JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1987, p. 1711-17160095-1137/87/091711-06$02.00/0Copyright © 1987, American Society for Microbiology

Enzyme-Linked Immunosorbent Assay for Detection ofAntibodies to the Venereal Disease Research Laboratory (VDRL)

Antigen in SyphilisNILS STRANDBERG PEDERSEN,1* OLE 0RUM,' AND S0REN MOURITSEN2

World Health Organization Collaborating Centre for Reference and Research on Treponematoses, Department ofTreponematoses,' and Autoimmune Laboratory,2 Statens Seruminstitut, DK-2300 Copenhagen S, Denmark

Received 9 March 1987/Accepted 27 May 1987

An enzyme-linked immunosorbent assay (ELISA) for detection of immunoglobulin G (IgG) and IgM tocardiolipin, lecithin, and cholesterol (VDRL [Venereal Disease Research Laboratory] ELISA) is described. Thespecificity of the VDRL ELISA for IgG and IgM was 99.6 and 99.5%, respectively, with sera from 1,008persons without syphilis. For a group of patients with false-positive results in traditional nontreponemal testsand for patients with autoimmune diseases, the VDRL ELISA for IgG had a higher specificity than the VDRLELISA for IgM. The sensitivity for IgG and IgM with 118 sera from patients with untreated syphilis was 96.6and 94.9%, respectively, which was equivalent to the sensitivities of the traditional nontreponemal tests. Theperformance of the VDRL ELISA was compared with that of an ELISA that uses cardiolipin as the antigen(cardiolipin ELISA). The VDRL ELISA was significantly more sensitive (P s 0.01) than the cardiolipin ELISAwith 25 sera from syphilis patients but was less sensitive (P - 0.01) with 53 sera from patients with autoimmunediseases. The antibody reactivity in the VDRL ELISA could not be absorbed out by lecithin and cholesterol,and the sera from patients with syphilis did not react in an ELISA that uses cholesterol and lecithin as theantigen. This indicates that cholesterol and lecithin, although not antigenic by themselves, may change thestructural form of the epitope on cardiolipin so that it becomes more recognizable for antibodies in syphilis andless recognizable for antibodies in autoimmune diseases. The results of the VDRL ELISA were expressed inpercentages of the absorbance value of a positive control. The VDRL ELISA gave, without titration of sera,quantitative results that correlated with the quantitative results of the traditional nontreponemal tests obtainedby titration. The VDRL ELISA will be well suited for large-scale testing for syphilis and may replace othernontreponemal tests.

Serological tests for syphilis can be divided in two majorgroups, the nontreponemal and the treponemal tests. Thenontreponemal tests or standard tests for syphilis detectantibodies against the phospholipid hapten cardiolipin, incomplex with lecithin and cholesterol. This antigen complexis used either in a complement fixation test (Wassermanntest) or in agglutination tests such as the Venereal DiseaseResearch Laboratory (VDRL) test, the automated reagintest (ART), or the rapid plasma reagin test (19). The sero-logical diagnosis of syphilis still depends to a large extent onthe results of a nontreponemal test, supplemented whennecessary by a confirmatory treponemal antibody test, i.e.,the fluorescent treponemal antibody absorption (FTA-ABS)test (19) or the Treponema pallidum hemagglutination test(13). The nontreponemal tests are extremely valuable forscreening purposes and may have even higher specificitiesfor syphilis than the treponemal tests (10, 15). The trepo-nemal tests should not be used as screening tests, but only as

confirmatory tests on highly selected serum samples frompatients who are suspected of a present or previoustreponemal infection (3, 15). Furthermore, whereas a

nontreponemal test generally becomes negative within 6 to12 months after treatment of early syphilis, the treponemaltests may remain positive for many years (2, 16). Quantita-tive nontreponemal tests are therefore useful both for

* Corresponding author.

posttreatment monitoring and for the detection of reinfec-tions (10).The nontreponemal tests presently used are not ideal.

None of the tests used today is able to distinguish betweenimmunoglobulin G (IgG) and IgM without previous separa-tion of serum. Detection of IgM is important for the diagno-sis of congenital syphilis, because IgM does not cross theintact placenta, and IgM in newborns indicates active syph-ilis. The Wassermann test used in Europe requires freshcomplement, which makes the test troublesome. The VDRLtest requires fresh and accurate preparation of antigen andmicroscopic visualization of agglutination. The ART and therapid plasma reagin test are easier to perform, but interpre-tation of weak-positive results may often be difficult. False-negative results may occur in the VDRL test, the ART, andthe rapid plasma reagin test when strongly positive sera are

tested, because of prozone phenomena (18). Recently a

radioimmunosorbent assay (7) and an enzyme-linked im-munosorbent assay (ELISA) (11) have been described forthe detection of anti-cardiolipin antibodies. These tests havenot been evaluated as syphilis diagnostic tests but have beenused to detect anti-cardiolipin antibodies in patients withautoimmune diseases, i.e., systemic lupus erythematosus(SLE).

In the present study, we have developed an ELISA fordetermination of IgG and IgM against the VDRL antigen.The sensitivity and specificity of this method were compared

1711

Vol. 25, No. 9

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from

1712 PEDERSEN ET AL.

with those of a cardiolipin ELISA, a cardiolipin Wasser-mann (CWRM) test, an ART, and an FTA-ABS test.

MATERIALS AND METHODS

Lipid antigens. Cardiolipin was purified by the method ofPangborn (12), and lecithin was purified by the method ofRiis Nielsen (14). Cholesterol was obtained from PfanstiehlLab, Waukegan, 111. The purity of all lipids was tested bythin-layer chromatography.

Conjugates. The conjugates used were horseradish perox-

idase-conjugated rabbit anti-human IgG (code P214, lot 091)and anti-human IgM (code P215, lot 0.25A), both fromDAKO-Immunoglobulins, Copenhagen, Denmark.VDRL ELISA for IgG and IgM. Each well of flat-bottom

polyvinyl chloride microtiter plates (Titertek, catalog no.

77-173-05; Flow Laboratories, Irvine, Scotland) was coatedwith 25 .1t of an ethanol solution containing 0.003% (wt/vol)cardiolipin, 0.09% (wt/vol) cholesterol, and 0.021% (wt/vol)lecithin (the VDRL antigen diluted 10 times in absoluteethanol). The solvent was evaporated at 4°C overnight. Theplates were washed three times with phosphate-bufferedsaline (PBS), pH 7.4, and then coated with 100 iil of 10%

(vol/vol) bovine serum in PBS (PBS-BS) at room tempera-

ture for 1 h. Before use, the bovine serum was tested andfound negative by the CWRM test and the ART. The wellswere aspirated and washed as described above, and 100 ,uI ofthe serum to be tested was diluted 1:100 in PBS-BS andadded to each of two wells. A positive and a negative controlserum were included in every experiment. The positivecontrol was obtained from a syphilis patient and had opticaldensities of 1.2 and 1.9 in the ELISAs for IgG and IgM,respectively. The negative control serum was obtained froma blood donor. The optical density values for this were 0.1and 0.15 in the ELISAs for IgG and IgM, respectively. Afterincubation for 2 h at room temperature, the plates were

aspirated and washed, and 100 jil of conjugate diluted inPBS-BS was added. Incubation took place for 2 h at room

temperature, the plates were then aspirated and washed, and200 1tl of substrate (41 mg of o-phenylenediamine [SigmaChemical Co., St. Louis, Mo.], 100 ml of citrate buffer [pH5], 25 Fd of 30% H202) was added to each well. Afterincubation for 15 min at room temperature, the reaction was

stopped by the addition of 50 ,ul of 3 M H2SO4 to each well.The A492 was read with a photometer (ImmunoReader NJ2000; Nippon, InterMed, Tokyo, Japan).

Cardiolipin ELISA and lecithin-cholesterol ELISA. TheELISA for antibodies against cardiolipin was done as de-scribed for the VDRL ELISA, but the plates were coatedwith 0.0045% (wt/vol) cardiolipin. In the lecithin-cholesterolELISA, the plates were coated with 0.09% (wt/vol) choles-terol and 0.02% (wt/vol) lecithin.

Absorption studies. As a measure of antigen specificity inELISA, four sera from patients with syphilis were absorbedas follows. One milliliter of 0.03% (wt/vol) cardiolipin-0.9%(wt/vol) cholesterol-0.21% (wt/vol) lecithin was added dropby drop to 0.8 ml of PBS, pH 6.0 (VDRL buffer), withstirring. The total volume was then adjusted to 10 ml withVDRL buffer. This antigen suspension (1 ml) was centri-fuged at 12,000 x g for 15 min. The pellet was suspended inPBS-BS, incubated for 15 min on a shaker, and centrifugedat 12,000 x g for 15 min. The pellet was then suspended in 1ml of the serum to be tested, diluted 1:100 in PBS-BS, andallowed to react for 15 min at room temperature on a shaker.It was then centrifuged at 12,000 x g for 15 min to remove

antigen and immune complexes and subsequently membrane

filtered through a 0.22-,um-pore-size filter (Millex-GS; Milli-pore, Molsheim, France). The absorbed serum was thentested as described above.

Routine syphilis serological methods. All sera were testedby the ART (19) and, if positive, quantified by the rapidplasma reagin test (19). All sera were also tested by theCWRM test as modified by Môrch (17), and sera fromsyphilis patients were also analyzed by the FTA-ABS test.The conjugate used in the FTA-ABS test was a rabbit anti-human IgG, IgA, IgM (code F1009; DAKO-Immunoglobulins).

Diagnoses of syphilis. The diagnoses of syphilis were madeon the basis of extensive centralized, serological, and clini-cal data collected at the Department of Treponematoses,Statens Seruminstitut, Copenhagen.

Sera. Sera were chosen from specimens submitted forroutine syphilis serological analysis during a 4-month periodin 1986, except for 25 samples from patients with secondarysyphilis (group 2), which were collected in 1980 and stored at-20°C, and 500 samples from blood donors, which werecollected in 1982. None of the sera tested was heat inacti-vated.Group 1: patients with untreated primary syphilis. There

were 46 patients with primary untreated syphilis. All patientshad lesions suspected ofbeing chancres, and all had sera thatwere reactive in the FTA-ABS test.Group 2: patients with untreated secondary syphilis. There

were 32 patients with untreated secondary syphilis. Allpatients had exanthema, and all gave sera that were reactivein the CWRM, ART, and FTA-ABS tests.Group 3: patients with untreated late syphilis. There were

five patients with untreated neurosyphilis. All patients hadcerebrospinal fluid that gave reactive results in the CWRMand FTA-ABS tests.Group 4: patients with untreated serologically diagnosed

early syphilis. There were 35 patients with serologicallydiagnosed syphilis. All patients were positive in the CWRM,ART, and FTA-ABS tests. None of the patients had clinicalfindings indicating syphilis, and because of a rapid serolog-ical response to treatment the diagnosis of early syphilis wasmade.Group 5: patients with treated syphilis. There were 146

patients with treated syphilis. The stages at the time ofdiagnosis differed, and both early and late syphilis caseswere represented. Six patients were CWRM positive andART negative. A total of 31 patients were CWRM negativeand ART positive, 91 patients were both CWRM and ARTpositive, and 24 patients were negative in both tests. All serawere reactive in the FTA-ABS test.Group 6: persons whose sera gave false-positive results.

There were 86 patients with no history or signs of syphiliswhose sera gave positive results in either the CWRM test orthe ART and negative results in the FTA-ABS test. Re-peated tests of sera from these persons confirmed that thereactivity in the previous samples was not caused by tech-nical errors or syphilis. Of the patients, 57 had sera that werepositive in the CWRM test and negative in the ART, 24 hadsera that were negative in the CWRM test and positive in theART, and 5 had sera that were positive in both the CWRMtest and the ART.Group 7: persons without syphilis (blood donors). There

were 500 sera from unpaid voluntary blood donors withoutsyphilis. All sera were nonreactive in the CWRM test andthe ART, and none of the persons in this group had a historyof past or present syphilis.Group 8: patients without syphilis. There were 508 sera

from 343 patients from venereal disease clinics, 85 hospital

J. CLIN. MICROBIOL.

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from

VDRL ELISA FOR SYPHILIS 1713

TABLE 1. Results of syphilis serological tests of sera from patients with different stages of syphilis and from controls

No. (%) of positive sera by:

Patient group No. of sera VDRL ELISA CWRM FTA-ABS~~~~~~~~~~~~estART t sIgG IgMtetes

1, Untreated primary syphilis 46 43 (93.5) 40 (87.0) 39 (84.8) 41 (89.1) 46 (100.0)2, Untreated secondary syphilis 32 32 (100.0) 32 (100.0) 32 (100.0) 32 (100.0) 32 (100.0)3, Untreated tertiary syphilis 5 5 (100.0) 5 (100.0) 5 (100.0) 5 (100.0) 5 (100.0)4, Untreated serologically diagnosed syphilis 35 34 (97.1) 35 (100.0) 35 (100.0) 35 (100.0) 35 (100.0)5a, Treated syphilis, CWRM positive, ART negative 6 1 (16.7) 0 (0.0) 6 (100.0) 0 (0.0) 6 (100.0)Sb, Treated syphilis, CWRM negative, ART positive 31 21 (67.7) 14 (45.2) 0 (0.0) 31 (100.0) 31 (100.0)Sc, Treated syphilis, CWRM and ART positive 91 88 (96.7) 71 (78.0) 91 (100.0) 91 (100.0) 91 (100.0)Sd, Treated syphilis, CWRM and ART negative 24 3 (17.5) 1 (4.2) 0 (0.0) 0 (0.0) 24 (100.0)6a, False-positive reactors, CWRM positive 57 16 (28.1) 32 (56.1) 57 (100.0) 0 (0.0) 0 (0.0)6b, False-positive reactors, ART positive 24 5 (20.8) 11 (45.8) 0 (0.0) 24 (100.0) 0 (0.0)6c, False-positive reactors, CWRM and ART positive 5 4 (80.0) 4 (80.0) 5 (100.0) 5 (100.0) 0 (0.0)7, No syphilis, blood donors 500 1 (0.2) 3 (0.6) 0 (0.0) 0 (0.0) Not

done8, No syphilis, patients 508 3 (0.6) 2 (0.4) 0 (0.0) 0 (0.0) Not

done

VDRL ELISA-positive sera were tested and found to be negative in the FTA-ABS test.

patients, and 80 patients consulting general practitioners.The sera were all tested by the CWRM test and the ART andfound negative. None of the patients had a history of pastsyphilis.Group 9: persons with rheumatoid factor-positive sera. We

randomly selected 30 sera among specimens found to bepositive for rheumatoid factor by the Laboratory forAutoimmune Diseases, Statens Seruminstitut. The sera con-tained rheumatoid factors in medium and high titers as

determined by ELISA (20).Group 10: patients with autoimmune diseases. A total of 53

sera from patients with autoimmune diseases were obtainedfrom various clinical hospital departments. Twenty-eightsera were from patients with SLE. Four sera were frompatients with progressive systemic sclerosis, nine sera were

from patients with primary biliary cirrhosis, three sera were

from patients with mixed connective-tissue disease, and ninesera were from patients with polydermatomyositis.

Statistical analysis. The sign test and the Spearman testwere used for statistical analyses.

Specificity and sensitivity of the test. The specificity wasdefined as the percentage of unaffected individuals whosesera were nonreactive in the test in relation to the number ofunaffected individuals in the population tested. The sensitiv-ity was defined as the percentage of syphilis patients whosesera were reactive in the test in relation to the number ofsyphilis patients in the population tested.

RESULTSCutoif point, specificity, and precision of the VDRL ELISA

for IgG and IgM. The cutoff values of the VDRL ELISA forIgG and IgM were determined by assaying 500 serum sam-

ples from blood donors (group 7). All sera had been previ-ously tested and found negative in the CWRM test and theART. The test results were expressed in percentages of theabsorbance value of a positive control. When a cutoff of 20was chosen for the VDRL ELISA for both IgG and IgM, 1 ofthe 500 blood donor sera was found positive in the VDRLELISA for IgG, with a value of 24, and 3 sera were foundpositive in the VDRL ELISA for IgM with values of 22, 24,and 27 (Table 1). All of the latter three sera were nonreactivein the FTA-ABS test.A total of 508 sera (group 8) submitted from venereal

disease clinics, hospital departments, and general practition-

ers for syphilis serology were tested in the VDRL ELISA.None of the sera came from patients with past or presenthistories of syphilis. All sera were found negative in theCWRM test and the ART. Three sera were reactive in theVDRL ELISA for IgG with values of 24, 27, and 33. Anothertwo sera were found positive in the VDRL ELISA for IgM,both having the value of 34. None of these five sera wasreactive in the FTA-ABS test. The specificity of the VDRLELISA for IgG was therefore 99.4%, and the specificity ofthe VDRL ELISA for IgM was 99.6% (Table 1).We also tested 30 sera containing rheumatoid factors

(group 9). One of these sera was strongly reactive in theCWRM test, and none was reactive in the ART. None ofthese was reactive in the VDRL ELISA for IgG but threewere reactive in the VDRL ELISA for IgM with values of37, 42, and 44. One of the IgM-positive sera was the onereactive in CWRM.The total assay precision of the VDRL ELISA was

determined by assaying two negative and three positive serain duplicates on 10 different days (Table 2).The VDRL antigen is composed of cardiolipin, lecithin,

and cholesterol, and to study the specificity of the VDRLELISA we absorbed four sera from syphilis patients eitherwith complete VDRL antigen or with cholesterol-lecithinbefore testing. Only the complete VDRL antigen was able toblock the ELISA reactivity (Table 3). Furthermore, micro-titer plates coated with cholesterol-lecithin were not able tobind either IgG or IgM in sera from controls or syphilispatients (data not shown).

TABLE 2. Total assay precision of the VDRL ELISA for IgGand IgM as determined by testing two negative and three

positive sera in 10 independent assays

Mean ELISA value (SD)Serum

IgG IgM

Negative1 2.7 (1.0) 1.9 (0.9)2 4.3 (1.3) 4.1 (1.0)

Positive3 43.7 (5.7) 38.1 (5.4)4 101.9 (12.7) 66.4 (9.7)5 149.9 (13.5) 109.2 (13.9)

VOL. 25, 1987

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from

1714 PEDERSEN ET AL.

TABLE 3. Specificity of the VDRL ELISA as shown by absorption of four sera with eithercardiolipin-lecithin-cholesterol (VDRL antigen) or lecithin-cholesterol

Residual VDRL ELISA reactivity in %c of ELISA values of nonabsorbed samples'Absorption A B C D Mean (SD)antigen

IgG IgM IgG IgM IgG IgM IgG IgM IgG IgM

VDRL antigen 5 5 10 il 10 4 9 5 8.8 (2.5) 6.3 (2.5)Lecithin-cholesterol 80 80 106 73 89 71 99 90 93.5 (11.4) 78.5 (8.6)

" Specimens A. B, C. and D showed nonabsorbed IgG ELISA values of 147, 124, 149. and 100 and IgM ELISA values of 196, 76, 160. and 100, respectively.

Comparison of the VDRL ELISA with the cardiolipinELISA. A panel of sera from patients with secondary syph-ilis or different autoimmune diseases was assayed in parallelby the VDRL ELISA and the cardiolipin ELISA for bothIgG and IgM (Table 4). The VDRL ELISA for both IgG andIgM was more sensitive with a group of sera from patientswith secondary syphilis than the cardiolipin ELISA (P -

0.01). On the other hand, the cardiolipin ELISA for both IgGand IgM was more sensitive than the VDRL ELISA for IgG(P < 0.05) and IgM (P < 0.01) with sera from patients withSLE. The cardiolipin ELISA was also more sensitive thanthe VDRL ELISA with the whole group of sera frompatients with the different autoimmune diseases (P - 0.01).Comparison of the VDRL ELISA with the CWRM test and

the ART. The results of the syphilis serological tests areshown in Table 1. A total of 270 sera from syphilis patientswas tested. Of the sera from patients with untreated primarysyphilis (group 1), 43 (93.5%) were reactive in the VDRLELISA for IgG and 40 (87.0%) were reactive for IgM. Noneof the sera was reactive only for IgM, and the combined useof VDRL ELISA for IgG and IgM compared with the use ofthe VDRL ELISA for IgG alone did not increase thesensitivity. The differences in sensitivity of the nontrepone-mal serological tests with sera from patients with primarysyphilis were not statistically significant. For untreatedsecondary syphilis (group 2) and untreated late syphilis(group 3), all tests had a sensitivity of 100%. In the group ofpatients with serologically diagnosed untreated early syphilis(group 4), the patients were selected on the basis of positiveCWRM, ART, and FTA-ABS test results. All of these serawere reactive in the VDRL ELISA for IgM, and all but onewere positive in the VDRL ELISA for IgG.The distribution of the ELISA values determined by the

VDRL ELISA for IgG is shown in Fig. 1. The ELISA valuesin syphilis range widely, from the cutoff of 20 to as high as210. The quantitative results of VDRL ELISA for both IgGand IgM showed an excellent correlation with the quantita-

tive results of both the CWRM test and the ART (Fig. 2;Table 5).

In the group of patients with treated syphilis whose serawere positive in both the CWRM test and the ART (group5c), the VDRL ELISA for IgG antibodies had a highersensitivity than the ELISA for IgM antibodies (P < 0.01).The differences in sensitivity between the ELISA for IgGand the CWRM test and the ART were not significant.

In the group of 57 patients with false-positive CWRMresults (group 6a), the VDRL ELISA for IgG had a higherspecificity than the VDRL ELISA for IgM (P < 0.05). Onlyfour (7%) of the sera in this group were positive for both IgGand IgM antibodies in the VDRL ELISA. Also, in the groupof 24 patients with false-positive ART results (group 6b), thespecificity of the VDRL ELISA for IgG was higher than thatfor IgM (P < 0.05). All patients with IgG in this group had IgM.

VDRL- ELISA% of positive control

220

200

180

160

140

120

100

80

60

40

TABLE 4. Comparison of cardiolipin ELISA results and VDRLELISA results for samples from patients with

syphilis and autoimmune diseases

No. of sera positive in:

Disease No. of Cardiolipin VDRLsera ELISA ELISA

IgG IgM IgG IgM

Syphilis 25 5 15 25 25SLE 28 7 25 1 10Progressive systemic sclerosis 4 0 1 0 0Primary biliary cirrhosis 9 1 9 0 4Mixed connective-tissue disease 3 t) 3 O 1Polydermatomyositis 9 1 9 0 3

20

1510.

5-

rl

:3

.3.ù'.

*

4

*

*:1

*VIL3Qi

.-dg. N 185 N 1i6

N277 N 302

1 2 3 4 5a 5b 5c 5d 7 8

FIG. 1. Distribution of VDRL ELISA lgG values in differentpopulations. Sera with values above the dashed line were regardedas positive. Groups: 1, patients with untreated primary syphilis; 2,patients with untreated secondary syphilis; 3, patients with un-treated late syphilis; 4, patients with untreated serologically diag-nosed syphilis; 5. patients with treated syphilis (a, CWRM positive,ART negative; b. CWRM negative, ART positive; c, CWRM andART positive; d, CWRM and ART negative); 7, persons withoutsyphilis (blood donors); 8, patients without syphilis.

J CI-I N. M ICROB IOL

S

:

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from

VDRL ELISA FOR SYPHILIS 1715

VDRL-ELISA% of positive control

200-

180-

160-

140-

120-

100-

80-

60-

40-

20-

O i 2 3 i 5 6 i 8 b10 l 14 15 1

CWRM degree of strength

FIG. 2. Relation between the quantitative results of the VDRLELISA for IgG and the CWRM test with 265 sera from patients withuntreated syphilis (Spearman's coefficient of correlation [r] = 0.882,t = 30.35). A total of 116 sera came from patients with untreatedsyphilis (r = 0.763, t = 12.61), and 149 came from patients withtreated syphilis (r = 0.824, t = 17.62).

DISCUSSIONThe results of this study indicate that the VDRL ELISA is

suitable for the diagnosis of syphilis and may be an alterna-tive to traditional nontreponemal tests. The sensitivities ofthe VDRL ELISA for IgG and IgM with the 118 sera frompatients with untreated syphilis were 96.6 and 94.9%, re-

spectively. These sensitivities were comparable with thoseof the traditional nontreponemal tests. The specificities ofthe VDRL ELISA as tested on 1,008 sera from personswithout syphilis (groups 7 and 8) were 99.6% for IgG and99.5% for IgM. In the group of patients with false-positivenontreponemal tests (group 6; Table 1) the VDRL ELISA for

TABLE 5. Correlation between quantitative results ofnontreponemal tests with 265 sera from patients

with untreated and treated syphilis

Spearman'sNontreponemal coefficient of t valuetests compared variation (r)

VDRL ELISA IgG and VDRL ELISA IgM 0.755 18.81VDRL ELISA IgG and CWRM test 0.882 30.35VDRL ELISA IgM and CWRM test 0.800 21.60VDRL ELISA IgG and ART 0.740 17.95VDRL ELISA IgM and ART 0.863 27.94ART and CWRM test 0.790 20.93

IgG had a higher specificity than the VDRL ELISA for IgM.The same result was found in sera from patients withautoimmune diseases (Table 4). Of 30 sera containing rheu-matoid factors, 3 were found positive in the VDRL ELISAfor IgM and none was positive in the VDRL ELISA for IgG.False-positive reactions in the traditional nontreponemaltests have previously proved to be predominantly an IgMresponse (1), and the present study confirms this observa-tion. For routine use, we therefore prefer the VDRL ELISAfor IgG; however, the VDRL ELISA for IgM may be usefulin posttreatment monitoring of patients and when congenitalsyphilis is suspected in a newborn child.The quantitative results of traditional nontreponemal tests

such as the CWRM test and the ART are based on titrationof serum samples. The VDRL ELISA, however, was quan-titative without further dilution of serum, and the quantita-tive results showed an excellent correlation with the quan-titative results of the CWRM test and the ART (Table 5; Fig.2). The discrimination between positive and negative resultswas good in the VDRL ELISA for both IgG and IgM. Theratio between the median ELISA values obtained with serafrom patients with untreated syphilis and sera from controlswas 25 (positive-negative ratio).The performance ofthe VDRL ELISA was compared with

that of an ELISA that uses cardiolipin as the antigen.Anti-cardiolipin antibodies have proved to be significantlyassociated with the lupus anticoagulant, thrombocytopenia,spontaneous abortions, and thrombosis in SLE patients (4,6-9, 11, 21). The cardiolipin ELISA is now used in severallaboratories to define this high risk subset of SLE patients.Although the cardiolipin ELISA from a theoretical point ofview would be ideal as a syphilis serological method, it hasnot been used as such.The results of the present study show that the VDRL

ELISA was significantly more sensitive than a cardiolipinELISA with sera from syphilis patients (Table 4). On thecontrary, the cardiolipin ELISA was more sensitive than theVDRL ELISA with sera from patients with autoimmunediseases. We showed that the antibody reactivity in theVDRL ELISA could not be absorbed out by lecithin andcholesterol (Table 3), and sera from patients with syphilis didnot react in an ELISA that uses cholesterol and lecithin asthe antigen. This indicates that cholesterol and lecithin,although not antigenic by themselves, may change the struc-tural form of the epitope on cardiolipin, so that it becomesmore recognizable for antibodies in syphilis and less recog-nizable for antibodies in SLE. Costello and Green (5), usinga different assay principle, also found that lecithin, althoughnot antigenic, increased the sensitivity of cardiolipin withregard to detection of antibodies in syphilis. Koike et al. (11)and Colaco and Male (4) did not find any correlation betweenanti-VDRL antibody titer and anti-cardiolipin antibody titerin patients with SLE. Colaco and Male (4) also demonstratedsignificantly different profiles of reactivity with syphiliticsera and SLE sera against different phospholipid antigens.These results indicate that antibodies in syphilis and SLE aredirected against different but cross-reactive phospholipidepitopes and are consistent with our findings.The VDRL ELISA may substitute traditional nontrepone-

mal tests, especially in laboratories using the ELISA princi-ple for other tests. The VDRL ELISA is especially useful forlarge-scale testing, i.e., in laboratories undertaking blooddonor, prenatal, and premarital screening for syphilis, and itwould also be useful in laboratories receiving a large numberof samples from venereal disease clinics. The VDRL ELISAhas other advantages. It is quantitative, even without further

VOL. 25, 1987

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from

1716 PEDERSEN ET AL.

dilution of serum, and the use of an automatic plate readereliminates the subjectivity of visual reading.

ACKNOWLEDGMENTS

Kirsten Lindboe is thanked for perfect technical assistance.Severin Olesen Larsen is thanked for statistical assistance.

LITERATURE CITED

1. Aho, K. 1968. Studies of syphilitic antibodies. Il. Substancesresponsible for biological false positive sero-reactions. Br. J.Vener. Dis. 44:49-53.

2. Brown, S. T., A. Zaidi, S. A. Larsen, and G. H. Reynolds. 1984.Serological response to syphilis treatment-a new analysis ofold data. J. Am. Med. Assoc. 253:1296-1299.

3. Center for Disease Control. 1973. Technical problems withFTA-ABS test for syphilis-Virginia. Morbid. Mortal. WeeklyRep. 22:102.

4. Colaco, C. B., and D. K. Male. 1985. Anti-phospholipid antibod-ies in syphilis and a thrombotic subset of SLE: distinct profilesof epitope specificity. Clin. Exp. Immunol. 59:449-456.

5. Costello, P. B., and F. A. Green. 1986. Reactivity patterns ofhuman anticardiolipin and other antiphospholipid antibodies insyphilitic sera. Infect. Immun. 51:771-775.

6. Englert, H. J., C. H. Hawkes, M. L. Boey, G. J. M. Derue, S.Loizou, E. N. Harris, A. E. Gharavi, R. G. Hull, and G. R. V.Hughes. 1984. Degos' disease: association with anticardiolipinantibodies and the lupus anticoagulant. Br. Med. J. 289:576.

7. Harris, E. N., M. L. Boey, A. E. Gharavi, B. M. Patel, C. G.Mackworthyoung, S. Loizon, and G. R. V. Hughes. 1983.Anticardiolipin antibodies: detection by radioimmunoassay andassociation with thrombosis in systemic lupus erythematosus.Lancet ii:1211-1214.

8. Harris, E. N., A. E. Gharavi, R. A. Asherson, M. L. Boey, andG. R. V. Hughes. 1984. Cerebral infarction in systemic lupus:association with anti-cardiolipin antibodies. Clin. Exp. Rheuma-tol. 2:47-51.

9. Harris, E. N., A. E. Gharavi, U. Hedge, G. Derue, S. H. Morgan,

H. Englert, J. K. H. Chan, R. A. Asherson, and G. R. V. Hughes.1985. Anticardiolipin antibodies in autoimmune thrombocy-topenic purpura. Br. J. Haematol. 59:231-234.

10. Hart, G. 1986. Syphilis tests in diagnostic and therapeuticdecision making. Ann. intern. Med. 104:368-376.

11. Koike, T., H. Sueishi, H. Funaki, H. Tomioka, and S. Yoshida.1984. Anti-phospholipid antibodies and biological false positiveserological test for syphilis in patients with systemic lupuserythematosus. Clin. Exp. Immunol. 56:193-199.

12. Pangborn, M. C., J. O. Almeida, F. Maltaner, A. M. Silverstein,and W. R. Thompson. 1955. Cardiolipin antigens. Monographseries, no. 6. World Health Organization, Geneva.

13. Ratlev, T. 1967. Haemagglutination test utilizing pathogenicTreponeina palliditin for the sero-diagnosis of syphilis. Br. J.Vener. Dis. 43:181-185.

14. Riis Nielsen, J. 1980. A simple chromatographic method forpurification of egg lecithin. Lipids 15:481-484.

15. Sand Petersen, C. 1986. Biological false positive reactions intreponemal serological tests used to diagnose syphilis. Genito-urin. Med. 62:356.

16. Sand Petersen, C., B. B. J0rgensen, and N. Strandberg Pedersen.1984. Treatment of early infectious syphilis in Denmark. Dan.Med. Bull. 31:70-72.

17. Schmidt, H. 1951. Cardiolipin antigen. 1. A quantitative exami-nation of sensitivity and specificity. Br. J. Vener. Dis. 27:23-33.

18. Spangler, A. S., J. H. Jackson, N. J. Fiumara, and T. A.Warthin. 1964. Syphilis with a negative blood test reaction. J.Am. Med. Assoc. 189:87-90.

19. U.S. Department of Health, Education, and Welfare. 1969.Manual tests for syphilis. Publication no. 411. U.S. GovernmentPrinting Office, Washington, D.C.

20. Vejtorp, M., M. H0ier Madsen, and P. Halberg. 1979. Enzymelinked immunosorbent assay for determination of IgM rheuma-toid factor. Scand. J. Rheumatol. 8:65-70.

21. Williams, R. C., Jr. 1984. Cerebral infarction in systemic lupus:association with anticardiolipin antibodies. Clin. Exp. Rheu-matol. 2:3-4.

J. CLIN. MICROBIOL.

on May 2, 2019 by guest

http://jcm.asm

.org/D

ownloaded from