Disadvantages of VDRL syphilis screening 
Laboratory and clinica/ study 
LIMITATIONS AND CONSEQUENCES 
OF BASING LATE SYPHILIS 
SEROASSESSMENTS ON VDRL TEST 
V. Muic, l. Vodopija, M. Ljubičic and V. Mayer 
ABSTRACT 
The purpose of this paper is to caution about a recommendation made in 1996 by the 39th Meeting of 
the European Committee on Health, which operates within the Council of Europe. It prescribes a screening 
for the agents of major microbial infections, syphilis included, that endanger transplantation. This leaves an 
alternative in choosing between the VDRL (non-treponemal) and TPHA (treponemal) tests. With a 
retrospective analysis of the findings of our tests collected in a proficiency testing program, we have attempted 
to estimate the size of a potential diagnostic miss that could result from such recommendation. The results 
were obtained on 18,094 subjects (nonvenereal inpatients, psychiatric asylum patients, persons undergoing 
preemployment examination, persons seeking medica! certificates or international health documents, pregnant 
women screened for syphilis) who were tested with both tests simultaneously and came from two public health 
laboratories. A false negative VDRL ratio calculated from the 265 infected (i.e., TPHA-reactive) was defined 
as a diagnostic miss. It amounted to 84.5%. A statistical procedure derived from the Bayes theorem was used 
to assess the active late syphilis, yielding a value as high as 21.0% per 265 subjects infected, which was 
significantly more than the percentage suggested by the VDRL reactive test (15.5% ). These results suggest 
that, for the needs of transplantation, organ and tissue donors ( and ali venerological or nonvenerological 
subjects) should also undergo the screening for syphilis with the TPHA test as is usual in transfusiology. 
KEY WORDS 
VDRL, screening disadvantages, late syphilis detection model. 
INTRODUCTION 
Recommendations on the methods to use in 
serological screenings for the most prevalent microbial 
infections ( one being syphilis) linked with tissue and 
organ transplantation came from the 39th Meeting 
acta dermatovenerologica A.P.A. Vol 6, 97, No 3 
of the European Health Committee (CDSP) held in 
Strasbourg in June 1996 (1). This committee operates 
within the Council of Europe. Although the risk of 
syphilis transmission through transplantations may 
be small, positive reaction in a syphilis test is 
indicative of the donor's hazardous sex behavior. 
87 
Disadvantages of VDRL syphilis screening 
Table l . The value of the VDRL test's nosological 
sensitivity (SExJa based on the TPHA reference testb, 
as well as the VDRL false negative (FN) ratec were 
determined for a public health population oj 18,094 
subjectsd. 
VDRL FN rate 
15.5% 84.5% 
•SEx = (VDRL+ /TPHA+)• 100 
b Ali 265 TPHA reactors were considered to be 
infected with T. pallidum, ali nonreactors being 
uninfected . 
c In the parallel strategy (both tests done simulta-
neously) users, the VDRL FN rate represented a 
hypothetical miss in the detection of those infected 
with syphilis. Conversely, in serial strategy (VDRL 
used as a screening test and TPHA as a confirmation 
test), the VDRL FN rate constituted a genuine diagnostic 
miss. 
d Of the 18,094 subjects tested the 265 syphilis 
infected comprised almost exclusively old syphilis 
either late treated syphilis or untreated syphilis. 
Further, the cost of the test is minimal, it is the 
most widely used serotest, and one with which there 
is a long experience. Besides, it unveils the risk of 
other known sexually and parenterally transmitted 
diseases and of the stili unexplored diseases for 
which efficient screening techniques have not become 
routine yet. 
There is thus every justification to perform an 
effective screening for syphilis e.g. at STD clinics, 
public health dispensaries, in blood donors' centers. 
Such selection of tests differs from the one undertaken 
to detect patients suspected to be infected with 
( early) syphilis. 
In screenings for syphilis in addition to the TPHA 
hemagglutination test (or an alternative test) VDRL 
(non-treponemal antigen serologic test) should be 
used. The latter is known for its lower specificity 
and considerably lower sensitivity at later stages of 
the disease, a fact which is mostly disregarded or 
overlooked. 
A different practice derived from some older 
misunderstandings, is found to cause negative public 
88 
health effects as confirmed by the recent findings in 
a proficiency testing program carried out by the 
Croatian National Institute of Public Health from 
1993 to 1995 in our clinical, public health and 
blood transfusion laboratories which perform syphilis 
serotests (2). As in most European countries, a 
request for serological test in Croatia assumes that 
syphilis will be investigated at any stage. A problem 
may arise under the influence of the above meeting's 
conciusions (1) or the routine in the USA, where 
unless accompanied by other indications, a request 
for syphilis screening will result in the use of 
algorithm needed to detect the epidemiologically 
more dangerous and markedly infectious early syphilis 
(3). This was stressed by one of us (LV.) at the 
Strasbourg meeting. According to its conclusion VDRL 
is the only test or the first in a series with the 
TPHA test (screening with VDRL and confirming 
with TPHA only the VDRL positives). If a general 
practitioner in the US wants to look for latent 
syphilis or for late active syphilis, he is expected to 
specially emphasize this on the request form or add 
1 1 
SEuN 
SErn SEx 
o o .... o X 1 
1 o 
1-X 
Figure l . Graphic model of the detection of late active 
untreated syphilis. 
SEx - VDRL test sensitivity observed in our study 
SETR - VDRL test sensitivity in treated late syphilis 
SEuN - VDRL test sensitivity in patients with untreated 
late syphilis 
x - proportion of untreated late active syphilis among 
the infected found in the study 
Note. Alt terms are expressed in terms of proportion. 
acta dennatovenerologica A.P.A. Vol 6, 97, No 3 
Disadvantages of VDRL syphilis screening 
Table 2. Nosological sensitivity and specificity of the tests used in testing far syphilis, and some characteristics of 
the 18,094 population tested. 
Syphilis stages with Estimated age 
highest presumed of VDRL 
prevalence among and TPHA 
the study reactors 
population 
Untreated cases x=60.1 yrs. 
of latent and Range 37 
late syphilis to 80 yrs 
N=28 
Cases of late 
treated or 
of self-healed syphilis 
Screening 
tests used 
VDRL 
TPHA 
VDRL 
TPHA 
Nosological 
sensitivity (SE) 
of tests (%) 
Range (in brackets) 
of literature given 
values (3,4) 
SEuN=?0• 
(34-94) 
100 (97-100)C 
SETR=1b (0-1) 
100 (97-100)C 
Nosological 
specificity 
(SP) of tests 
(%) 
98 (96-99)c 
99 (98-100)° 
98 (96-99)C 
99 (98-1 00)c 
a SEuN - VDRL test sensitivity in untreated latent and late syphilis. 
b SErn - VDRL test sensitivity in treated late and self-healed syphilis. 
c For simplicity it was assumed that the above values equalled 100%. Also ignored was the 0.3% of the cases 
registered as isolated VDRL reactors (biological false positive reactions or early syphilis). 
such a remark to it. After that, the laboratory will 
employ other tests, i.e. those for antitreponemal 
antibody demonstration (mainly of the IgG class) 
such as the TPHA test, and the tests for the 
demonstration of process activity (19 S IgM tests) -
for the individuals whose infection had previously 
been demonstrated. 
As our old population is practically free from 
early syphilis the combination VDRL+ and TPHA+ 
primarily indicates late active untreated syphilis. 
MATERIALS AND METHODS 
Based on our proficiency testing program findings 
(2), we have attempted to assess the size of two 
potential misses: 
l. The miss due to interlaboratory differences in the 
effectiveness of detection of individuals infected 
with T pallidum at any tirne in their lives. Miss of 
this kind can be readily detected (and is demonstrable 
directly or indirectly) by using in addition the parallel 
testing strategy on the same or similar population. 
In this case both tests are done simultaneously on 
the same analytical sample. Mathematically, this is 
acta dennatovenerologica A.P.A. Vol 6, 97, No 3 
expressed as VDRL test's false negative rate in 
percentages as determined on a population of the 
TPHA positive patients (i.e., on the syphilis-infected 
in the course of their lives regardless of the stage 
of the disease or process activity) (Table 1). 
2. The second type of potential miss occurs when 
detecting late, and probably active, forms of syphilis. 
It was assumed that the VDRL test has a maximum 
specificity, but various sensitivity (ranging from 0.01 
to 0.70) which depends on the stage of the disease 
and its status of treated or untreated (3,4) (Table 
2). From this we proceeded to making the assessment 
using the statistical procedure derived from the 
Bayes theorem (5), which is based on a ratio (x) of 
two differences, i.e., the nosological sensitivity (SEx) 
observed in our study for VDRL, and the VDRL 
sensitivity already known from the literature (3) for 
persons treated for late or latent syphilis (SEm). Ali 
this is shown in the numerator. The second difference 
is that between the VDRL test's sensitivity to the 
persons untreated for latent and late syphilis (SEuN) 
and VDRL's sensitivity to the persons treatect for 
latent and late syphilis (SEm)· Both of these have 
been previously established by other workers ( 4). 
This difference is represented by the denominator. 
89 
Disadvantages of VDRL syphilis screening 
Table 3. The number of cases infected with T. pallidum who were detectable by means of a parallel (P), respectively 
serial (S) testing strategy of combining the VDRL and TPHA tests. 
Subject 
total tested 
No. of infected 
detected by strategy 
p 
No. of infected 
detected by strategy 
s 
The infected total 
detected by both strategies 
(S+P) 
18,098 265 41 265 
x Parallel strategy (P) involves simultaneous use of the VDRL and TPHA tests on a sample. 
In the serial strategy (S), only VDRL test reactors are additionally checked with the TPHA test. 
The thus calculated quotient (x), is the proportion 
of presumed late active but untreated syphilis among 
the infected. 
RESULTS AND DISCUSSION 
Graphically, a solution to "x" is illustrated in Fig. 
l. The error first mentioned, or a potential miss, 
was calculated from a public health laboratory's test 
findings on 18,094 subjects (without clinical or 
epidemiological data). Of the 265 infected whose T. 
pallidum infection was demonstrated using the TPHA 
test, 41 were simultaneously also reactive in the 
VDRL test. Consequently, we regarded them (because 
of their average age of 60.1 yrs; range 37 to 80; 
N = 28) as having late untreated syphilis, or as (not 
very likely) cases of recently treated but not yet 
seronegativized syphilis. Had the infected been detected 
by the often recommended serial strategy, which 
combines the VDRL and TPHA tests, and not with 
the parallel strategy, as few as 41 (15.5%) infected 
individuals would have been found instead of the 
265, with the miss rate amounting to 84.5% (Tables 
1,3). 
The second error or miss was assessed by taking 
the same material. It consisted in making a statistical 
assessment of the share of presumably late active 
syphilis in our sample of the 265 infected. On 
introducing in the afore-mentioned formula (5,6,7) 
an VDRL SEx of 15.5% previously noted in our 
study (Table 1), that of 70% (3) cited by most 
reports for the untreated with late syphilis (SEuN), 
and one of 1 % ( 4) for the treated cases (SETR) 
(Table 2), the assessment showed that the 265 
infected who had been detected included as many 
as 56 (21 % ) people affected by late active syphilis 
(Table 4). This is 15 persons more (56 instead of 
41, i.e. 21 % instead of 15.5% of 265 infected) than 
screening with the VDRL test only, would reveal 
(although the method does not permit an iden-
tification of affected individuals) (6,7). Moreover, 
the method enables calculations of the confidence 
Table 4. The results of a retrospective analysis of a screening far syphilis covering 18,094 public health subjects: 
different numbers of late active syphilis suspects under standard interpretation and according to the Bayes theorem. 
90 
Subject total 
tested 
18,094 
No. of infected 
detected 
256 
No. of late 
active syphilis 
(VDRL+ and TPHA+) 
41 
Estimate of the number of late 
active syphilis cases based on 
Bayes theorem 
56 
acta dennatovenerologica A.P.A. Vol 6, 97, No 3 
Disadvantages of VDRL syphilis screening 
limits of an assessment, ignored here for the sake 
of clarity. 
Both diagnostic misunderstandings may be avoided 
by planning the screening with regard to the type of 
test(s), the strategy of combining them and by the 
presumed prevalence of individual stages of the 
syphilitic process in a patient population. 
ACKNOWLEDGMENT 
CONCLUSIONS 
An attempt has been made to evaluate a diagnostic 
miss in detecting old ( and probably also active) 
syphilis in the routine use of VDRL and TPHA 
tests (VDRL test being the first in serial testing) . 
Our data confirm the size of the miss to be speci-
fically dependent on the mode of combining the two 
tests. By using an original statistical procedure we 
have uncovered a probably additional diagnostic miss. 
The authors would like to thank Victoria Pope, PhD, Chief, Treponemal Pathogenesis and Immunology Branch, Division of AIDS, STD 
and TB Laboratory Research National Center far Infectious Diseases, CDC, Atlanta, USA for valuable advice helping in the completion 
of this work. 
They also thank Vitim Crlenjak, BA, from the Croatian National Institute of Public Health for the translation. 
REFERENCES 
l. Anonymous, Serological screening methods far the 
most relevant microbiological diseases of organ and 
tissue donors. European Health Committee, Council of 
Europe, 39th Meeting, Strasbourg, 25-27 lune 1996, 
CDSP (96) 21, p.5. 
2. Muic V, Vodopija I Serotesting far syphilis, diagnostic 
proftciency and health care quality surveillance (in 
Croatian). Liječ. Vjesn 1996; 118: 184-6. 
3. Larsen S, Steiner E, Rudolph A. Laboratory diagnosis 
and interpretation of tests far syphilis. Clin Microbiol 
Rev 1995; 8: 1-21. 
4. Tramont E. Treponema pallidum (Syphilis). In: 
Mandel! GL, Benett JE, Dolin R, eds. Mandel!, Douglas 
and Benett's Principles of Infectious Diseases. New 
York, NY: Churchill Livingstone, 1995: 2117-33. 
5. Muic V, Petres JI, Telisman Z. Validity of a 
diagnostic test designated by a single function. Meth 
Inform Med 1973; 12: 244-8. 
6. Bayes T, An assay toward solving a problem in the 
doctrine of chance. Philo Trans Roy Soc 1763; 53: 
370-418, in: Galen RS. Beyond Normality. New York: 
Wiley, 1975, 123-5. 
7. Deželic Gj, Deželic N, Jakšic Z, Muic V, Vuletic S. 
The prevalence of rheumatoid factors in the sera of 
open rural populations and its epidemiological significance 
(in Croatian), Liječ Vjesn 1978; 100: 695-9. 
AUTHORS' ADDRESSES 
Vladimir Muic, MD, PhD, consultant, Croatian National Institute of Public Health, 10000 Zagreb, 
Rockefellerova 7, Croatia. 
Ivan Vodopija, MD, professor and consultant, Zagreb Public Health Institute, Mirogojska 16, 
10000 Zagreb, Croatia 
Mate Ljubičic, MD, PhD, director, Croatian National Institute of Public Health, 10000 Zagreb, 
Rockefellerova 7, Croatia. 
Vladimir Mayer, MD, PhD, Head Health Care Planning and Programming Dept., Croatian National 
Institute of Public Health, 10000 Zagreb, Rockefellerova 7, Croatia. 
acta dennatovenerologica A.P.A. Vol 6, 97, No 3 91