CLIN. CHEM. 35/8, 1688-1693 (1989)

1688 CLINICAL CHEMISTRY, Vol. 35, No. 8, 1989

Lipase Activity Measured in Serum by a Continuous-MonitoringpH-Stat Technique-anUpdateNorbert W. Tletz, J. Rex Astles,1 and Denlee F. Shuey

Using recent knowledge regarding the roles of colipase, bileacids, Ca2t and emulsifiers, we optimized a previouslypublished pH-Stat method for lipase (EC 3.1.1.3) activitymeasurements. The recommended assay conditions are:olive oil/triolein, 100 mL/L; sodium glycocholate, 35 mmol/L;Ca2t 8.5 mmol/L; and colipase, 6.0 mg/L. The samplevolume is 0.10 mL, the reaction pH 9.0, the temperature30#{176}C,and the concentration of titrant 15 mmol/L. Hydroxy-propyl methylcellulose, 20 g/L, replaces acacia as emulsifierto avoid inhibition by excess Ca2. The standard curve islinear to >4566 U/L. The reference interval with olive oil assubstrate is 30-235 U/L. Upase activities withtnoleinsub-strate are 9.9% greater than with olive oil. Interference bypancreatic carboxylesterase (EC 3.1.1.1) activity is inhibitedby incubating the sample with diisopropylfluorophosphate.Results correlate well with those by the optimizedSingleVial#{174}method of Boehringer Mannheim Diagnostics(r = 0.997) and the immunochemical assay of BeckmanInstruments, Inc. (r = 0.995). Correlation with the acamethod (E.I. DuPont de Nemours & Company) is lesssatisfactory (r = 0.892), probably owing to lack of colipase inthe latter method.

Increased activity of lipase (triacylglycerol acylhydro-lase, EC 3.1.1.3) in serum is widely accepted as a goodindicator of pancreatitis(1-3). Measurement of lipase isclaimed to be more specific than measurement ofamylase,which is increased not only in pancreatitis but also invarious nonpancreatic abdominal and non-abdominal ill-nesses (4). There are many published methods for measur-ing lipase activity in serum, but no one method has beenwidely accepted. We previously published a continuous-monitoring method in which a pH-Stat is used to addstandard NaOH automaticallywhen fattyacids,liberatedby the actionof lipase,decrease the pH of the reactionmixture (5). Increased understanding ofthereactionmech-anisms forlipasenow makes it possibleto improve signif.icantly the specificity and sensitivityof thisassay.Suchimprovements are reported inthiscommunication.

Pancreatic lipase is rapidly and irreversiblyinactivatedat the interface of water and insoluble substrate because ofthe unfolding of its three-dimensional structure by the highsurface tension at the hydrophobic interface (6). Addition ofbile salts or albumin at low concentrations prevents thisinactivation, so we added bile salts in a concentration of2 mmoIJL to our previously recommended reaction mixture(5). This method, like other titrimetric assays, detects notonly lipase but also post-heparin lipolytic activity fromlipoprotein lipase (triacylglycero-protein acylhydrolase, EC

Division of Clinical Chemistry, Department of Pathology,Uni-versity of Kentucky Medical Center, Lexington, KY.

1 Present address: Dept. of Pathology,Medical College of Vir-ginia, Richmond, VA 23298.

ReceivedApril6,1989;acceptedJune 2, 1989.

3.1.1.34) and hepatic lipase. However, the addition of bileacids to the reaction mixture at supramicellar concentra-tions has been shown to inhibit lipase (7-9) and post-heparin lipolytic activity (10, 11) by physically excludingthese enzymes from the substrate-water interface (12-14).Lipase (but not lipoprotein lipase) activity can be fullyrestored by adding colipase to the reaction mixture. Suchrestorationis brought about by the formation of a bileacid-lipase--colipase complex that has a high affinity forthe substrate-water interface that anchors lipase to thesubstrate surface (9, 12). Formation of this complex notonly increases the specificity of the assay, it also increasesthe stability of lipase (15).

Because colipase is now recognized as an essential com-ponent of a specific lipase assay (16,17) its concentration inthe reaction mixture must be optimized, and its concentra-tion must be adequate to compensate for the differentamounts of colipase found in individual serum specimens(18, 19). Unfortunately, most of the experimental workwith colipase has been done with purified lipase prepara-tions (7-9) rather than with serum. In 1982, Hockebornand Rick (10), using a pH-Stat method, showed the benefi-cial effects of adding colipase to the reaction mixture whenserum specimens are analyzed. But in 1984 the sameauthors were unable to obtain linear reaction rates withtheir assay and recommended that colipase not be added tothe assay until a preparation of consistent purity hadbecome available (20). Junge (21) also added colipase to hisassay mixture, but our experiments have shown that hisprocedure issuboptinialwith regard to the bile acid con-centration and therefore detects lipoprotein lipase. In ad-dition,the assay is not optimized for Ca2. Thus, thereremains an urgent need for a reliable method of measuringlipase activity in serum, a method that takes into consid-eration the new knowledge regarding the roles of colipase,bile salts, Ca2, and emulsifiers.

Materials and Methods

instrumentation

We used a pH-Stat (Radiometer A/S, DK-2400 CopenhagenNV, Denmark; U.S. distributor: The London Company,Cleveland, OH 44145) consisting of a PHM 63 digital pHmeter with G2040B and K4040 electrodes, a TFA 60 titrationassembly, and a ITF 60 titrator. Parameters were set asfollows: temperature setting, 30 #{176}C;titration, upscale;propor-tional band, pH-Stat, delay shut-off, infinity; and endpoint,pH 9.00. The temperature in the reaction vessel was main-tained at 30#{176}Cby a Lauda K-2t’RD circulating water bath(Brinkmann Instruments, Inc., Westbury, NY 11590).

We added standard sodium hydroxide with an auto-burette, Model ABU 12 (Radiometer) equipped with a250-L burette and a 25-mL reaction vessel. The deliveryspeed ranged from 10 to 160. The autoburette was linkedthrough a mechanical connection to the Servograph re-corder, type REC 61 (Radiometer), operated with the penset at 28 mm/full scale and the chart speed set at “free.”

CLINICALCHEMISTRY, Vol. 35, No. 8, 1989 1689

Reagents

Purified olive oil and C02-free water. Both reagents wereprepared as described in our previous communication (5).

Olive oil (cat no. 0-1500) and triolein (99% pure, cat no.T-7140) were both obtained from Sigma Chemical Co., St.Louis, MO 63178.

Hydroxypropyl methylcellulose, 20 gIL. Dissolve 20 g ofhydroxypropyl methylcellulose (Sigma Chemical Co.; cat.no. H-8384, premium grade, USP) in 340 mL of CO2-freewater, heated to 80-90 #{176}C.Use a 1.L beaker, a magneticstirrer, and a 35-mm stirring bar. When solutioniscom-plete, add 640 mL of C02-free, chilled water to the beakerand stir. Store at 4-8 #{176}Cin a glass-stoppered bottle.

Substrate emulsion. To 28.6 mL of triolein or purifiedolive oil in a high-speed blender (e.g., Osterizer “Galaxie”14-speed blender) add 171.4 mL of hydroxypropyl methyl-cellulose solution, previously cooled to 4-8 #{176}C.Blend athigh speed for 5 mm, cool in refrigerator for 1 h, andcontinue the emulsification process at high speed foranadditional 5 miii. Store the reagent, tightly stoppered, at4-8 #{176}C.This emulsion is stable for five days.

Because the optimal substrate concentration is not de-termined by the amount but by the surface area of the oliveoil (triolein), the substrate must be sufficiently emulsified.Confirm the adequacy of the surface area by measuring thelipase activity with either a previously standardized con-trol specimen or with a patient’s specimen that has lipaseactivity �4000 UIL, using substrate concentrations at both100 and 150 mLIL. No increase in activity should beobserved with the higher substrate concentration.

Combined sodium glycocholate, 0.125 moliL, and CaCI2,0.030 moliL. Dissolve 6.095 g of sodium glycocholate(Behring Diagnostics, La Jolla, CA 92037; cat. no. 360512)and 0.4463 g of calcium chloride dihydrate (Fisher Scien-tific Co., Fair Lawn, NJ 07410; cat. no. C-79) in about 80mL of CO2-free water, stirring constantly. When solutioniscomplete,transferthe solutionto a 100-mL volumetricflask and ifilto the mark with CO2-free water.Storedat4-8 #{176}Cand protected from light,thissolutionisstableforatleast six weeks.

Colipase. Prepare a stock solution of colipase (BoehringerMannheim Diagnostics, Indianapolis, IN 64250; cat. no.644099) at a concentration of 1200 mg/L in sterile isotonicsaline (Travenol Laboratories, Inc., Deerfield, IL 60015; cat.no. 1F7124). Lyophilized preparations of colipase are gener-ally only about 60% colipase, so an appropriately greateramount must be weighed out to achieve the indicated con-centration. Store this reagent, in appropriate aliquots, at-70 #{176}C.Prepare a working solution of colipase, 600 mgfL, bydiluting one part of stock solution with one part of the sterilesaline. Colipasesolutionisstableforlongerthan a year at-70 #{176}C,and forat least two weeks when refrigerated.

Sodium hydroxide, 1.000 inoliL. Prepare thissolutionwith sodium hydroxidepellets(FisherScientificCo.,cat.no. S-318) and standardize.

Sodium hydroxide, 30 and 15 mmollL. Prepare fromsodium hydroxide stock solution, 1.000 molJL, by appropri-ate dilution with C02-free water. Confirm the titer of the15 mmolJL solution by titration against a standard solutionof potassium hydrogen phthalate.

Hydrochloric acid, 100 and 50 mmol/L. Prepare bothsolutions by appropriate dilution of standard HC1 withC02-free water.

Heparin, sodium (1000 USP unitslmL; Elkins-Sinn, Inc.,Cherry Hill, NJ 08034).

Diiosopropylfluorophosp hate, 3 mmol/L. In aqueous solu-tion, this has been recommended as a carboxylesteraseinhibitor (21a, 21b). Prepare an aqueous solution of diiso-propylfiuorophosphate (Sigma Chemical Co., cat. no. D-0879),approximately3 mmol/L, and storeina brown bottleat 4-S #{176}C.Refrigerated in the dark,thissolutionisstableforfivedays.

Diethyl-4-nitrophenyl phosphate (paraoxon, cat. no. D-4503), eserine (cat. no. E-8375), phenylmethyl.sulfonyl fluo-ride (cat. no. P.7626), p-arsanilic acid (atoxyl, cat. no.A-9258), and p-hydroxymercuribenzoic acid (cat. no. H-0642) were all from Sigma Chemical Co. Lanthanum chlo-ride (cat. no. L-9) was from Fisher Scientific Co.

Release of Lipoprotein Lipase

Six volunteershad fastedfor 12 h and alsoremainedseatedforatleast20 mm beforethesampling procedure, todecrease hemoconcentration. Blood was then sampled be-fore and 10 mm after an intravenous injection of sodiumheparmn, 10 units per kilogram body weight, to releaselipoprotein lipase (17).

Preparation of Instruments

Adjust the titrator settings, set the recorder to specifica-tions listed above, and fill the burette with 15 mmol/Lsodium hydroxide. Standardize the pH meter at 30#{176}Cwitha standard buffer. After each measurement, rinse thereaction vessel, the electrode, and all parts under the coverof the reaction vessel with de-ionized water. Keep theelectrodes wet at all times to prevent the drying ofoliveoilor triolein on the electrode surface. Between runs, place theelectrodes into pH 4 buffer (potassium biphthalate; FisherScientific, cat. no. SO-B-98). If results are not reproduciblewithin the desired precision goal or if reaction-rate curvesare not linear, soak the glass electrode overnight in HC1,0.10 molJL. At the end of each working day, dry theelectrodes with acetone and clean them with petroleumether to remove all lipid particles. Place the glass electrodein pH 4 buffer or in 0.10 moIJL HCI solution, as needed.Place the reference electrode into saturated KC1 solution.Meticulous cleaning and conditioning of the electrodes areessential for satisfactory performance of the procedure.

Assay

Final concentrationin reactionmixture:substrate(tn.oleinor oliveoil),100 mLIL; Na glycocholate,35 mmol/L;CaCl2,8.5mmolfL; and colipase,6 mg/L.

1. With a 10-mL syringe, dispense 7.0mL ofsubstratesolution into the reaction vessel.

2. Add 2.8 mL of sodium glycocholate-CaC12 solution.3. Add 0.10 mL of colipase, 600 mg/L.4. Secure the reaction vessel in its proper place. Guide a

gentle but constant stream of nitrogen gas over the surfaceof the mixture by inserting an appropriate tube throughthe port in the cover of the reaction vessel. This purgeremoves CO2 gas from the reaction mixture and preventsCO2 absorption from the atmosphere.

5. Adjust the pH of the mixture to about 9.00 withNaOH, 30 mmol/L. Mix for 4 or 5 mm to assurethat all CO2isremoved from the reaction vessel and that temperatureequilibriumisreached.The temperaturecan be monitoredwith a thermistor (Fisher Scientific Co., cat. no. 1517624)connected to a Model 43TA Tele-Thermometer (YellowSprings Instrument Co., Inc., Yellow Springs, OH 45387)and inserted through a port in the top of the reaction vessel.

-J

C,

00.

-J

4100

4000

1600

I500

200

100 -

1690 CLINICAL CHEMISTRY, Vol.35, No. 8, 1989

6. Add 0.10 mL of NaC1, 155 mmol/L, to the blank, and0.10 mL of specimen to the unknown.

7. Readjust the pH with either HC1, 50 mmolJL, orNaOH, 30 mmo]JL, until the pH meter reads between 8.99and 9.00.

8. Reset the burette volume counter to 0.00, and simul-taneously push the start button of the titrator and start astop watch. Note the burette reading every minute until aconstant rateismaintained during at least5 mm. Checkthe reaction-rate curve on the recorder-chart paper byplacing a ruler next to the line to confirm linearity, anddetermine the consumption of NaOH per minute for thetime during which the reaction rate curve was linear. Forexample,ifthereactionislinearduringtheinterval3 mmto 8 mm, subtractthe burette readingat 3 mm from theburette readingat 8 mm and divide by 5 mm.9. Determine the blanksinduplicateatthebeginningof

the first run of the day and recheck at several intervalsduring the day. If duplicates do not check within thedesired precision, additional blanks should be titrated; ifstill unsatisfactory, electrode maintenance may be re-quired.

Calculations

Lipase activity, U/L = (T/t.r - B/tB) X 0.015 X i0= (T/tT - B/tB) x 150

where U is defined as micromoles of free fatty acids formedper minute; T and B represent the volume of a 15 mmol/Lsolution of NaOH, in microliters, used to titrate the testand blank, respectively; t. and tB are the reaction times, inminutes, for the test and blank, respectively; and 10converts sample volume to liters, for a 100-p.L sample.

Results with the Proposed Method

Optimization of Reaction Parameters

Hydroxypropyl methykellulose. This emulsifier is a cel-lulose ether with no ionic charge, and thus it will notcomplex with metallic salts and ionic organic compounds. Itstabilizes the emulsion by decreasing the surface andinterfacial tensions and by increasing the viscosity of thewater phase (22). We confirmed the earlier reports ofSem#{233}rivaet al. (23), who found a 20 g/L solution mosteffective. Concentrations of 30 and 40 g/L render thereaction mixture too viscous and cause uneven mixing ofthe titrant and thus irregular reaction rate curves.

Triolein or olive oil emulsions. We determined optimalconcentrationsofsubstrateby measuring the lipase activ-ity with emulsions containing 50,80, 100, 120, and 150 mLof triolein or purified olive oil per liter. We obtainedmaximum activities with substrate emulsions containing80 mLfL of triolein or olive oil, and with the highersubstrate concentrations listed above. A concentration of100 mL/L was selected to provide for excess substrate. Weoptimized the olive oil substrate concentration by usingspecimens with slightly and with greatly increased lipaseactivities, in all cases using hydroxypropyl methylcelluloseconcentrations of 20 g/L.

Olive oil vs triolein. Lipase activities in 26 sera fromhospital patients, varying in activity from 55 to 3544 U/L,were compared by using purified olive oil and triolein, 99%,both at a final concentration of 100 mLIL. The meanactivity measured by using triolein was 9.9% greater thanthat obtained with olive oil as substrate. The regressionequation was y (triolein) = 1.lx (olive oil) + 5.4, the

correlation coefficient (r) was 0.998, and S = 45.8.These findings are in contrast to those of Hockeborn and

Rick (10), who found the same activities with olive oil andtriolein. The disparity isperhaps due to differencesin thequalitiesofoliveoil(especially)and triolein.These authorsused triolein,95% (Serva no. 37085), and triolein, A grade,>90% (Calbiochemno.6450),asopposed to the 99% triolein(Sigma) we used in our experiments. Triolein is the pre-ferred substrate when highest accuracy is desired, but oliveoil, because of its lower cost, is preferred for routine assaysand has been used in these experiments except whenotherwise specified.

pH optimum. The pH-optimization curves showed abroad plateau between pH 8.8 and 9.2. A reaction pH of 9.0(±0.1) isoptimal,because it is near the maximum activityfor all specimens tested. There is, however, little change inactivity from the optimal between pH 8.8 and 9.2.

Ca!2 optimization. The role of Ca2 in lipase activitymeasurements has been debated for many years. It isnowreasonably well established that Ca2 has one or more ofthe following roles: it protects the active center againstpoisoning by reaction products through precipitation of freefatty acids as calcium soaps at the substrate interface; itinteracts with emulsified substrate particles to reduce theirnegative electrostatic charge; it facilitates adsorption of theenzyme onto the substrate in the presence of bile salts; andit shortens the lag phase (24-27). Calcium concentrationsbetween 8 and 9 minol/L were found optimal for all speci-mens regardless of their lipase activity or the disease state(Figure 1); therefore we chose a concentration of 8.5 mmollL for use in our proposed method.

The inhibition of lipase activity by Ca2 concentrationsof �9.0 mmol/L is of interest because gum acacia, usedhitherto as an emulsifier in lipase assays, may containsubstantial amounts of calcium and magnesium. The solu-tion of gum acacia used in our experiments (cat. no. G-85,Fisher Scientific Co.) contained, per liter, 13.5 mmol ofCa2, 7.4 mmol of Mg, and 15 mmol of K. By contrast,hydroxypropyl methylcellulose contains none of these ions.We believe that these contaminants are responsible for thelower activity observed with gum acacia emulsions ascompared with emulsions prepared with hydroxypropyl

I I I7 8 9 10

Co2, rnmol/L

Fig. 1. Ca2 optimization curvespreparedby usingfour sera withdifferent lipase activities

>

U

C)

C0.

-J

150

100

50

.75 10 20 30 40 mmol/L

Acacla

160543

130124632290

183624

28052473

14.414.911.313.98.0

Mean 12.5

n, U/LSD, IJ/LCV, %

Within-day precision15

144325.5

16246

13.55.5

Day-to-dayprecision10 10

244 144515.8 27.6

1.8 6.4 1.9

CLINICALCHEMISTRY, Vol. 35, No. 8, 1989 1691

methylcellulose (Table 1).Glycocholate concentration. The bile salt concentration in

the reaction mixture should exceed the critical micellarconcentration in order to “clear” the substrate surface ofproteins, to inhibit the activity of interfering enzymes, andto form the lipase-colipase-bile salt complex. Figure 2shows results of experiments with sera from an individualbefore (curves A and B) and after administration of heparin(curves C and D). Curve A demonstrates the completeinhibition of lipase by sodium glycocholate concentrationsof �30 mmoL’L. The addition of colipase, 6 mgfL (curve B),restores lipase activity to a level exceeding that seen incurve A. This increase in activity is probably ascribable tosaturation of the enzyme with cofactor, protection of theenzyme from inactivation, and improved anchoring of theenzyme to the substrate. Post-heparin specimens (curves Cand D) show considerably higher lipolytic activity at lowconcentrations of bile salts, presumably owing to the re-lease of lipoprotein (and possibly hepatic) lipase. Addingbile salts to the reaction mixture to give a concentration�35 mmol/L causes complete inhibition of all lipolyticactivity (curve C). However, addition of colipase, 6 mg/L(curve D), restores lipase activity to near that seen inpre-heparmn specimens. We conclude from these experi-ments that bile salt concentrations of 35 mmol/L in thepresence of colipase, 6 mgfL, successfully inhibit all post-heparin lipoprotein lipase activity. Comparison with pre-heparmn values indicates that lipase activity is fully re-stored by the presence of colipase. The results of theseexperiments were confirmed on pre- and post-heparmnspec-imens from five other individuals. (See also Colipase con-centration, below.)

Although deoxycholate has been recommended by otherauthors (16,28) for use in lipase activity measurements, itis unsuited for our reaction mixture. Optimal calciumconcentrations of 8.5 mmolJL in the presence of deoxycho-late produce a heavy precipitate that interferes with titra-tion of the reaction mixture. Neumann et al. (29) also foundthat deoxycholate precipitates with calcium concentrations>2 mmol/L. (We observed precipitates in the reactionmixture proposed by Neumann et al. even at Ca2 concen-trations of 1 minol/L.)

Hockeborn and Rick (10) reported difficulties with spe-cific lots of glycocholate. We checked various lots of thisreagent, including the lot (Behring lot no. 810044, cat. no.360512) that was found unsatisfactory by these authors,but we saw no differences in results with our assay system.

Colipase concentration. Although glycocholate optimiza-tion experiments (Figure 2) established the need for acolipase concentration of 6 mgfL, we conducted additionalexperiments with reaction mixtures containing 3,6, and 12

Table 1. Lipase ActivIties (U/L) Compared with Acaclaand Hydroxypropyl Methyiceliulose as Emulslflers and

Olive Oil as SubstrateHydroxypropylmethylcellulose Difference, %

1448

No Glycocholote

Fig. 2. Glycocholate optimization with use ofsera from an individualbefore and after administration of hepannCurveA (*_*), pre-heparinserum samplewithoutcolipase; curve B (0-0),pre-heparinserumsamplewithcolipase,6mg/I; curve C(s- . .), post-heparinserumsamplewithout colipase;curve D (0- - -0) post-hepannserum samplewithcolipase,6 mg/I

mg of colipase per liter. Lipase activity was determined byusing specimens with moderately and greatly increasedlipase activity. These experiments confirmed that a coli-pase concentration of 6 mg/L is adequate to maintainoptimal lipase activity, even in specimens with a very high(2400 U/L) lipase activity.

Linearity. We checked the linearity of the method byanalyzing a serum specimen with 4566 UIL of lipaseactivity and making serial dilutions with either heat-inactivated serum or NaC1, 155 mmol/L, or by taking asmaller sample. We observed linearity throughout theentire range, regardless of the diluent or sample size used.

The ability of the system to accept sera diluted witheither heat-inactivated serum or NaC1 solution is a signif-icant advantage over previous assay systems adapted to thepH-Stat technique (5, 10). This improvement, broughtabout by the change in glycocholate concentration and theaddition of colipase, also decreases the relative dependenceof the system on the use of a consistent sample size.However, a substantial increase in sample size-e.g., to 200p.L-results in a slight decrease in activity, possibly owingto the competition between serum proteins and lipase forthe substrate surface.

Precision. In Table 2 we compare day-to-day and within-day precision for serum specimens with different lipaseactivities.

Reaction temperature. In accordance with the recommen-dations of the International Federation for Clinical Chem-istry (IFCC), the method was optimized at 30 #{176}C.However,lipase activity measurements performed at 25, 30, 35, 37,and 40#{176}Cindicated a linear increase in activity between 25

Table 2. PrecIsion Studies with Two Serum Specimens

IS

l0UCI’3

L5

Fig. 3. Histogram of lipase activity valuesobtainedfor 116 healthyadults

. 116

1692 CLINICAL CHEMISTRY, Vol. 35, No. 8, 1989

and 37#{176}C.The rate of increase with increase in tempera-ture becomes smaller at a temperature of 40#{176}C,as anArrhenius plot showed. These findings are in contrast toour earlier experiments (5), in which we observed temper-ature stability only up to 30#{176}C.This increase in stabilitycan be attributed to the presence of colipase, which stabi-lizes lipase at the substrate-water interface (15). The Q10for lipase between 25 and 35#{176}C,as measured with ourreaction mixture, is 1.7.

Effect of NaCI on Lipase Activity

Arzoglou et al. (30) reported that adding sodium chlorideto the reaction mixture in both the Boehringer MannheimDiagnostics (Indianapolis, IN 64250) turbidimetric lipaseprocedure and a slightly modified titrimetric procedure (31)results in an increase in activity that is affected by both thesodium chloride concentration and the type of lipase isoen-zyme present. According to these authors, lipase activitiesof sera from healthy subjects and from patients withoutpancreatitis peak in the presence of NaCl, 80 mmol/L.However, sera from patients with pancreatitis exhibitedthe greatest increase in activity in the presence of NaC1,140 mmol/L. Arzoglou et al. therefore recommended thatsodium chloride be added to the reaction mixture as ameans of differentiating hyperlipasemia resulting fromacute pancreatitis from that attributable to other causes.When we added NaC1 to our recommended reaction mix-ture, regardless of the NaCl concentration or the type ofspecimen used, there was a decrease in lipase activity.Thus this approach, when applied to our method, does notaid in the differentiation of the causes of hyperlipasemia.

Schandi and Pittner (32) demonstrated with immobiliza-tion techniques that both alkali-earth and alkali metalions-especially Ca2, Mg2, and Na-stabilize the ac-tive center of the enzyme and promote adsorption of theenzyme onto the substrate surface. These authors foundcalcium ions to be particularly effective in this regard.Because our proposed method is optimized with regard tocalcium ions, and because calcium ions are more effectivestabilizers than sodium ions, our method does not benefitfrom the presence of additional sodium ions.

Reference Interval

We established the normal reference interval by analyz-ing sera from 116 rested, fasting laboratory employees(Figure 3). The central 95th percentile of results was 30 to235 UIL with olive oil as substrate. Values obtained withtriolein as substrate were -9.9% higher.

Inhibition of Carboxylesterase Activity

If the procedure is used as a reference method, pancreaticcarboxylesterase activity can be inhibited by pre-incu-bating 200 L of serum with 40 pL of diisopropylfluoro-phosphate stock solution for 30 mm at room temperature.Use 100 L of the inhibited sample for the assay and makea mathematical correction for the dilution, if the method isused for clinical purposes, diisopropylfluorophosphate neednot be added because pancreatic carboxylesterase is in-creased in the same clinical conditions as lipase; in addi-tion, there is little (<12%) interference by this enzymeunder the proposed reaction conditions. Alternatively, car-boxylesterase interference can be eliminated by replacingsodium glycocholate with taurodeoxycholate, 35 mmol/L(33). In this case, however, lipase activity is decreased byabout 45%.

Junge (21) reports that “nonspecific pancreatic carbox-ylesterase” (no EC no.) is also detected by methods involv-ing the use of triglyceride substrates. However, use ofcarboxylesterase inhibitors (diisopropylfluorophosphateand eserine) does not inhibit this enzyme, suggesting thatit may not be a carboxylesterase. With use ofother inhib-itors, the enzyme behaves similarly or identically to lipase.Diisopropylfiuorophosphate, eserine, phenylmethylsulfo-nyl fluoride, atoxyl, and LaCl3 do not inhibit either en-zyme, whereas p-hydroxymercuribenzoic acid partly inhib-its both enzymes. Paraoxon inhibits the nonspecific carbox-ylesterase totally but causes only partial inhibition oflipase and nonlinear reaction kinetics. Thus, the identity ofthis enzyme isnot clearand it cannot be ruled out that it isa lipase isoenzyme.

Method Comparisons

The proposed method was compared with a “SingleVial”method (Boebringer Mannheim Diagnostics), modified asdescribed previously (34). We measured the activity in serafrom 151 hospitalized patients with normal, moderately,and greatly increased lipase activities. As reported in ourprevious communication (34), results by these methodsagreed well except in the low portion of the referenceinterval. The correlation coefficient (r) is 0.997, and theregression equation is y = 0.57x + 52.2 (S = 63.9).

We also compared our proposed procedure with the acamethod (DuPont Co., Wilmington, DE 19898) and an im-munochemical assay developed by Beckman Instruments,Inc. (35). Experiments were carried out with the sameserum specimens used in the comparison with the modifiedBoehringer Mannheim Diagnostics method. The correla-tioncoefficientforthe pH-Stat vs Beckman method is 0.995(S),,, = 42.9). A graphic representation of the data is givenin Figure 4. The correlation coefficient for the pH-Stat vsthe aca method is r = 0.892 (S = 17.4) (see Figure 4). Therelatively poor correlation of the aca method with othermethods isattributed to the lack of colipase in the reactionmixture of the aca method. This interpretation agrees withresults reported by Junge et al. (19), who found a signifi-cantly better performance of the aca method when colipasewas added to the aca reaction mixture.

Colipase used inour experiments was kindly provided by Dr.K.Wulff, Boehringer Mannheim GmbH, Tutzing, F.R.G., and by Dr.Stephan Hesse, Vice-President, Boehringer Mannheim GmbH,Mannheim, F.R.G. Human pancreatic carboxylesterase (EC3.1.1.1; M 100 000) was kindly provided by Dr.Dominique Lom-bardo, Marseilles, France, and nonspecific pancreatic carboxyl-esterase (noEC number; Mr 54000) by Dr. W. Junge, Kiel, F.R.G.

ng/mL10,000

n 151r#{176}0.995

B 1,000#{149}C . .,.

Iim . .

o100

Ia10 100 1,000 IQ000 U/L

pH - Stat

U/d L

0c 100 -0

l0.

n 151

r#{176}0.892

CLINICAL CHEMISTRY, Vol. 35, No. 8, 1989 1693

100 1,000 10,000 U/L

p1-1 -Stat

Fig.4. Comparison oflipase activities from 151 hospitalized patientsmeasured with the pH-Stat method and (top) the immunochemicalassay developedby BeckmanInstruments, Inc. (35) or (bottom) theDuPontaca method

References1. Kolars JC, Ellis CJ, Levitt MD. Comparisonofserum amylase,pancreatic isoamylase and lipase in patientswith hyperainyl-asemia.Dig Die Sci 1984;29:289-93.2. Orda R, Orda S, Baron J, Wiznitzer T. Lipase turbidimetricassay and acute pancreatitis. Dig Dis Sci 1984;29:294-6.3. Song H, Tietz NW, Tan C. Usefulness ofserum lipase, esterase,and arnylase estimation in the diagnosis of pancreatitis-.a corn-Sparison. Clin Chem 1970;16:264-8.4. Salt WB II, Schenker S. Amylase-its clinical significance: areview of the literature [Review]. Medicine 1976;55:269-81.5. Tietz NW, Repique EV. Proposedstandard method for measur-ing lipase activity in serum by a continuous sampling technique.Clin Chern 1973;19:1268-75.6. Brockerhoff H. On the function of bile salts and proteins ascofactors of lipase. J Biol Chem 1971;246:5828-31.7. BorgstrOm B. On the interactions between pancreatic lipaseand colipase and the substrate, and the importanceof bile salts.JLipid Res 1975;16:411-7.8. Vanderrneers A, Vanderrneers..Piret MC, Rath#{233}J, ChristopheJ. Effect of colipase on adsorption and activity of rat pancreaticlipase on emulsified tributyrin in the presence of bile salt. FEBSLett 1975;49:334-7.9. Momsen WE, Brockman HL. Inhibition of pancreatic lipase Bactivity by taurodeoxycholate and its reversal by colipase. J BiolChern 1976;251:384-8.10. Hockeborn M, Rick W. Zur Bestimxnung der katalytischenAktivit&t der Lipase mit dem kontinuierlichen titrimetrischenTest. J Clin Chem Clin Biochem 1982;20:773-85.11. Shirai K, Matsuoka N, Saito Y, Yoshida S. Post-.heparinplasma hepatic triacylglycerol lipase-’catalyzed hydrolysisof tribu-tyrin. Biochim BiophysActs 1984;795:1-8.12. Chapus C, Sari FH, S#{233}m#{233}rivaM, DesnuelleP. Roleof colipase

in the interfacial adsorption of pancreatic lipase at hydrophilicinterfaces. FEBS Lett 1975;58:155-8.13. BorgstrOm B, Erlanson C. Pancreatic lipase and co-lipase.Interactions and effects of bile salts and other detergents.Eur JBiochem 1973;37:60-8.14. Lairon D, Nalbone G, Lafont H, et al. Inhibition of lipaseadsorption at interfaces. Role of bile salt micelles and colipase.Biochemistry 1978;17:205-8.15. Borgstrom B. The temperature-dependent interfacial inactiva-tion of porcine pancreatic lipase. Effect of colipase and bile salts.Biochim Biophys Acta 1982;712:490-7.16. Neumann U, Knitsch KW, Ziegenhorn J, R#{233}derA, Zwez W,Kramer W. Reagens zur Lipasebestimmung und Verfahren zuseiner Herstellimg. German patent no. 2904305, US patent no.4343897.17. Hoffmann GE, Weiss L. Specific serum pancreatic lipase de-termination, with use of purified colipase. Clin Chem1980;26:1732-3.18. Junge W, Leybold K. Detection of colipase in serum and urineof pancreatitis patients. Clin Chim Acts 1982;123:293-302.19. Junge W, Leybold K, Kraack B. Influence of colipase on theturbidimetric determination of pancreatic lipase catalytic activity.J Clin Chem Clin Biochem 1983;21:445-51.20. Rick W, Hockeborn M. Katalytische AktivitAt der Serumli-pase im kontinuierlichen titrimetrischen Test in Abhflngigkeitvon der Temperatur. J Clin Chem Clin Biochem 1984;22:357-70.21. Junge W. Pancreatic lipase, titrimetric method. In: BergmeyerU, ed. Methods of enzymatic analysis. Vol IV. Weinheim: VerlagChemie, 1984:15-25.21a. Mayli#{233}MF, Charles M, Desnuelle P. Action of organophos-phates and sulfonyl halides on porcine pancreatic lipase. BiochimBiophysActs 1972;276:162-75.21b. LombardoD. Modification of the essential amino acids ofhuman pancreatic carboxylic-esterhydrolase. Biochirn BiophysActs 1982;700:67-74.22. Handbook on methocel, cellulose ether products. Midland, MI48640: Dow Chemical Co.23. S#{233}m#{233}rivaM, Benzonana G, Desnuelle P. Some properties of alipase from Rhizopus arrhizus. Separation of a glycopeptide boundto the enzyme. Biochim Biophys Acts 1969;191:598.-610.24. Verger R. Pancreatic lipases. In: BorgstrOm B, Brockman HL,eds. Lipases. Amsterdam: Elsevier Science Publishers, 1984:84-150.25. Leger C, Charles M. Pancreatic lipase [Review]. World RevNutr Diet 1980;35:96-128.26. Brown WJ, Belmonte AA, Melius P. Effects of divalent cationsand sodium taurocholate on pancreatic lipase activity with gumarabic-emulsified tributyrylglycerol substrates. Biochim BiophysActs 1977;486:313-21.27. Benzonana G. Sur le role des ions calcium durant l’hydrolysedestriglycerides insolubles par Ia lipase pancreatique en presencede eels biliaires. Biochim BiophysActs 1968;151:137-46.28. Kannisto H, Lalla M, Lukkari E. Characterization and elim-ination of a factor in serum that interferes with turbidimetry andnephelometry of lipase. Clin Chem 1983;29:96-9.29. Neumann U, Kaspar P, Ziegenhorn J. Pancreatic lipase,turbidimetric method. In: Bergmeyer U, ed. Methods of enzymaticanalysis. Vol. VI. Weinheim: Verlag Chemie, 1984:26-34.30. Arzoglou PL, F#{233}rardG, M#{233}taisP. Differentiating two forms ofplasma lipase by use of media with different ionic strengths. ClinChern1984;30:360-3.31. Tietz NW, Fiereck EA. Measurement of lipase activity inserum.Standard Methods Clin Chem 1972;7:19-31.32. Schandi A, Pittner F. The role of Na and Ca2’ ions on theaction of pancreatic lipase studied with the help of immobilizationtechniques. Eur J Biochem 1984;140:547-51.33. Lombardo D, Guy 0. Studies on the substrate specificity of acarboxylester hydrolase from human pancreatic juice. BiochimBiophysActs 1980;611:147-55.34. Tietz NW, Shuey DF, Astles JR. Turbidimetric measurementof lipase activity-problems and some solutions. Clin Chem1987;33:1624-9.35. RoodFW, Coleman PF. Solid-phase enzyme immunoassay forthe quantitation of pancreatic lipase in human serum [Abstract].Clin Chem 1986;32:1129.