Vol. 13, No. 4JOURNAL OF CLINICAL MICROBIOLOGY, Apr. 1981, p. 698-7040095-1 137/81/040698-07$02.00/0

Diagnosis of Recent Rubella Virus Infection by Demonstrationof Specific Immunoglobulin M Antibodies: Comparison ofSolid-Phase Reverse Immunosorbent Test with Sucrose

Density Gradient CentrifugationG. A. DENOYEL,I* A. GASPAR,' AND D. PEYRAMOND2

Service des Virus, Institut Pasteur de Lyon, 69365 Lyons Cedex 2,' and Infectious Diseases Division, Hopitalde la Croix Rousse, 69004 Lyons Cedex 2,2 FranceReceived 23 June 1980/Accepted 24 October 1980

A solid-phase reverse immunosorbent test (SPRIST) based on the addition ofan excess of rubella virus hemagglutinin was evaluated for the demonstration ofrubella-specific immunoglobulin M (IgM), and the results were compared withthose of the density gradient centrifugation technique. In a retrospective study inwhich 157 sera were tested, the two techniques yielded identical results (55 IgM-positive and 102 IgM-negative samples). In a prospective study, 592 sera wereexamined; 8 IgM-positive results by SPRIST corresponded to a recent rubellainfection or vaccination. Neither rheumatoid factor nor heterophil antibodyseemed to interfere with the results of SPRIST. This test would be a useful andrapid routine technique for demonstration of the presence of virus-specific IgM inserum samples, particularly for viruses with a hemagglutinin. Except for anti-human IgM, no more reagents are needed than for widely used hemagglutinationinhibition procedures.

A useful approach for detection of a recentviral infection is the demonstration of virus-spe-cific antibody of the immunoglobulin M (IgM)class. The separation ofIgM from IgG by sucrosedensity gradient centrifugation (DGC) (3, 17) orgel filtration (4) is useful, but these techniquesare cumbersome. Another approach is to detectvirus-specific IgM by using labeled anti-humanIgM in immunofluorescence (3), radioimmu-noassay (6), or enzyme immunoassay (EIA) (9).Unfortunately, the specificity of these tests canbe impaired by the presence of rheumatoid fac-tor (7, 14). Furthermore, the sensitivity of im-munofluorescence can be reduced by competi-tion for antigen between IgM and IgG, whichare usually both present in the serum sample.Another rapid method involves the treatment ofserum by staphylococcal protein A, which re-moves about 95% of IgG (1). However, protein Afails to remove the IgG3 subclass and removes10 to 15% of IgM (15). More recently, the use ofsolid-phase immunosorbents was extended tothe separation of IgM from whole serum immu-noglobulins (2). In this laboratory, we evaluatedfor the detection of rubella-specific IgM anti-body a solid-phase reverse immunosorbent test(SPRIST) based on the addition of rubella he-magglutinin (HA). In the present report, the

results are compared with those of the sucroseDGC method.

MATERIALS AND METHODSSera. (i) Retrospective study. A total of 157 sera

which were received during the last 3 years for rubellaIgM antibody testing by the sucrose DGC techniquewere examined. Among them, 47 sera were drawn from28 subjects (20 pregnant women, 4 newborns, and 4children) with clinical symptoms of rubella infection(rash, involvement of lymph nodes, arthralgia, or con-genital rubella). A total of 19 paired sera were avail-able, 12 of which demonstrated seroconversions, and7 of which exhibited high stable titers (>1:640). Theremaining 9 sera were single, with antibody titers of>1:320. All sera were collected from day 1 to day 32after the clinical onset. Eight sera were obtained fromtwo previously seronegative young women who re-ceived a recent rubella vaccination (RA27/3 Rudivax;Merieux, France). For the first vaccinated woman,sera were collected at days 10, 22, 38, and 85, and forthe second, sera were collected at days 10, 28, 64, and98 postimmunization. A total of 102 sera were collectedfrom pregnant women without any symptoms of ru-bella but who demonstrated moderate to high rubellaantibody titers (1:160 to 1:2,560).

In addition, 24 sera with positive rheumatoid factor(Waaler-Rose test positive, ranging in titer between 1:64 and 1:2,048), 15 sera from patients with infectiousmononucleosis (Paul-Bunnell-Davidsohn test posi-tive), and 20 sera from individuals with a recent mea-sles infection (measles-specific IgM positive) were also

698

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

DIAGNOSIS OF RUBELLA INFECTION 699

assayed.All sera were aseptically dispensed and stored at

-60°C before testing.(ii) Prospective study. A total of 592 sera which

were received in the laboratory for routine rubellaantibody determinations were assayed. The hemagglu-tination inhibition (HI) antibody titers ranged from<1:10 to 1:2,560.HI test (16). All rubella antibody titers indicated

above were obtained by HI. Nonspecific inhibitorswere removed by treatment with kaolin in boratebuffer (pH 9) (12), and the heteroagglutinins wereabsorbed with a 50% adult chicken erythrocyte (RBC)suspension. A micromethod was performed by using0.05 ml of serum dilution (starting dilution, 1:10), 4 Uof HA (Behringwerke, Marburg, Germany), and a0.20% 1-day-old chicken RBC suspension.

Fractionation of serum immunoglobulin bysucrose DGC. For separating IgM from IgG, a DGCtechnique was used. Briefly, 0.3 ml of serum previouslyabsorbed with chicken RBC (resulting in a dilution of1:2) was layered onto a 4.7-ml 12 to 33% sucrosegradient in phosphate-buffered saline (pH 7.2). Aftercentrifugation at 35,000 rpm for 18 h (BeckmanSW50.1 rotor), 12 0.3-mi fractions were collected bypuncturing the bottom of the tube, checked for thepresence of IgM or IgG by using radial immunodiffu-sion (Tri-Partigen plates, Behringwerke), and testedfor HI antibody. Rubella IgM antibody was consideredpositive only if the HI titer was -1:4 and the IgM-containing fraction was separated by at least one frac-tion from IgG-containing fractions. DGC titers wereexpressed as the reciprocal dilution of the fractiongiving the maximum HI reactivity.2-ME treatment. Sera and DGC fractions were

sometimes treated as described by Scott and Gershon(13). Briefly, a 0.1 M solution of 2-mercaptoethanol(2-ME) (Eastman Organic Corp., Rochester, N.Y.) wasprepared in phosphate-buffered saline, and the pHwas adjusted to 7.2. Then 0.1 ml of a serum samplepreviously absorbed with chicken RBC was mixedwith 0.1 ml of 2-ME solution and incubated at 37°Cfor 1 h before the HI test or SPRIST. This procedurewas previously shown to reduce only the antibody titerof IgM-containing sera or fractions.SPRIST. (i) Coating of plates. Standard polysty-

rene round-bottomed microplates were used through-out this work. Cooke microtiter M24A plates (Dyna-tech, Rungis, France) were selected because of theiracceptable performance in EIA tests as previouslydetermined in the laboratory.A 0.1-ml amount of dilutions of rabbit anti-human

IgM (u-chain specific; Dako Immunoglobulins, Copen-hagen, Denmark) or goat anti-human IgM (CappelLaboratories, Inc., Cochranville, Pa.) in either carbon-ate buffer (pH 9.6) or TSA buffer (0.05 Mtris (hydroxymethyl) aminomethane - hydrochloride,pH 7.4, 0.15 M NaCl, and 0.03% NaN3) were added tothe wells. The plates were held at 37°C in a moistchamber for 16 h and thoroughly washed with phos-phate-buffered saline containing 0.05% Tween 20, andthen the wells were filled (0.3 ml) with phosphate-buffered saline containing 0.5% bovine serum albuminfraction V. The plates were incubated at 37°C for 2 h,

the liquid was removed, and the plates were ready foruse.

(ài) Pretitration and testing ofthe specificity ofanti-human IgM used for solid-phase coating.Although the specificity of all labeled or nonlabeledanti-human IgM or IgG used in this laboratory wassatisfactory as determined by immunoelectrophoresis,an additional control experiment was performed asfollows. Plates were coated as described above withincreasing dilutions (1:200 to 1:12,800) of anti-humanIgM (12 wells per dilution). Then 0.1 ml of humanserum dilutions (1:20 to 1:20,480) in phosphate-buffered saline-0.05% Tween 20 and 1% bovine serumalbumin were added to each dilution of anti-humanIgM, except for one row, which served as controls.After 2 h at 37°C and extensive washing, the wellswere filled with 0.1 ml of either peroxidase-labeledanti-human IgM or IgG (Cappel Laboratories) at anoptimal dilution as determined in EIA, i.e., 1:300 and1:500, respectively. After 2 h at 37°C and washing, theenzyme activity was detected by a color change in asolution containing H202 and 5-aminohydroxybenzoicacid (11). Readings were made after 30 min at 450 nmwith a Multiscan apparatus (Flow Laboratories, Inc.,Rockville, Md).

(iii) Determination of the optimal concentra-tion of anti-IgM coating antibody, rubella HA,and chicken RBC suspension. Based upon resultsof the specificity test described above, the plates werecoated with various dilutions of anti-IgM (1:200 to 1:3,200). After extensive washing, sera (previously ab-sorbed with chicken RBC) were diluted from 1:20 to1:20,480, and 0.1 ml of each dilution was added to thesensitized wells. For standardization of the test, onerubella IgM-positive serum and one negative serum,as demonstrated by DGC, were used. After 2 h at37°C, the plates were washed. A 0.05-ml amount of arubella HA dilution in borate-bovine serum albuminbuffer saline (pH 9) and 0.1 ml of chicken RBC sus-pension in phosphate-buffered saline (pH 6.2) weresuccessively added to each well. Increasing concentra-tions of HA (1, 2, 4, 6, and 8 U) and RBC suspension(0.20,0.40,0.60,0.80, and 1%) were evaluated. A serumcontrol (1:20 serum dilution without HA and withRBC) and a control of HA activity (dilutions of HAwithout serum and with RBC) were included. Aftergentle shaking at 4°C for 30 min, the plates were heldat 20°C until the RBC were completely settled in theserum controls. Two patterns were observed: a stronghemagglutination with a clumping RBC which flowedwhen the plate was tilted (this was considered to

FIG. 1. SPRIST results. (A) Positive pattern; (B)negative pattern.

VOL. 13, 1981

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

700 DENOYEL, GASPAR, AND PEYRAMOND

indicate an IgM-negative sample) and a hemadsorp-tion with strong adherence of a homogeneous sheet ofRBC onto the bottom of the well, without any slidingwhen the plate was tilted (this was considered toindicate an IgM-positive sample (Fig. 1).

RESULTSAssay conditions. Different variables were

first evaluated to insure the validity of the test.(i) Coating procedure. By using an EIA

method as described above, the influence of twodifferent buffers on the anti-IgM coating proce-dure was examined (Fig. 2). It appeared that thecoating was slightly better with the carbonatebuffer (pH 9.6) than with the TSA buffer (pH7.4). Furthermore, the reproducibility of thecoating was increased when the alkaline bufferwas used. The influence of the concentration ofthe coating antibody was evaluated by usingincreasing dilutions of anti-IgM. Dilutions be-tween 1:200 and 1:800 gave approximately thesame results, since IgM was detected in thehuman serum sample up to a dilution as high as1:20,480. The specificity of the anti-IgM was

0.5

0,~0.4 *.

E -

< 0.3 "

go0n-.

demonstrated when a peroxidase-labeled anti-IgG was added, resulting in a colored reactionno different from that of the background. Theanti-human IgM's were from two different com-mercial sources, and they gave approximatelythe same results.

(ii) Determination of optimal concentra-tions of anti-IgM, rubella HA, and RBC sus-pension. The best results, i.e., the maximumpattern difference between the IgM-positive andIgM-negative samples, were obtained by using 6to 8 HA U and a 0.60% RBC suspension. Thereaction was uninterpretable with only 1 U ofHA, indicating that an excess of HA was re-quired. The same IgM antibody titer (1:10,240)was obtained for the rubella IgM-positive sam-ple with 1:200 to 1:800 dilutions of coating anti-IgM, whereas a negative result (<1:20) was ob-tained for the rubella IgM-negative serum. A 1:500 dilution of coating anti-IgM was selected forfurther tests from the standpoint of economyand maximal reactivity. Based upon the resultsof the specificity test with EIA, the lowest serum

20 40 80 160 320 640 1280 2560 5120 10240 20480

SERUM DILUTION

FIG. 2. Evaluation of the coatingprocedure with an EIA system withperoxidase-labeled anti-human IgM.( ) Carbonate buffer, pH 9.6; (-----) TSA buffer, pH 7.4. The coating anti-IgM antibody was used atdilution of 1:200 (O), 1:400 (A), 1:800 (-), and 1:1,600 (A). (a and b) Absorbance values were obtained byusing a peroxidase-labeled anti-human IgG. The coating anti-human IgM was used at a dilution of 1:200.

J. CLIN. MICROBIOL.

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

VOL. 13, 1981

TABLE 1. Specificity ofSPRIST; comparativeresults obtained by SPRIST and HI for DGCfractions from a rubella IgM-positive serum

Radial im-munodif- Rubella antibody tier

DGC fusionfrac-tion SPRIST

IgM IgG SPRIST after 2- HI HI afterME 2M

1 + - >640 10 8 <22 + - >640 10 32 <23 + - >640 <5 16 <24 + - 320 <5 8 <25 + - 160 <5 4 <26 tr tr 80 <5 8 47 - + 40 <5 32 648 - + 20 <5 64 649 - + <5 <5 64 6410 - + <5 <5 128 128

IgM + Gradient

FIG. 3. SPRIST on DGC fractions of a rubellaIgM-positive serum. Fractions I to 8 were positive.CF, Control fraction (without hemagglutinin).

dilution which was considered as positive was 1:40.

(iii) Additional specificity determina-tions. Rubella IgM-positive and IgM-negativesera were fractionated by using DGC, and the

DIAGNOSIS OF RUBELLA INFECTION 701

fractions were tested by SPRIST. Positive re-sults were obtained only with the IgM-contain-ing fractions as determined by radial immuno-diffusion, except for fractions 7 and 8. Wheneach fraction was treated by 2-ME before per-forming SPRIST, the antibody titer of IgM-con-taining fractions and also of fractions 7 and 8was abolished or considerably reduced (Table 1)(Fig. 3).

(iv) Reproducibility of the test. RubellaIgM-positive and IgM-negative sera were ex-amined by SPRIST each day during a 10-dayperiod. Although the same titer (1:5,120) wasconsistently obtained with the positive sample,we occasionally found that certain plates failedto regularly retain the coating anti-IgM sincegaps were observed through the serial dilutionsof serum (for example, 1:40 to 1:3,20, positive; 1:640, negative; and 1:1,280, positive). The nega-tive serum was always found to be negative.Retrospective study. Table 2 shows results

obtained when 47 sera drawn from individualswith recent rubella infections were tested. Ru-bella IgM-positive results were found by usingSPRIST whenever positive results were dem-onstrated by DGC. The IgM antibody titersdemonstrated by SPRIST were far higher (20-to 1,000-fold) than those obtained with DGC,and there was not an evident correlation be-tween the antibody titers obtained by the twomethods since a SPRIST titer of 1:10,240 cor-responded to DGC titers of from 1:8 to 1:64.Furthermore, low-positive SPRIST titers (1:40)were found, although DGC titers were negative(<1:4) (Table 2, sera 17 and 29), and low- tomiddle-positive titers were demonstrated bySPRIST when no HI antibody (<1:10) was de-tected (sera 1, 3, 7, 17, and 29). Ail of the positiveresults obtained by SPRIST were negative afterthe sera were treated with 2-ME. With the 102sera drawn from women who had no symptomsof recent rubella infection and were DGC nega-tive, no positive results were demonstrated withSPRIST (titer, <1:20).Table 3 shows the comparative results ob-

tained when sera collected from vaccinatedwomen were tested. The presence of specificIgM was demonstrated by both methods, andthe duration ofIgM after the rubella vaccinationwas identical, except for the last serum (day 85)of patient 1, which was DGC negative and ex-hibited a low titer (1:80) by SPRIST.Rheumatoid factor, heterophil antibody, or

measles IgM antibody did not seem to interferein the rubella SPRIST since all sera from thesegroups were found negative. They were alsonegative when rubella IgM detection was per-formed by DGC.

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

TABLE 2. Retrospective study; comparative results obtained by DGC and SPRIST for 47 sera from 28patients with recent rubella infection

Days HI an- IgM antibody ti- Days HI an- IgM antibody ti-

Patient Serum clnia tibody ter Patient Serum clinical tibodyonset DGC SPRIST onset DGC SPRIST

1 1 3 <10 NDa 401I2 12 320 16 5,120

2 3 3 <10 ND 804 18 1,280 32 10,240

3 5 5 40 4 1606 15 2,560 32 10,240

4 7 2 <10 ND 808 12 320 16 10,240

5 9 4 20 4 64010 13 1,280 64 20,480

6 il 3 10 ND 16012 15 1,280 16 5,120

7 13 6 10 4 16014 18 640 64 10,240

8 15 1 <10 <4 <2016 16 640 16 10,240

9 17 2 <10 <4 4018 15 640 64 5,120

10 19 4 40 4 64020 17 1,280 32 5,120

il 21 3 10 4 8022 14 1,280 32 2,560

12 23 4 20 4 32024 18 2,560 16 10,240

13 25 4 80 8 1,28026 13 320 16 10,240

a ND, Not done.b NewbornS with congenital rubella.

14 27 5 40 16 64028 17 640 32 5,120

15 29 1 <10 <4 4030 16 320 16 1,280

16b 31 3 1,280 32 5,12032 20 1,280 16 2,560

17b 33 2 1,280 8 10,24034 28 2,560 32 5,120

18b 35 10 640 16 5,12036 30 2,560 8 5,120

19b 37 2 1,280 32 1,28038 10 2,560 8 1,280

20 39 18 320 8 10,240

21 40 35 2,560 8 5,120

22 41 24 320 16 2,560

23 42 12 320 8 2,560

24 43 16 1,280 32 5,120

25 44 18 5,120 16 5,120

26 45 22 2,560 16 2,560

27 46 31 1,280 8 5,120

28 47 26 1,280 16 5,120

Prospective study. Of 592 sera which were at random and tested by DGC for rubella IgM,received for routine rubella HI antibody deter- with negative (<1:4) results (Table 4).mination and were prospectively examined forrubella IgM by SPRIST, 8 were found positive, DISCUSSIONand the positive titers were abolished by 2-ME The technique of capturing IgM by anti-hu-treatment. When tested by DGC, seven sera man IgM adsorbed to the solid phase has beenwere found positive, and one serum contained evaluated by several authors (2, 10). Krech andonly a trace of IgM (1:2). After obtaining clinical Wilhelm (5) previously described a SPRIST forinformation, it appeared that six patients suf- the rapid demonstration of rubella IgM. A 1-Ufered from a recent infection resembling rubella, amount of HA was used, and HI was observedone was recently vaccinated for rubella, and one for the IgM-positive samples. Our test seemedwas asymptomatic. Among the remaining 584 to be rather different. Positive sera caused ad-sera found negative by SPRIST, 72 were taken sorption of RBC onto the bottom of the well,

702 DENOYEL,' GASPAR, AND PEYRAMOND J. CLIN. MICROBIOL.

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

DIAGNOSIS OF RUBELLA INFECTION 703

TABLE 3. Comparative durations of specific rubellaIgM antibodies as demonstrated by using DGC and

SPRIST after rubella vaccination

Day after Rubella antibody titer

Patient vaccina-Ig gtion HI SPRIST DGC

1 -a <10 <20 <410 20 160 422 80 1,280 838 80 640 485 40 40 <4

2 - <10 <20 <410 10 320 428 40 640 1664 160 640 898 40 <20 <4

a Before vaccination.

and an HA excess was needed since poor resultswere obtained with only 1 U of HA. The firststep of the reaction probably was the coating ofRBC by an excess of rubella HA, followed bythe specific binding of the HA-RBC complexonto the virus-specific IgM coupled with anti-IgM in the solid phase. The positive pattern wasspecifically correlated with the presence of IgMsince it was observed only in the IgM-containingfractions of sera fractionated by DGC. The dis-crepancy in fractions 7 and 8 (Table 1) could beexplained by the greater sensitivity of SPRISTfor detecting IgM as compared with radial im-munodiffusion or DGC. In addition, the positivereactions were abolished by 2-ME treatment.The lowest serum dilution which was consideredas positive was 1:40 based upon the results ofspecificity tests on the coating anti-IgM withEIA.

In our experience, the presence of rheumatoidfactor or heterophil antibody did not seem tointerfere in SPRIST. Although encouraging,

these results need further confirmation with a

greater number of samples.The IgM antibody titers detected with

SPRIST were 20- to 1,000-fold higher than thoseobtained with DGC. Furthermore, low SPRIST-positive titers were detected in a few cases whenHI or DGC titers were negative. Since the posi-tive reactions regularly disappeared after 2-MEtreatment, they probably resulted from the in-creased sensitivity of SPRIST.Concerning the duration of specific rubella

IgM, DGC and SPRIST gave comparable resultsfor the two vaccinated women. Unfortunately,we were not able to test successive sera after a

naturally acquired infection. It is possible thatan occasional prolonged IgM response might beobserved (8). In our experience, we have seen no

such results which would hamper the value ofthe method for determining a recent infection,since all positive results were obtained with sera

taken no more than 60 days after infection and98 days after vaccination. Nevertheless, furtherstudy with sera taken later after the onset ofrubella is required to confirm this point.From the above data, it appeared that

no false-positive results were obtained withSPRIST. The major drawback of the techniquewas the possible occurrence of false-negativeresults due to the occasional lack of fixation ofthe anti-IgM onto the plastic wells. Although weselected plates which gave optimal results inEIA, we occasionally experienced this pitfallwhen testing the reproducibility of the method.Consequently, it is of essential importance tofind a reliable technique for fixing anti-IgM, andthis point deserves further investigation. Pres-ently, it seems possible to minimize this possibledrawback by testing several dilutions of eachserum in duplicate by using two different plateswith a positive serum control on each plate.

TABLE 4. Prospective study; comparative results obtained by SPRIST and DGC for 592 sera received in thelaboratory for routine rubella HI antibody determinationIgM IgDC Day after Patient

Reaction SPRIST igMterC HI antibody titer Clinical symptoms ()agetiter titr ose

IgM positive 2,560 16 640 Rash 30-35 42,560 8 1,280 Rash, arthralgia 15-20 12640 8 640 Mild fever, rash 30-35 3320 8 640 Rash 12-15 18320 4 2,560 Fugace rash 40-50 4.5320 4 1,280 Asymptomatic 2.5160 2 2,560 Rash? >60 3160 8 160 Vaccination 22 24

IgM negative -a <4b <10-2,560a _, The remaining 584 sera were negative by SPRIST.b Negative for 72 out of 72 samples selected at random from among those negative by SPRIST (see the text).

VOL. 13, 1981

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

704 DENOYEL, GASPAR, AND PEYRAMOND

In our hands, SPRIST seemed to be a simple,rapid, and reliable routine screening techniquefor detecting rubella-specific IgM. However, it isessential to carefully determine the ability ofplates to bind the anti-IgM, the specificity ofthis reagent, and the optimal dilution for use.These variables drastically influence the validityof the test. Taking these precautions into consid-eration, SPRIST should work for other virus-specific IgM antibodies and particularly viruseswith HA. Except for the anti-IgM, no additionalreagents are needed in addition to those usuallyused in the HI test.

ACKNOWLEDGMENTS

We thank A. Coffy, A. C. Malanca, and F. Serbonnet forexcellent technical assistance and G. Joly for photographicwork.

LITERATURE CITED

1. Ankerst, J., P. Christensen, L. Kjellen, and G. Kron-vall. 1974. A routine diagnostic test for IgA and IgMantibodies to rubella virus: absorption of IgG withStaphylococcus aureus. J. Infect. Dis. 130:268-273.

2. Duermeyer, W., F. Wielaard, and J. Van Der Veen.1979. A new principle for the detection of specific IgMantibodies applied in an ELISA for hepatitis A. J. Med.Virol. 4:25-32.

3. Forghani, B., N. J. Schmidt, and E. H. Lennette. 1973.Demonstration of rubella IgM antibody by indirectfluorescent antibody staining, sucrose density gradientcentrifugation and mercaptoethanol reduction. Intervi-rology 1:48-59.

4. Gupta, J. D., V. Peterson, M. Stout, and A. M. Mur-phy. 1971. Single-sample diagnosis of recent rubella byfractionation of antibody on Sephadex G-200 column.J. Clin. Pathol. 24:547-550.

5. Krech, U., and J. A. Wilhelm. 1979. A solid-phaseimmunosorbent technique for the rapid detection ofrubella IgM by haemagglutination inhibition. J. Gen.Virol. 44:281-286.

6. Meurman, O. H., M. K. Vilijanen, and K. Granfors.

1977. Solid-phase radioimmunoassay of rubella virusimmunoglobulin M antibodies: comparison with sucrosedensity gradient centrifugation test. J. Clin. Microbiol.5:257-262.

7. Meurman, O. H., and B. R. Ziola. 1978. IgM-class rheu-matoid factor interference in the solid-phase radioim-munoassay of rubella-specific IgM antibodies. J. Clin.Pathol. 31:483-487.

8. Pattison, J. R., D. S. Dane, and J. E. Mace. 1975.Persistence of specific IgM after natural infection withrubella virus. Lancet i:185-187.

9. Prevot, J., and J. L. Guesdon. 1977. Titrage immunoen-zymatique des anticorps IgG et IgM spécifiques de larubéole. Ann. Microbiol. (Inst. Pasteur) 128:531-535.

10. Roggendorf, M., G. G. Frosner, F. Deinhardt, and R.Scheid. 1980. Comparison of solid-phase test systemsfor demonstrating antibodies against hepatitis A virus(anti-HAV) of the IgM-class. J. Med. Virol. 5:47-62.

11. Ruitenberg, E. J., P. A. Steerenberg, J. M. Brossi,and J. Buys. 1974. Serodiagnosis of Trichinella spir-alis infections in pigs by enzyme-linked immunosorbentassay. Bull. W.H.O. 51:100-109.

12. Schmidt, N. J., and E. H. Lennette. 1970. Variables inthe rubella hemagglutination-inhibition test system andtheir effects on antigen and antibody titers. Appl. Mi-crobiol. 19:491-504.

13. Scott, D. W., and R. K. Gershon. 1970. Determinationof total and mercaptoethanol-resistant antibody in thesame serum sample. Clin. Exp. Immunol. 6:313-316.

14. Shirodaria, P. V., K. B. Fraser, and F. Stanford. 1973.Secondary fluorescent staining of virus antigens byrheumatoid factor and fluorescein-conjugated anti-IgM.Ann. Rheum. Dis. 32:53-57.

15. Skaug, K., and E. TMotta. 1974. Diagnosis of recentherpes simplex infections. A modified immunofluores-cent test for the detection of specific herpes simplexIgM antibodies after staphylococcal adsorption of IgG.Acta Pathol. Microbiol. Scand. Sect. B. 82:323-328.

16. Stewart, G. L., P. D. Parkman, H. E. Hopps, R. D.Douglas, J. P. Hamilton, and H. M. Meyer, Jr. 1967.Rubella virus hemagglutination-inhibition test. N. Engl.J. Med. 276:554-557.

17. Vesikari, T., and A. Vaheri. 1968. Rubella. A methodfor rapid diagnosis of a recent infection by demonstra-tion of the IgM antibodies. Br. Med. J. 1:221-223.

J. CLIN. MICROBIOL.

on April 5, 2020 by guest

http://jcm.asm

.org/D

ownloaded from