane electrodes containing DNNS:NPOE:M48 1:49.5:49.5 by weight in 50 mmol/L Tris-HCl, pH 7.4, containing 120 mmol/L NaCl. The average response of three electrodes ± SD is shown.

View larger version (28K): [in this window] [in a new window]   Figure 6. Typical changes in whole-blood heparin concentrations over the duration of heart surgery as monitored by titrations with protamine with the protamine-sensitive membrane electrode. Blood samples S1 to S5 were heparinized; S0 was the preheparinized sample and S6 was obtained after neutralization with protamine. The average response of three electrodes containing DNNS:NPOE:M48, 1:49.5:49.5 by weight is shown.

The reproducibility of the sensor-based titration method was assessed by performing multiple titrations on whole-blood samples supplemented with 1 and 2 kU/L porcine heparin. Heparin concentrations of 1.03 ± 0.15 kU/L (±SD) (n = 5) and 1.93 ± 0.25 kU/L (±SD) (n = 5) respectively were determined by the sensor by performing manual titrations on these samples.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Currently, the Hepcon assay is the only available method for determining heparin concentrations in whole blood. The speed of the assay (30 to 250 s) and a wide measurable range of heparin concentrations (from 0.0 to 8.2 kU/L) make it particularly attractive as a point-of-care test method. However, the appropriate cartridge must be chosen for a given assay, such as to encompass the heparin concentration in the sample to be tested. Thus, if the heparin concentration is determined as a result of clot formation in the first (or last) channel within the cartridge, then the assay must be repeated with a cartridge capable of measuring a lower (or higher) heparin concentration. Hence, multiple assays may be required in such cases. Further, erroneous positive deviations were noted in the Hepcon assay in five samples within this study, where the Hepcon showed an increase (1 kU/L) in heparin concentration even when no additional heparin was administered to the patient. Such deviations were likely due to systematic errors associated with the Hepcon (e.g., plunger problems). The protamine sensor-based method, however, enables determination of heparin over the entire range of concentrations that are likely to be encountered during bypass surgery. Further, the sensor can be used to follow the titration and thus concentrations of several low-molecular-weight heparin preparations (e.g., Fragmin by Pharmacia Hepar) (data not shown). Such low-molecular-weight heparins are currently used clinically for patients with deep venous thromboembolism and cannot be monitored by clotting time-based assays such as ACT (20) or activated partial thromboplastin time (21), since they do not increase clotting times significantly.

Upon infusion into the patient, the functional blood concentration of heparin is dependent on the binding of heparin to various plasma proteins as well as heparin metabolism. Whole-blood heparin concentrations therefore continually decrease over time during CPB (see Fig. 6 ), often requiring the administration of an additional dose of heparin to the patients to maintain adequate systemic anticoagulation on the basis of ACT times. Titrations with the protamine sensor can follow these blood heparin concentration changes over the duration of the surgical procedure. Absence of heparin in preheparinized (S0) or postprotamine (S6) samples is reflected by a lack of a clear break point in the resulting titration curves.

As shown in Table 1 , the Hepcon and potentiometric sensor-based methods show good correlation between each other, and reasonable correlation with heparin measurements with the chromogenic anti-Xa assay. Minimal bias was found in heparin measurements between the heparin and protamine sensors (0.17 kU/L; P = 0.001), and between the protamine sensor and the Hepcon (0.08 kU/L; P = 0.02). However, a significant positive bias was observed with the anti-Xa assay compared with the other methods. This may be due to the continuous metabolism of heparin during surgery that yields lower-molecular-weight heparin fragments. These fragments bind only weakly to protamine and therefore may not be detected in titrations with protamine. However, such fragments may still possess significant anti-Xa activity and thus be detected via the chromogenic assay.

The ACT has long been used in clinical practice to assess the protamine dose required for heparin neutralization after CPB. Typically, the desired protamine doses are calculated from the infused heparin concentrations on the basis of a dose of 1 mg of protamine for every 100 units of infused heparin. However, because of the aforementioned metabolism of heparin during surgery and binding of heparin to various plasma proteins, a reduced protamine dose should normally be sufficient to fully neutralize the remaining heparin. Previous reports indicate that reduced protamine doses after surgery can decrease perioperative blood losses (22), perhaps by reducing complement activation. The excellent detection limits of the DNNS-based polycation-sensitive electrode towards protamine (< mg/L) makes it possible to detect a protamine overdose, which is reflected by an increased baseline EMF value when the electrode is placed into postprotamine blood samples. Indeed, in this study, the postprotamine samples showed an average increase of about 10 mV in the absolute starting potentials when compared with those of the heparinized samples, indicating a protamine overdose in all cases examined. As suggested by others (1)(23), to maintain an appropriate state of hemostasis in patients undergoing open heart surgery, the actual blood heparin concentrations, and not just the ACT values, should be closely monitored.

Although the titrations performed in this study involved manual addition of protamine to the blood samples, resulting in a rather lengthy assay time (3 min/titration point; 30–45 min for entire manual titration, depending on the number of points), preliminary attempts were made also to automate such titrations by using a syringe pump for continuous protamine infusion into heparinized samples. Fig. 7 presents a typical set of automated titration curves obtained for three different heparin concentrations in buffer solution. Heparin concentrations can be easily and rapidly determined from the end point of these curves, providing the concentration and infusion rate of the protamine solution are known. Preliminary results with whole-blood samples indicate that the heparin concentrations determined in this manner are in excellent agreement with those found by the manual titrations (data not shown). Multiple automated titrations performed on blood samples supplemented with heparin (1 and 2 kU/L) yielded concentrations of 1.00 ± 0.17 (±SD) (n = 8) and 1.97 ± 0.25 kU/L (±SD) (n = 8) respectively. Indeed, a full clinical evaluation of this new automated methodology is currently in progress.

View larger version (29K): [in this window] [in a new window]   Figure 7. Titration of 0.0 (), 1.0 (), and 2.0 () kU/L, respectively, of porcine heparin performed by continuous infusion (25 µL/min) of a protamine solution (1 g/L) in 10 mL of buffer (50 mmol/L Tris-HCl, pH 7.4, 120 mmol/L NaCl) as monitored by polycation-sensitive membrane electrodes containing DNNS:NPOE:M48, 1:49.5:49.5 by weight. The average response of three electrodes is shown. Inset: The titration curves obtained by continuous infusion (25 µL/min) of protamine (1 g/L) in 5 mL of whole blood containing 0.0 (), 1.0 (), and 2.0 () kU/L heparin.

In summary, an improved membrane electrode that shows significant potentiometric response to protamine has been optimized and used to determine heparin concentrations in whole-blood samples via protamine titration. Such heparin determinations show good correlation with other currently available methods, including the Hepcon and the anti-factor Xa assays. Because this method is specific for heparin and does not require clot formation for the detection of titration end points, it can be used in samples devoid of clotting factors (e.g., serum) or blood samples containing other anticoagulants (e.g., EDTA, citrate). In addition, the electrode allows for the determination of heparin on a continuous scale (unlike the Hepcon method), rendering it suitable for the measurements of a wider range of heparin concentrations. By using the mode of continuous protamine infusion via a syringe pump, we envision that a simple portable system equipped with disposable DNNS-based electrodes could be adapted for bedside heparin monitoring.

	Acknowledgments

 
We thank Theresa Ambrose for her helpful review of this manuscript. This work was supported in part by NIH grants GM 28882, HL 38353, and HL 55461 and a research grant from Medtronics Inc.

	Footnotes

 
¹ Dedicated to the memory of Dr. Jong Hoon Yun.

² Nonstandard abbreviations: CPB, cardiopulmonary bypass surgery; ACT, activated clotting time; PVC, polyvinyl chloride; TDMAC, tridodecylmethylammonium chloride; KTpCIPB, potassium tetrakis(4-chlorophenyl)borate; EMF, electromotive force (phase boundary potential); DNNS, dinonylnaphthalenesulfonate; CaTMBPP, calcium bis[4-(1,1,2,3-tetramethylbutyl)phenyl] phosphate; DOS, bis(2-ethylhexyl) sebacate; NPOE, 2-nitrophenyloctyl ether; TOP, tris(2-ethylhexyl) phosphate; DOPth, dioctyl phthalate; DOA, dioctyl adipate; and THF, tetrahydrofuran.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Despotis GJ, Joist JH, Goodnough LT. Monitoring of hemostasis in cardiac surgical patients: impact of point-of-care testing on blood loss and transfusion outcomes. Clin Chem 1997;43:1684-1696. [Abstract/Free Full Text]
2. Cohen EJ, Camerlengo LJ, Dearing JP. Activated clotting times and cardiopulmonary bypass: I. The effect of hemodilution and hypothermia upon activated clotting time. J Extra Corpor Technol 1980;12:139-141.
3. Culliford AT, Gitel NS, Starr N. Lack of correlation between activated clotting time and plasma heparin level during cardiopulmonary bypass. Ann Surg 1981;193:105-111. [ISI][Medline] [Order article via Infotrieve]
4. Teien AN, Lie M, Abildgaard U. Assay of heparin in plasma using a chromogenic substrate for activated Factor X. Thromb Res 1976;8:413-416. [ISI][Medline] [Order article via Infotrieve]
5. Despotis GJ, Summerfield AL, Joist JH, Goodnough LT, Santoro SA, Spitznagel E, et al. Comparison of activated coagulation time and whole blood heparin measurements with laboratory plasma anti-Xa heparin concentrations in patients having cardiac operations. J Thorac Cardiovasc Surg 1994;108:1076-1082. [Abstract/Free Full Text]
6. Ma SC, Yang VC, Fu B, Meyerhoff ME. Electrochemical sensor for heparin: further characterization and bioanalytical applications. Anal Chem 1993;65:2078-2084. [Medline] [Order article via Infotrieve]
7. Meyerhoff ME, Fu B, Bakker E, Yun JH, Yang VC. Polyion-sensitive membrane electrodes for biomedical analysis. Anal Chem 1996;68:168A-175A. [Medline] [Order article via Infotrieve]
8. Yun JH, Meyerhoff ME, Yang VC. Protamine-sensitive polymer membrane electrode: characterization and bioanalytical applications. Anal Biochem 1995;224:212-220. [ISI][Medline] [Order article via Infotrieve]
9. Fu B, Bakker E, Yun JH, Yang VC, Meyerhoff ME. Response mechanism of polymer membrane based potentiometric polyion sensors. Anal Chem 1994;66:2250-2259. [Medline] [Order article via Infotrieve]
10. Han IS, Ramamurthy N, Yun JH, Schaller U, Meyerhoff ME, Yang VC. Selective monitoring of peptidase activities with synthetic polypeptide substrates and polyion-sensitive membrane electrode detection. FASEB J 1996;10:1621-1626. [Abstract]
11. Meyerhoff ME, Yang VC, Wahr JA, Lee LM, Yun JH, Fu B, et al. Potentiometric polyion sensors: new measurement technology for monitoring blood heparin concentrations during open heart surgery. Clin Chem 1995;41:1355-1356. [Free Full Text]
12. Telting M, Collison ME, Meyerhoff ME. Simplified dual-lumen catheter design for simultaneous potentiometric monitoring of carbon dioxide and pH. Anal Chem 1994;66:576-583. [Medline] [Order article via Infotrieve]
13. Sergeant EP. Potentiometry and potentiometric titrations. Kolthoff IM Elwing PJ eds. Chemical analysis 1985;Vol. 69:362-364 John Wiley and Sons New York. .
14. Wahr JA, Yun JH, Yang VC, Lee LM, Fu B, Meyerhoff ME. A new method of measuring heparin levels in whole blood by protamine titration using a heparin-responsive electrochemical sensor. J Cardiothorac Vasc Anesth 1996;10:447-450. [ISI][Medline] [Order article via Infotrieve]
15. Bland MJ, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-310.
16. Harrell JB, Jones AD, Choppin GR. A liquid ion-exchange membrane electrode for polyvalent ions. Anal Chem 1969;41:1459-1462.
17. Martin CR, Freiser H. Response characteristics of ion-selective electrodes based on dinonylnaphthalenesulfonic acid. Anal Chem 1980;52:562-564.
18. Anderson JM, Marchant KK. Platelet interactions with biomaterials and artificial devices. CRC Crit Rev Biocompat 1985;1:111-204.
19. Yun JH, Ma SC, Fu B, Yang VC, Meyerhoff ME. Direct potentiometric membrane electrode measurements of heparin binding to macromolecules. Electroanalysis 1993;5:719-724.
20. Greiber S, Weber U, Galle J, Bramer P, Schollmeyer P. Activated clotting time is not a sensitive parameter to monitor anticoagulation with low molecular weight heparin in hemodialysis. Nephron 1997;76:15-19. [ISI][Medline] [Order article via Infotrieve]
21. Carter CJ, Kelton JG, Hirsh J, Gent M. Relationship between the antithrombotic and anticoagulant effects of low molecular weight heparin. Thromb Res 1981;21:169-174. [ISI][Medline] [Order article via Infotrieve]
22. Despotis GJ, Joist JH, Hogue CW, Alsoufiev A, Kater K, Goodnough LT, et al. The impact of heparin concentration and activated clotting time monitoring on blood conservation. J Thorac Cardiovasc Surg 1995;110:46-54. [Abstract/Free Full Text]
23. Despotis GJ, Joist JH, Hogue CW, Alsoufiev A, Joiner-Maier D, Santoro SA, et al. More effective suppression of haemostatic system activation in patients undergoing cardiac surgery by heparin dosing based on heparin blood concentration rather than ACT. Thromb Haemost 1996;76:902-908. [ISI][Medline] [Order article via Infotrieve]