CUN. CHEM. 39/6, 1064-1068 (1993)

1064 CUNICAL CHEMISTRY, Vol. 39, No. 6, 1993

Stability of Intact Chorionic Gonadotropin (hCG) in Serum, Liquid Whole Blood, andDried Whole-Blood Filter-Paper Spots: Impact on Screening for Down Syndrome byMeasurement of Free f3-hCG SubunitKevin Spencer,”3 James N. Macri,’ Paul Carpenter,’ Robert Anderson,2 and David A. Krantz2

The use of multiple maternal serum biochemical markersin screening for Down syndrome is gaining worldwideacceptance. We sought tostudy the impact of the poten-tial instability of intact human chorionic gonadotropin(hCG) on free -hCG subunit, a marker that has recentlybeen used successfully in such screening. We found that,in practice, any changes in free -hCG due to theinstability of intact hCG do not inhibitthe effectiveness offree -hCG as a marker for Down syndrome. This wasproven by controlled laboratory experiments at variousstress temperatures, freeze-thaw studies, and analysis ofa large set of screening data with particular reference totime in transit for indMdual samples. Data from controlleddissociation studies demonstrate that any apparent in-crease in free -hCG due to the instability of intact hCGcannot be attributed simply to the dissociation of intacthCG. Finally, for large-scale mass populationscreening inareas of the world where transport delays, safety con-cerns, and high temperatures preclude the shipment ofliquidwhole blood, dried whole-blood spots on filter paperprovide a suitable delivery system with many advantages.

IndexIng Terms: sample handling . variation, source offetal status peptide hormones

We have recently established, in both independent(1-3) and joint studies (4), that the free a-subunit ofchorionic gonadotropin (hCG) is the marker of choice fordetecting chromosomal trisomies in early gestation,even during the first trimester (5). It has been sug-gested, based on anecdotal evidence (6), that the con-centrations of free 13-hCG increase as a result of im-proper sample storage, transport of samples at ambienttemperature, and leaving samples at ambient tempera-ture for 30 mm or more. This is allegedly due to theinstability of intact hCG, resulting in release of the-hCG subunit. Gau et al. (7) and Bidart et al. (8) havedescribed increased concentrations of free 13-and freea-hCG in serum samples that had been heated to de-plete the serum complement components. However, thisinvolved heating serum to 56#{176}Cfor 30 mm-clearly nota realistic environment for Down syndrome screeningsamples.

Most biological materials consist of large and complexmolecules and it is characteristic of such materials that

1Endocrine Unit, Clinical Biochemistry Department, OldchurchHospital, Romford, Essex RM7 OBE, UK

2Research l)ivision, NTD Laboratories Inc., 383 Old Country

Rd.,Carle Place, NY 11514.3Address correspondenceto this author.Received October 6, 1992; accepted January 27, 1993.

they are susceptible to thermally induced dissociative orconformational changes that significantly alter theiractivity. Although every precaution should be taken totransport, prepare, and store biological materials underconditions that favor stability, almost inevitably somechange in activity will occur. It is essential to be able toestimate the rate at which such changes occur and tounderstand the impact of such changes, if any, on theclinical interpretation of the analyte in question (9).

Stability testing has its roots in the pharmaceuticalindustry, where various techniques, including thermaldegradation studies, have been used since the early1950s to determine the shelf life of pharmaceuticalproducts (10) and more recently have been implementedin the production of biological standards (9) and themanufacture of quality-control material (11).

The Arrhenius equation is the basis of most methodsof accelerated thermal stability testing (12). In essence,this equation states that the speed of a reaction isdependent upon temperature; hence, the log of thereaction rate is a linear function of the reciprocal ofabsolute temperature. it is this relationship that allowsthe data obtained at high temperatures to be used to

predict the degradation rate at room temperature (orany other temperature) and hence to estimate stability.

Throughout this paper, we will use the standardisednomenclature with respect to stability and accelerateddegradation studies; however, for free p-hCG, thechanges seen are not those of degradation but are“apparent” increases in free f3-hCG concentration,brought about as a result of the instability of intacthCG.

Little attention has been paid to the stability ofanalytes used in clinical chemistry for the monitoring,diagnosis, and treatment of disease, other than briefstudies more than one or two decades ago (13-16).Furthermore, the practice of storing rare samples orsamples from various clinical experiments over longperiods, without knowledge of the stability of the vari-ous analytes, must cast doubt on the validity of somestudies. The observationsof Wilding et al. (16) 16 yearsago, that “very few publications are available whichdocument the degree of deterioration of various constit-uents over periods of time and at different tempera-tures,” is sadly still true today. We have thereforesought to use proven stability-testing protocols to gainscientific data on the stability of intact hCG and itssubunits, not only in the controlled environment of theresearch laboratory, but in the practical environment ofa Down syndrome screening program.

CLINICAL CHEMISTRY, Vol. 39, No. 6, 1993 1065

Materials and Methods

Sample Collection Conditions

The study population consisted of women attendingthe Antenatal Clinic for their routine screening tests forneural tube defect and Down syndrome at 15-18 weeksof gestation. All patients’ samples were collected (afterinformed consent) into 7-mL plain Vacutainer Tubescontaining no anticoagulant or preservative (BectonDickinson U.K Ltd., Cowley, Oxford, UK). Serum washarvested from the samples after clot formation andaliquoted into polypropylene tubes in amounts suitablefor each experiment. Aliquots of samples were stored atvarious temperatures, depending upon the require-ments of the specific experimental protocol.AU of thepregnancies involved in this study resulted in the birthof a normal baby.

For dried filter-paper spot studies, heparinized whole-blood samples were spotted onto Schleicher and Schuellspecimen collection paper (Grade 903) to cover a 12-mm-diameter ring. The spots were allowed to dry at roomtemperature and then storedunder the respective ex-perimental conditions.

Measurement of Free 13-hCG

Free p-hCG was measured with a solid-phase two-siteiinmunoradiometric assay [ELSA-Fb HCG; CIS (UK)Ltd., High Wycombe, Bucks, UK] in which the monocle-nal antibodies were raised against sterically remoteepitopes on the 3-hCG molecule that are obscured whenthe p-hCG is in combination with the a-subunit. Themonoclonal capture antibody used in this assay isnamed FBT11. The analytical performance of this assayhas been previously described (2). The assay was cali-brated against the First International Reference Prep-aration (75/551) for free /3-hCG.

We also measured free f3-hCG with the microtiterplate ELISA described by Macri et al. (1, 17). Themicrotiter plate imxnunoassay was performed in dupli-cate, and absorbance measurements were made with anSLT plate reader [Tecan (UK) Ltd., Berkshire, UK]. TheCV of assays of duplicates of each patient’s sample was<5%.

The specificity of both the CIS assay and the ELISAhas been previously described (17) and shown to besimilar (4). Assays involving the FBT11 monoclonalantibody have been shown to measure both nicked andnonnicked forms of free p-hCG (18).

For filter-paper spot studies, we punched 5-mm-diam-eter spots from the collection cards and analyzed themin a modified form of the ELISA, using whole-blood spotcalibrators. The between-assay precision (CV) of thisassay at 11.2 lUlL was 2.9% and at 17.8 lU/L was 4.2%.

Measurement of Free a-Subunit

Free a-hCG was measured with a solid-phase two-siteimmunoradiometric assay (Biocode Free Alpha SubunitIRMA, Metachem Diagnostics Ltd., Northampton, UK).The assay was calibrated against the First Interna-tionalReference Preparation (75/569) for a-hCG. The

quoted cross-reactivity with intact hCG was <0.1%, andfree J3-hCGwas undetectable.

All assays were performed in duplicate according tothe manufacturer’s conditions. The bound radioactivitywas counted with an NE 1600 multihead gammacounter (NE Technology Ltd., Reading, UK); the countswere processed by using the WHO mass action curve-fitting routine (19). Counting times were adjusted toaccumulate 10000 counts at a concentration of 2 lU/Lfor free (3-hCG and 0.5 lU/L for free a-hCG.

Protocols

Accelerated degradation study. The degradation rateconstant for free /3-hCG was determined at 60#{176}C(333K), 37 #{176}C(310 K), 20#{176}C(293 K), and 4 #{176}C(277 K). Foreach temperature to be analyzed, 10 patients’sampleswere divided into 500-tL aliquots and placed in waterbaths or a refrigerator at the respective experimentaltemperature.

For the 60#{176}Cstudy, aliquots of the samples wereremoved to and placed at 4#{176}Cat 5-min intervals for 1 h.All aliquots were then analyzed within the same assaybatch.

For the 37#{176}Cstudy, aliquots were removed to 4#{176}Cat24-h intervals for 1 week. All aliquots were then ana-lyzed within the same assay batch. In a separate exper-iment we repeated the 37#{176}Cstudy, using samples withadded sodium aside, 1 gIL.

For the 20#{176}Cstudy, aliquots were removed at dailyintervals for the first week and at weekly intervals for afurther 4 weeks. Aliquots taken over the first week werestored at 4#{176}Cand analyzed in the same batch. Eachseries of subsequent weekly aliquots were analyzedtogether at the end of each week. The between-assayprecision over the assay period was monitored with ourinternal quality-control procedures, which demon-strated a between-assay CV for free f3-hCG of 6.3% at1.09 lU/L, 4.9% at 11.73 lU/L, 3.8% at 21.09 lU/L, and4.4% at 40.89 IU/L.

For the 4#{176}Cstudy, aliquots were removed at weeklyintervals for 4 weeks. Each series of weekly aliquotswere analyzed together, along with those from the 20#{176}Cstudy.

Liquid whole-blood stability study. To simulate delaysin sample separation, such as might occur in postal

transport of specimens to the receiving laboratory, wedivided whole blood from 10 patients into two portions.The first portion was allowed to clot and the serum wasleft on the clot at 22#{176}C.Aliquots were removed fromeach whole-blood sample after gentle inversion, thealiquots were centrifuged, and the serum aliquot wasstored at 4#{176}C.This cycle was repeated at daily intervalsfor 8 days. For the second portion, the serum wasseparated from the clot and left at 22#{176}C.Aliquots wereremoved from each sample at daily intervals (for 8 days)and stored at 4#{176}C.AU serum aliquots were analyzed

within the same assay batch.Dried filter-paper blood-spot stability study. To assess

the stability of whole blood transported as dried whole-blood spots on filter paper, we spotted blood from five

1.25C)x

1.2uJ

uJLU 1.15U-

zi.i

z4I 1.05C)‘U

0.950 20 40 60 80 100 120 140

Storag. tsinp., ‘C

20104

-20-40

1066 CUNICAL CHEMISTRY, Vol. 39, No. 6, 1993

patients onto the collection cards and dried the cards atroom temperature. The liquid whole blood from eachpatient was aliquoted into four lots and then stored atthe respective storage temperatures (4#{176}C,25#{176}C,and37#{176}C)along with the respective dried filter-paper bloodspot. After 9 days at the storage temperature, the liquidwhole-blood samples were measured in one batch in theconventional ELISA procedure and the dried bloodspotswere analyzed in the modified blood-spot ELISA proce-dure.

Freeze/thaw study. To assess the impact of repeatedfreezing and thawing of samples on their free p-hCGconcentration, we subjected 10 patients’ samples to aseries of freeze/thaw cycles. Samples were allowed tothaw at room temperature, an aliquot was removed to4#{176}C,and the samples were returned to the freezer(-20 #{176}C)until they had frozen again. This cycle wasrepeated 25 times over the next 48 h and the aliquotswere analyzed together within the same assay batch.

Influence of stability on Down syndrome screening. Toassess the impact of a delay in receipt of patients’samples (as both whole blood and serum) on the medianvalues obtained in a nationwide screening program inthe US, we analyzed the median values for free -hCG,broken down by the time delay to receipt of the samples,for a group of 2972 patients.

To ascertain the impact of significant artifactual in-creases of free -hCG values on screening programs, wetook 685 consecutive results from our screening pro-gram and calculated the individual patient’s specificrisk of Down syndrome and the initial positive rate forthe screened population. The free p-hCG values werethen hypothetically increased by 10%, and the respec-tive calculations repeated.

Results

To calculate the activation energy, one must firstcalculate the degradation rate constant at each “stress”temperature, which can be obtained by plotting therelative changes of analyte concentration with time.Figure 1 shows such a plot for the 20#{176}Cstudy. The slopeof the regression line is the degradation rate constant.For each of the “stress” temperatures-4, 20, 37, and60#{176}C-thedegradation rate constants were 0.000023,0.0016, 0.0077, and 1.105, respectively. Figure 2 showsthe Arrhenius plot obtained by plotting the variousdegradation rate constants against the reciprocal of theabsolute temperature. The slopeof the regression line isthe activation energy (Ea). The stability of any analyteis commonly held to be the time required to initiate a10% change from the original analyte concentration(11). The stability of free $-hCG is therefore calculatedfrom the relationship:

Stability = 0.105/K1

where K1 is the reactionrate constant at the stabilitytemperature, derived from the regression equation ofthe Arrhenius plot. Table 1 showsthe predicted stability

HOURS AT 293 KFig. 1. Relativechangeof serum free p-hCG withtimeof storageat293 K (20 ‘C)Eachpoint (#{149})Is the mean for 10 patients; (+) Is 2 x SEM, measuredwith theCISassay

2.8 2.9 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8

1000 * 1/TEMPERATURE (K)Fig. 2. Arrhenlus plotfor the stability of free -hCG(+) measured rate constant, (#{149})regressed data point

Table 1. StabIlIty of Free p-hCG In Serum, as PredIctedfrom the Arrhenlus Equation

StMIlfty

87.3 h30.3 days115 days131 years51131 years

of free 13-hCG in separated serum for various storagetemperatures.

To eliminate any involvement of enzymatic break-down of intact hCG at high temperatures, we repeatedthe 37 and 60#{176}Cincubation studies with five serumsamples containing sodium aside. The results (notshown) were not significantly different from those with-out aside.

Figure 3 shows the relative change of free p-hCG with

0 20 40 60 80 100 120 140 160

HOURS AT 295 K

CUNICAL CHEMISTRY, Vol. 39, No. 6, 1993 1067

2.C)

‘ULU

LU

IL.

zLU

z4zC)Lu>I-<1.2-JLU

1

Fig. 3. Relative change of free p-hCG in serum (*) and in liquidwhole blood(0) stored for varioustimesat 295 K (22 ‘C)Each point Is the meanfor 10 patients measured with the CIS assay

time at 22#{176}Cfor storage as liquid whole blood comparedwith separated serum. The degradation rate constant inserum, 1.51 x iO, yields a projected stability of 69.5 h.For liquid whole blood, the degradation rate curveshows a logarithmic change with time, with a muchfaster degradation occurring after 70 h, at which timeall the samples had become hemolyzed. The degradationrate constant from the early linear portion of the deg-radation curve was 3.09 x iO, for a projected stabilityof 34.0 h.

Table 2 shows data for a series of five samples forwhich, after accelerated degradation at 37 and 60#{176}C,the relative change in free a-hCG and free $-hCG isexpressed in molar terms. Molarity was calculated fromthe established primary structures (20): 15 kDa for freea-subunit and 23 kDa for free p-subunit. At both tem-peratures the mean free a/free molar ratio increasewas 0.515: i.e., for every 1-mol increase in free a-, therewas a 2-mol increase in free /3-hCG.

The freeze/thaw study showed a 2.5% increase in freep-hCG after 10 freeze/thaw cycles, with an 11% increaseafter 25 freeze/thaw cycles. Therefore, prior thawingand refreezing of stored Down syndrome samples is

highly unlikely to influence free p-hCG concentrations

Table 2. Molar Increases In Free a- and Free $-hCGafter IncubatIon at Various Temperatures

Increase, nmol, after Incubation at

37’C forQ6h 60’Cfor5mln

Sampi. Frsea Free Frei a Fri p1 1.05 2.01 0.87 2.972 1.22 2.34 1.67 1.733 0.83 1.90 1.22 1.734 1.16 1.83 0.79 2.175 1.04 2.02 1.10 1.65Mean 1.06 2.02 1.10 2.17SD 0.149 0.196 0.360 0.5302 x SEM range 1.00-1.12 1.98-2.06 0.96-1.24 1.95-2.38

in these samples. However, we did observe some be-tween-sample variability in stability (range 1.46-4.74%at 10 freeze/thaw cycles).

In a large US screening program, the variability ofthe population median value for free p-hCG in samplestaking various lengths of time to reach the screeninglaboratory was small, with the regression line for re-sults over an 8-day period showing a 2.8% increase inthe median per day. The size of this change is consistentwith an experimentally derived 10% stability of 87 h atroom temperature (i.e., a 2.8% change in 24.36 h).

In the study of the impact of a 10% artifactualincrease in free p-hCG in samples from 685 consecutivescreened patients, we determined that an increase ofonly 1% (from 5.1% to 6.1%) in the initialpositive ratewould have been observed.

Study of the stability of free j3-hCG in dried whole-blood filter-paper spots compared with the same samplesstored as liquid whole blood showed that storage for 9days at 4#{176}Cdid not affect the stability in either mate-rial. At 25#{176}C,the liquid whole blood showed a twofoldincrease in free p-hCG over the 9 days, but that in theblood spot showed no increase. Even after storage at37#{176}Cfor 9 days, the dried blood-spot free p-hCG wasstifi stable, whereas the liquid whole blood showed a13-fold increase in free p-hCG concentration.

DIscussion

Our study of the stability of free p-hCG in serumindicates that, contrary to anecdotal citation (6), serumconcentrations of free p-hCG do not increase in samplesleft at ambient temperature for 30 mm. Our data athigh temperatures are consistent with the publisheddata of Bidart et al. (8), who also demonstrated at 56#{176}Cincreases in both free a- and free f3-hCG concentrations.By using assays that recognize both nicked and non-nicked free p-hCG, we have alsotaken into account anyaltered stability of either of these forms.

We attempted to clarif’ the source of the increase infree /3-hCG at high temperatures by also monitoringfree a-hCG. If the increased free p-hCG were occurringas a result of dissociation of intact hCG, then by observ-ing both free a- and free -hCG in samples exposed toaccelerated degradation at high temperatures, onewould expect to see a parallel mole-for-mole increase infree a- and free p-hCG. However, the experimental datashow that for each molecule of free a-subunit released,two molecules of free p-hCG are released. There areseveralpossibleexplanationsfor this observation. First,

at the high temperatures, in addition to a dissociation,conformational changes may be taking place in theintact molecule that expose the free epitope, so thatconformationally disturbed intact hCG is also beingmeasured as free p-hCG. Alternatively, released freea-subunit could be combining with p-subunits of otherpituitary glycoprotein hormones, thereby being redis-tributed. However, there is no evidence to suggest thateither of these mechanisms is taking place.

Although our experimental work on accelerated ther-mal degradation of intact hCG was carried out with

1068 CUNICAL CHEMISTRY, Vol. 39, No. 6, 1993

normal pregnancy samples, it is unlikely that thegreater concentrations of intact hCG seen in Downsyndrome pregnancies will influence the rate of anythermal degradation. Even if different carbohydrate-substituted forms of intact hCG had altered rates ofdegradation, to date there is no evidence to show thatsuch forms are produced in abnormal pregnancies.

Thawing and refreezing of stored Down’s samplesminimally influences free p-hCG concentrations. As

shown, the impact of freeze-thawing will not influencesignificantly initial positive rates or detection efficiencyin retrospective studies.

The laboratory experiments reveal the changethrough time in free p-hCG concentrations from theirinitial in vivo value to the value detected in an assay.This is not significant in a screening setting. Screeningresults rely on comparison of prospective samples withnormative data (when both data sets have similarchanges from in vivo to in vitro values); therefore, theimpact of instability is not significant. Furthermore,because free p-hCG is but one of several variables in ascreening protocol, even a 10% change in its value willresult in no more than a 1% increase in the initialpositive rate. Earlier studies (1-4) make it clear thatdetection efficiency in a screening protocol incorporatingfree p-hCG can surpass that achieved by any of theother reported markers.

Studies with whole blood show that after --72 h thedegradation of free p-hCG does not obey Arrheniuskinetics; as significant lysis takes place, the rate ofdegradation accelerates exponentially. In the short term(0-72 h), whole-blood stability is approximately halfthat of separated serum. For large-scale mass popula-tion screening in areas of the world where transportdelays and high temperatures preclude the speedy de-livery of whole blood to the laboratory, the use offilter-paper spot collection of whole blood will becomethe method of choice: in this medium, free p-hCG isstable at 37#{176}Cfor >9 days. a-Fetoprotein, anothermarker used in conjunction with free p-hCG, was alsostable after 9 days at 37#{176}C(data not shown). Filter-paper spot collection procedures also provide other ad-vantages, particularly patient/clinician convenience,low cost, and, more importantly, reduced potential fortransmission of disease in the collection, transportation,and laboratory handling of biological fluids.

By using standard sample collection and transporta-tion protocols that can avoid artifactual changes in allanalytes, the superior clinical utffity of the free p-hCG

marker in screening for Down syndrome can be fullyrealized.

References1. Macri JN, Kasturi RV, Krantz DA, Cook LI, Moore ND, YoungJA, et al. Maternal serum Down syndrome screening, free betaprotein is a more effective marker than human chorionic gonado-tropin. Am J Obstet Gynecol 1990;163:1248.-53.2. SpencerK. Evaluationof an assay of the free beta subunit ofchoriogonadotropun and its potential value in screening for Down’ssyndrome. Clin Chem 1991;37:809-14.3. Spencer K, Coombes LI, Mallard AS, Milford Ward A. Freebeta human choriogonadotropinin Down syndrome screenung amulticentre study of its role compared with other biochemicalmarkers. Ann Clin Biochem1992;29:506-184. SpencerK, MacriJN. Earlydetectionof Down syndrome usingfree betahuman choriogonadotropun. Ann Clin Biochem1992;29:349-50.5. SpencerK,Macri JN, Aitken DA, Connor MJ. Free beta (hCG)as a first trimester marker for fetal trisomy [Letter]. Lancet1992;339:1480.6. Knight GJ, Cole LA. Measurement of choriogonadotropin freebeta subunit: an alternative to choriogonadotropun in screening forfetal Down’ssyndrome[Editorial].CliiiChem 1991;37:779-82.7. Gau G, RiceA, Chard P. Increase of HCG levels followingheating of serum. J Obstet Gynecol 1984;5:21-3.8. BidartJ-M, Troalen F, Lazar V, Vauzelle P, BeIlet D. Pitfalls inmeasuring choriogonadotropin and its free subunits in decomple-mented sera [Letter]. Clin Chem 1991;37:1467-8.9. Kirkwood TEL. Predicting thestability of biological standardsand products.Biometrics 1977;33:736-42.10. Zakowski J. Determination of stability. Clin Chem 1991;27:313-4.11. Anderson G, ScottM. Determination of product shelf life andactivation energy for five drugs of abuse. Clin Chem 1991;27:398-402.12. Connors KA, Amidon GL, Kennon L. Chemical stability ofpharmaceuticals-a handbook for pharmacists. New York: JohnWiley, 1973:8-119.13. Winston S. Collection and preservation of specimens. In:MeitesS, ed.Standard methods of clinical chemistry, Vol.5. NewYork Academic Press, 1965:1-21.14. Baer DM, Krause RB. Spurious laboratory values resultingfrom simulated mailing conditions. A study of time and tempera-ture variables. Am J Clin Pathol 1968;50:1ll-9.15. Zilva JF. Collection and preservationof specimens for chemi-cal pathology.Br J Hoap Med 1970;4:845-52.16. Wilding P, Zilva JF, Wilde CE. Transportof specimens forclinical chemistry analysis. Ann Clin Biochem 1977;14:301.-6.17. Macri JN, Spencer K, Anderson 1W!, Cook LI. Free betachorionic gonadotropin: a cross reactivity study of two immuno-metric assays used in prenatal maternal serum screening forDown’s syndrome. Ann Clin Biochem1993;30:94-8.18. Kardana A, Cole LA. Polypeptide nicks cause erroneous re-sultsin assaysofhuman chorionicgonadotropin free beta subunit.Clin Chem 1992;38:26-33.19. Edwards PR.,Ekins RP. Mass action model based microproces-sor program for RIA data processing. In: Hunter WM, Corrie JET,eds. Immunoassay for clinical chemistry, 2nd ed. Edinburgh:Churchill Livingstone, 1983:640-52.20. BahI OP, Carlsen RB, Bellisario R, Swaminathan N. Humanchorionic gonadotropin; amino acid sequence of the alpha andbetasubunits. BiochemBiophysRes Commun 1972;48:416-22.