ORIGINAL STUDY
Criteria for the Diagnosis of Neurosyphilis in Cerebrospinal Fluid: Relationships With Intrathecal Immunoglobulin Synthesis and BloodYCerebrospinal Fluid Barrier Dysfunction Nadezhda Levchik, MD, PhD,* Marina Ponomareva, MS,* Vera Surganova, MD, PhD,Þ Natalia Zilberberg, MD, PhD,þ and Nikolai Kungurov, MD, PhDþ
Background: The origin of cerebrospinal fluid (CSF) syphilis antibodies (intrathecal or blood-derived) is in doubt. Little is known about CSF test behavior under the condition of physiological or disturbed functioning of blood-CSF barrier (BCB) and intrathecal immunoglobulin (Ig) production. Methods: We collected 126 serum/CSF pairs from patients with serological evidence of syphilis. We explored the relationships between the established facts of intrathecal Ig synthesis and/or BCB dysfunction and the results of CSF diagnostic tests: the Treponema pallidum hemagglutination (TPHA) test, the fluorescent treponemal antibody absorption (FTA-Abs) test, the Venereal Disease Research Laboratory (VDRL) test, and white blood cell counts. We checked the criteria used either to support or refute the diagnosis of neurosyphilis. Results: Reactive CSF-VDRL tests, elevated CSFYwhite blood cell counts, and elevated CSF-TPHA titers/indices were associated with the signs of intrathecal Ig synthesis, whereas nonreactive CSFYfluorescent treponemal antibody absorption, nonreactive CSF-TPHA tests, and CSF-TPHA titers from 1:4 to 1:160 were associated with cases where the intrathecal synthesis was not detected. There were some peculiarities of the tests toward BCB dysfunction. Most of reactive CSF-VDRL test samples and CSF samples with pleocytosis were also meeting at least 1 of the CSF-TPHA titer/indicesbased criteria. T. pallidum hemagglutination indices were in no better conformity with the facts of intrathecal immune response than CSFTPHA titers. Conclusions: Our findings have shown that all the examined criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity and support the hypothesis that high CSF treponemal-specific antibody titers are a consequence of inflammatory pathology of the central nervous system.
T
he laboratory verification of the central nervous system (CNS) involvement in syphilis constitutes a problem. The purpose of cerebrospinal fluid (CSF) analysis in the neurosyphilis
From the Departments of *Laboratory Medicine, †Venereology, and ‡Science, Urals Institute of Dermatovenereology and Immunopathology, Yekaterinburg, Russia
Conflict of interest: None declared. Funding: None declared. Correspondence: Nadezhda Levchik, MD, PhD, Department of Laboratory Medicine, Urals Institute of Dermatovenereology and Immunopathology, Shcherbakova 8, 620076 Yekaterinburg, Russia. E-mail: [email protected]. Received for publication April 26, 2013, and accepted September 20, 2013. DOI: 10.1097/OLQ.0000000000000049 Copyright * 2013 American Sexually Transmitted Diseases Association All rights reserved.
Sexually Transmitted Diseases
&
diagnosis is to detect pathogen-driven intrathecal immune response. Cerebrospinal fluid pleocytosis is a standard marker of active inflammation within CNS, but it is nonspecific and low sensitive. The other diagnostic tools are CSF nontreponemal and treponemal tests that identify antiphospholipid and antitreponemal antibodies, respectively. Cerebrospinal fluid syphilitic antibodies may have a dual origin. On the one hand, small amounts of immunoglobulins (Igs) are constantly supplied by the blood; on the other hand, Igs may be synthesized by plasma cells of perivascular inflammatory infiltrates in response to CNS infection. Only the latter ones have a diagnostic value. However, it is theoretically possible that an increase in blood-derived Ig concentration in CSF will lead to misdiagnosing intrathecal synthesis. The increase may be caused by high blood antibody level and/or the so-called blood-CSF barrier (BCB) dysfunction (a disorder of blood-CNS barrier integrity and/or impairment of CSF circulation). It is assumed that CSF antiphospholipid antibodies have intrathecal origin because their blood level is low. Reactive CSF nontreponemal tests are considered highly specific, and one of them, the Venereal Disease Research Laboratory (VDRL) test, is currently the ‘‘gold standard’’ for neurosyphilis.1,2 However, the gold standard is not perfect because a significant part of patients, who were clinically defined as having neurosyphilis, had a nonreactive CSF-VDRL test. The intrathecal origin of CSF antitreponemal antibodies is questionable because their blood level is considerable, and it has been shown that there is a relationship between the CSF test reactivity and a blood antibody titer.3 Some have suggested that the level of CSF antitreponemal antibodies is crucial to the distinction between neurosyphilis and nonneurosyphilis. A CSF titer 1:320 in the Treponema pallidum hemagglutination (TPHA) test is recommended as a quantitative cutoff.4 Another variant of this approach is to use the indices, such as the TPHA index and the intrathecally produced T. pallidum antibody (ITpA) index.4 The indices not only take into account the CSF-TPHA titer but also seem to minimize the influence of BCB dysfunction and blood antibody concentrations and are regarded as indicators of specific intrathecal synthesis. These TPHA titer/indices tests are more sensitive but not widely approved and are declared as requiring further validation.5 Currently, only nonreactive results of CSF treponemal tests are widely accepted criteria for excluding neurosyphilis.1,2 However, a recent study has shown a possibility of neurosyphilis with nonreactive CSF treponemal tests.6 To better understand the influence of CNS disturbance on CSF status in syphilis, we explored the relationships between the CSF test results and the established facts of intrathecal Ig synthesis and/or BCB dysfunction in patients with serological evidence of syphilis. We also verified the criteria currently mentioned in the guidelines.1,2,7
Volume 40, Number 12, December 2013
917
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
MATERIALS AND METHODS
showed evidence of both intrathecal Ig synthesis and BCB dysfunction (n = 16).
Data Collection We collected paired CSF and blood samples from 140 consecutive patients with serological evidence of syphilis that were undergoing diagnostic lumbar puncture in the venereology department at the Urals Institute of Dermatovenereology and Immunopathology in Yekaterinburg, Russia, from May 2006 through November 2011. The patients were deemed by the managing clinicians as possibly having neurosyphilis or were followed up after the treatment of neurosyphilis. The patients were referred from outpatient sexually transmitted diseases clinics, neurology units of inpatient hospitals, and psychiatry inpatient hospitals. We focused our attention on laboratory aspects and so did not limit our study to particular clinical forms or previous treatment of syphilis/neurosyphilis. The local ethics committee approved our study protocol. After giving their written informed consent, the patients underwent neurological examination, lumbar puncture, venipuncture, and subsequent standard diagnostic procedure. After being diagnosed, patients were treated and then were referred to outpatient sexually transmitted diseases clinics for monitoring of follow-up. Each pair of CSF and blood samples was collected within a single day. We used a portion of each sample for standard neurosyphilis diagnostic procedures. Then we aliquoted the remainders and stored them at j80-C for analysis. After collecting all the samples, we submitted them to an in-depth laboratory study. The clinical diagnosis of the samples was unknown to the operators of the laboratory assays. We obtained the patients’ characteristics from standardized medical records when laboratory analyses had been completed. We excluded 14 pairs of samples from the analysis for any one of the following technical reasons: blood contamination of CSF samples (6 subjects), blood and CSF samples collection on different days (5 subjects), and insufficient CSF sample quantity (3 subjects). The final number of patients in the study group was 126.
Diagnostic CSF Tests The diagnostic CSF examination included the CSFYfluorescent treponemal antibody absorption (FTA-Abs), CSF-TPHA, CSF-VDRL, and CSFYwhite blood cell (WBC) tests. We diluted CSF 1:5 with sorbent for the CSFYFTA-Abs test and considered the CSF sample as ‘‘nonreactive’’ when there was no or minimal reaction. We considered the CSF-TPHA reactive at dilutions of 1:4 or greater. We calculated the TPHA index by dividing the CSF-TPHA titer by the albumin quotient (CSF albumin 103/serum albumin). We calculated the ITpA index as the ratio of ratios: (CSF-TPHA/serum TPHA) to (total CSF-IgG/total serum IgG). We considered the CSF sample as ‘‘reactive’’ for the CSF-VDRL test if it was weakly reactive or reactive. We evaluated the 126 CSF samples as meeting or not meeting the ‘‘rule out’’ or ‘‘rule in’’ criteria for the diagnosis of neurosyphilis in CSF. By the rule-in criteria, we meant the criteria used in support of the diagnosis of neurosyphilis: a reactive CSF-VDRL test,1,2 a CSF-WBC count greater than 5 cells/KL plus a reactive CSF-TPHA and/or CSFYFTA-Abs tests,1,2 a reactive CSF-TPHA test at a titer greater than 1:320,2 a TPHA index greater than 70.0,2 and an ITpA index greater than 3.0.7 By the rule-out criteria, we meant the criteria used to refute the diagnosis of neurosyphilis: a nonreactive CSFYFTA-Abs test,1,2 a nonreactive CSF-TPHA test,2 and a reactive CSF-TPHA test at a titer lower than 1:320.2
Statistical Methods We used dichotomous variables (meeting/not meeting the criteria) for our analysis. We calculated the proportions of CSF samples meeting each of the criteria within each of the 4 groups and then analyzed the differences in proportions among the groups using W2 test, corrected for multiple comparisons using the BonferroniHolm method. We used MedCalc version 12.2 (MedCalc Software bvba, Belgium) to compute all our statistical tests.
RESULTS Detection of Intrathecal Ig Synthesis and BCB Dysfunction We determined serum and CSF albumin concentrations by nephelometric methods. We used enzyme-linked immunosorbent assays to quantitate the total IgA, IgM, and IgG. We analyzed the CSF and serum albumin and Ig concentrations within the same analytical series to improve the accuracy of calculating CSF/serum quotients. All tests were performed in duplicate. In those cases, where the values of secondary tests differed by more than 15%, a third study was conducted and the arithmetic mean value was used for the analysis. We estimated the BCB status with the CSF/serum albumin quotient (QAlb) using the age-related reference range [QAlb = (4 + age/15) 10j3] as the cutoff.8 Values above the cutoff were an indication of the BCB dysfunction. We performed a quantitative assessment of intrathecal Ig synthesis using Reiber’s hyperbolic formula and calculating the intrathecal Ig fraction.8 We diagnosed intrathecal Ig synthesis if the intrathecal fraction of any Ig class (IgG and/or IgM and/or IgA) was greater than 0 mg/L. We used intrathecal Ig synthesis and BCB dysfunction to classify the 126 CSF samples into 4 groups: group 1, which showed no evidence of either BCB dysfunction or intrathecal Ig synthesis (n = 51); group 2, which showed evidence of BCB dysfunction solely (n = 20); group 3, which showed evidence of intrathecal Ig synthesis solely (n = 39); and group 4, which
918
Patient Characteristics We examined CSF and blood samples from 126 patients with suspected neurosyphilis. All patients were reactive in 2 serum treponemal tests, and 111 had reactive serum nontreponemal tests. All but 2 patients were HIV negative. Twentytwo of our patients with untreated syphilis had the late-stage disease or the disease of unknown duration, and 1 had secondary syphilis. Eighty-eight patients had the history of previous treatment of nonneurological syphilis; 47 of them had more than 1 episode of treatment or retreatment. Patients received parenteral penicillin (long, intermediate, or short acting) and ceftriaxone. Information on serological response to prior treatment was available for 22 patients, of whom 1 had a 4-fold rise and 14 did not have a 4-fold drop in the titer over the recommended period of follow-up, and 7 had an adequate serological decline, but the titers did not revert to negative. Sixty-six patients did not have documented information about nontreponemal test titers evolution; 9 of them showed high titers (Q1:32), 31 showed titers from 1:4 to 1:16, 13 showed titers less or equal to 1:2, and 13 were nonreactive. Fifteen patients had a follow-up visit after neurosyphilis treatment, and 3 of them were examined before and after treatment of neurosyphilis. Eighty-five patients had any neurological and/or ophthalmologic and/or psychiatric symptoms or signs. Table 1 shows how patients with different demographic and clinical characteristics and syphilis histories were distributed among the 4 groups.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Criteria for the Diagnosis of Neurosyphilis in CSF
TABLE 1.
Distribution of 126 Patients With Different Characteristics Among 4 Groups
Characteristic
Group 1 (n = 51) Group 2 (n = 20) Group 3 (n = 39) Group 4 (n = 16)
Sex Male Female Age, M (R), y Treatment before study entry No history of syphilis treatment* Had history of nonneurological syphilis treatment Follow-up visit after neurosyphilis treatment (6 moY5 y) Clinical abnormalities Vision loss Hearing loss Headache seizures Cerebrovascular accident Neuropsychiatric disorders Cranial nerve disorders Motor function disorders Sensory disorders Coordination disorders Gait abnormalities Others diseases Lyme disease history or seroposivity Tuberculosis history HIV infection
16 (31) 35 (69) 37.1 (10.6)
15 (75) 5 (25) 42.0 (8.4)
27 (69) 12 (31) 42.7 (8.7)
15 (94) 1 (6) 45.4 (7.8)
5 (10) 45 (88) 1 (2)
2 (10) 18 (90) 0 (0)
7 (18) 19 (49) 13 (33)
9 (56) 6 (38) 1 (6)
8 (16) 3 (6) 9 (18) 1 (2) 3 (6) 7 (14) 16 (31) 8 (16) 0 (0) 4 (8) 1 (2)
1 (5) 0 (0) 4 (20) 0 (0) 0 (0) 1 (5) 5 (25) 3 (15) 0 (0) 3 (15) 0 (0)
6 (15) 1 (3) 7 (18) 0 (0) 1 (3) 12 (31) 16 (41) 9 (23) 3 (8) 1 (3) 1 (3)
4 (25) 2 (13) 1 (6) 0 (0) 2 (13) 9 (56) 11 (69) 4 (25) 2 (13) 0 (0) 1 (6)
5 (10) 1 (2) 0 (0)
1 (6) 0 (0) 0 (0)
1 (3) 1 (3) 1 (3)
0 (0) 1 (6) 1 (6)
Definition of groups: group 1 represents patients with no evidence of either BCB dysfunction or intrathecal Ig synthesis; group 2, with evidence of BCB dysfunction solely; group 3, with evidence of intrathecal Ig synthesis solely; group 4, with evidence of both intrathecal Ig synthesis and BCB dysfunction. The indicator of BCB dysfunction was elevated QAlb (age adjustment according to QAlb = (4 + age/15) 10j3, whereas QAlb is the CSF/serum albumin quotient); the indicator of intrathecal synthesis was a minimally detectable (90 mg/L) level of Reiber intrathecal fraction of any Ig class (IgG and/or IgM and/or IgA). *One patient with secondary syphilis and 22 patients with late syphilis (including syphilis of unknown duration).
Distribution of CSF Test Results by Groups To find which CSF test results were most typical for the groups, we counted the number of CSF samples meeting each of the criteria within each of the 4 groups and compared the proportions among the groups (Table 2). The reactive CSF-VDRL test was significantly more likely in groups 3 and 4, and there was a trend toward a higher proportion in group 4 compared with group 3, but the difference was not statistically significant. Cerebrospinal fluid pleocytosis was more common in group 4. The CSF-TPHA titer greater than 1:320 and the TPHA index greater than 70.0 were significantly more likely in groups 3 and 4. The ITpA index greater than 3.0 was often
TABLE 2.
found not only in groups 3 and 4 but also within group 2. Nonreactive CSFYFTA-Abs test was significantly more likely within groups 1 and 2. Nonreactive CSF-TPHA test samples were relatively common only within group 1. Reactive CSFTPHA samples at a titer below 1:320 were significantly more likely within group 2.
Combinations of CSF Test Results To compare the results that were obtained by applying different criteria to the same sample, we created a complete list of combinations of CSF test results that meeting (or not meeting) the rule-in and the rule-out criteria (Table 3).
Results of Laboratory Investigations
No. CSF Samples With the Indicated Diagnostic Test Result
CSF Test Results Reactive CSF-VDRL CSF-WBC 95/KL* CSF-TPHA titer Q1:640 TPHA index 970 ITpA index 93 Nonreactive CSFYFTA-Abs Nonreactive CSF-TPHA CSF-TPHA titer 1:4Y1:160
Group 1 Group 2 Group 3 Group 4 (n = 51) (n = 20) (n = 39) (n = 16) 0 (0) 0 (0) 3 (6) 4 (8) 14 (27) 40 (78) 22 (43) 24 (47)
1 (5) 2 (10) 1 (5) 1 (5) 14 (70) 12 (60) 2 (10) 16 (80)
16 (41) 10 (26) 29 (74) 30 (77) 26 (67) 3 (8) 0 (0) 7 (18)
10 (63) 10 (63) 13 (81) 13 (81) 11 (69) 0 (0) 0 (0) 2 (13)
Statistical Tests Bonferroni-HolmYCorrected Post Hoc Tests (P value) W2
P
42.0 35.9 67.1 66.6 20.1 60.2 33.2 27.5
G0.001 G0.001 G0.001 G0.001 0.002 G0.001 G0.001 G0.001
1 vs. 2 1 vs. 3 1 vs. 4 2 vs. 3 2 vs. 4 3 vs. 4 NS NS NS NS 0.009 NS 0.02 0.048
G0.001 0.001 G0.001 G0.001 0.001 G0.001 G0.001 0.026
G0.001 G0.001 G0.001 G0.001 0.03 G0.001 G0.001 0.036
0.015 NS G0.001 G0.001 NS G0.001 G0.001 G0.001
G0.001 0.006 G0.001 G0.001 NS G0.001 G0.001 G0.001
NS 0.043 NS NS NS NS NS NS
Definition of groups: see Table 1. *In combination with reactive CSF-TPHA and/or CSFYFTA-Abs.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
919
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
TABLE 3.
Combinations of CSF Test Results Meeting Different Criteria Used to Support (rule in) or Refute (rule out) the Diagnosis of Neurosyphilis
No. CSF Samples With the Indicated Combination Group 1 (n = 51)
Group 2 (n = 20)
Group 3 (n = 39)
Group 4 (n = 16)
Total (n = 126)
ITpA index 93 V
0 (0)
0 (0)
1 (3)
3 (19)
4 (3)
0 (0)
0 (0)
3 (8)
3 (19)
6 (5)
ITpA index 93 V
0 (0)
0 (0)
8 (21)
2 (13)
10 (8)
0 (0)
0 (0)
2 (5)
0 (0)
2 (2)
ITpA index 93 V
0 (0)
1 (5)
1 (3)
2 (13)
4 (3)
0 (0)
0 (0)
1 (3)
0 (0)
1 (1)
ITpA index 93 V
0 (0)
0 (0)
0 (0)
2 (13)
2 (2)
0 (0)
0 (0)
3 (8)
1 (6)
4 (3)
ITpA index 93 V
0 (0)
2 (10)
1 (3)
0 (0)
3 (2)
0 (0)
0 (0)
2 (5)
1 (6)
3 (2)
3 (6)
1 (5)
11 (28)
2 (13)
17 (14)
0 (0)
0 (0)
1 (3)
0 (0)
1 (1)
1 (2)
0 (0)
2 (5)
0 (0)
3 (2)
10 (20)
10 (50)
2 (5)
0 (0)
22 (17)
14 (27) 37 (73)
14 (70) 6 (30)
38 (97) 1 (3)
22 (43)
2 (10)
0 (0)
0 (0)
24 (19)
Combinations of CSF Test Results Rule-in criteria Reactive CSF-WBC CSF-VDRL 95/KL Reactive CSF-WBC CSF-VDRL 95/KL Reactive V CSF-VDRL Reactive V CSF-VDRL Reactive V CSF-VDRL Reactive V CSF-VDRL V CSF-WBC 95/KL V CSF-WBC 95/KL V CSF-WBC 95/KL V CSF-WBC 95/KL V V V
V
V
V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640 CSF-TPHA Q1:640 CSF-TPHA Q1:640 V
TPHA index 970 TPHA index 970 TPHA index 970 TPHA index 970 V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640 V
TPHA index 970 TPHA index 970 V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640
TPHA index 970 V
V
TPHA index 970 V
ITpA index 93 ITpA index 93 ITpA index 93 ITpA index 93
Total Results not meeting the rule-in criteria
16 (100) 0 (0)
82 (65) 44 (35)
Rule-out criteria Nonreactive CSFYFTAAbs Nonreactive CSFYFTAAbs V
Nonreactive CSF-TPHA
Nonreactive CSFYFTAAbs
V
V
CSF-TPHA 1:4Y1:160
16 (31)
10 (50)
2 (5)
0 (0)
28 (22)
V
CSF-TPHA 1:4Y1:160 V
8 (16)
6 (30)
5 (13)
2 (13)
21 (17)
1 (2)
0 (0)
1 (3)
0 (0)
2 (2)
47 (92) 4 (8)
18 (90) 2 (10)
8 (21) 31 (79)
2 (13) 14 (88)
75 (60) 51 (40)
Total Results not meeting the rule-out criteria Definition of groups: see Table 1. ‘‘V’’, results not meeting the criteria.
All but 1 (96%) reactive CSF-VDRL test samples and all but 3 (86%) CSF samples with pleocytosis were also meeting at least 1 of the TPHA titer/indices-based criteria. Group 3 had a higher proportion of samples with elevated CSF-TPHA titers and/or indices and nonreactive CSF-VDRL tests than did group 4. All samples with elevated CSF-TPHA titers had an elevated TPHA index and/or ITpA index. Elevated TPHA indices at low CSF-TPHA titers were found in 1 sample in group 1 and 2 samples in group 3. Elevated ITpA indices at low CSF-TPHA titers were found significantly more often: in 11 samples in group 1, in 13 samples in group 2, and in 6 samples in group 3. Samples, nonreactive in both the CSFYFTA-Abs and CSF-TPHA tests, referred primarily to group 1 (n = 22) and in
920
less extent to group 2 (n = 2). Twenty-eight samples, nonreactive in the CSFYFTA-Abs, had low titers in the CSF-TPHA test. There were no samples, reactive in the CSFYFTA-Abs test and nonreactive in the CSF-TPHA test. All samples in group 4 (100%) and all but 1 in group 3 (97%) met at least 1 of the rule-in criteria. All but 4 samples within group 1 (92%) and all but 2 (90%) within group 2 met at least 1 of the rule-out criteria.
CSF Data of 3 Patients Before and After Treatment of Neurosyphilis Of the 15 patients who were punctured for neurosyphilis therapy control, 3 were included in the study both after and
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Criteria for the Diagnosis of Neurosyphilis in CSF
TABLE 4.
CSF Data of 3 Patients With Neurosyphilis Before and After Treatment*
Patient A CSF Data Group CSF-WBC/KL CSF-VDRL CSF-TPHA titer TPHA index ITpA index CSFYFTA-Abs
Patient B
Patient C
Before
After + 29 mo
Before
After + 8 mo
Before
After + 27 mo
3 62 Reactive 1:640 119 1 Reactive 4+
3 2 Nonreactive 1:320 107 7 Reactive 4+
4 22 Reactive 1:10,240 875 0.3 Reactive 4+
4 3 Reactive 1:5120 391 12 Reactive 3+
4 34 Reactive 1:163,840 12,412 8 Reactive 4+
3 2 Nonreactive 1:40,960 6606 161 Reactive 2+
*Patients were treated with intravenous benzyl penicillin G (10 million units twice daily for 28 days).
before their treatment of neurosyphilis. The CSF data on these patients are shown in Table 4. All the 3 patients had more than a 4-fold decline in a serum RPR titer at a follow-up visit. There was positive dynamics of all CSF tests except ITpA indices.
DISCUSSION The diagnosis of the CNS involvement is difficult when a patient with suspected neurosyphilis has no typical clinical manifestations. There is no definitive laboratory test to diagnose neurosyphilis. The CSF-VDRL test is highly specific but insensitive; the CSF-WBC test is both nonspecific and insensitive. These tests are accepted to rule in the diagnosis. The CSF treponemal-specific tests are highly sensitive and used mostly to rule out the diagnosis. There is no consensus on the use of TPHA titer/indices-based criteria for diagnosing neurosyphilis. Current Centers for Disease Control and Prevention standards do not favor the use of TPHA titer/indices for the diagnosis of patients with suspected neurosyphilis. Although some published research9 has shown that treponemal antibodies may be synthesized intrathecally, there is insufficient evidence that these tests always reflect the immune response in the CNS and that peripheral antibodies do not imitate positive results. Our study has been designed to explore the accuracy of current criteria used to diagnose neurosyphilis for assessment of intrathecal immune response. Our findings have shown that all the examined criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity. We have found that elevated CSF-TPHA titers and indices, elevated CSF-WBC counts, and reactive CSF-VDRL tests are associated with the signs of intrathecal Ig synthesis, whereas nonreactive CSFYFTAAbs, nonreactive CSF-TPHA tests, and CSF-TPHA titers from 1:4 to 1:160 are associated with cases showing no intrathecal synthesis. These findings agree with the previous studies that have reported that intrathecal Ig synthesis is a very common laboratory symptom of neurosyphilis10 and that there is a correlation between intrathecal Ig production and the presence of CSF syphilis antibodies and pleocytosis in patients with syphilis.11 Some experts recommend to interpret the combination of intrathecal Ig synthesis, BCB dysfunction, and an increased CSF cell count as an active infection, whereas intrathecal Ig synthesis without BCB dysfunction may be a consequence of past infection.8 Therefore, we have distinguished 2 separate intrathecal Ig synthesis groups. We consider group 4 as a group of high probability of active infection and group 3 as a group with possible past infection. A confirmation of such approach is the fact that the proportion of patients with late syphilis without any history of syphilis treatment was greater within group 4, Sexually Transmitted Diseases
&
whereas the proportion of patients followed up after their treatment of neurosyphilis was greater within group 3. We have found a tendency to higher incidence of the reactive CSF-VDRL test in group 4 compared with group 3, whereas the proportions of elevated CSF-TPHA titers/indices in these groups were almost equal. In most cases, the reactive VDRL test and CSF pleocytosis manifested in the presence of elevated CSF-TPHA titers/indices. Group 3, compared with group 4, had a higher percentage of samples that had elevated CSF-TPHA titers and indices but were nonreactive in the CSF-VDRL test. These findings do not contradict to the previously published studies that have reported that nonreactive treponemal tests predict nonreactive VDRL test6 and that CSF-TPHA titers and indices are more sensitive than the VDRL test.4,5 We presume that the findings are biologically plausible. All forms of neurosyphilis are the result of T. pallidum invasion into the CNS owing to the blood dissemination. Antibody production is an integral part of the immune response against this extracellular pathogen. The immune reaction against T. pallidum proteins is strong, obligatory, and permanent, and theoretically, intrathecal treponemal antibody production should be in all immunocompetent patients with neurosyphilis because it is a response to the invasion. The invasion does not mean the disease is active, but it is a prerequisite. The immune response against lipoid (VDRL) antigens is more associated with the disease activity because such antibodies are typically associated with the tissue damage and the ‘‘scar syndrome’’ is less pronounced than in treponemal antibodies. We have found that the TPHA indices are in no better conformity with the facts of intrathecal immune response than the TPHA titers. The TPHA index demonstrates concordant results with the CSF-TPHA titer criterion in most cases. The ITpA index greater than 3.0 often occurs not only in the presence of intrathecal Ig synthesis but also in its absence and BCB dysfunction. Isolated BCB dysfunction is not a common symptom for infectious diseases of CNS and may occur even in nonneurological patients.12,13 A possible explanation for the finding could be higher sensitivity of ITpA index because the total intrathecal IgG synthesis is inferior to antigen-specific intrathecal IgG synthesis.14 Another explanation could be false identification of intrathecal antibody production by ITpA index in low CSF-TPHA titer cases because of the complexity of this parameter. Changes of ITpA index after neuroshyphilis treatment in our 3 patients have also raised doubts about its accuracy. Further studies are required. We have also found the difference between CSFYFTAAbs and CSF-TPHA tests. A significant number of samples are nonreactive by CSFYFTA-Abs test and reactive by CSF-TPHA at titers from 1:4 to 1:160. The nonreactive CSFYFTA-Abs test
Volume 40, Number 12, December 2013
921
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
is more likely among cases with undetected intrathecal synthesis regardless of BCB function, whereas low-titer CSF-TPHA test samples are often found among cases with undetected intrathecal synthesis and impaired BCB. This difference may be explained by the CSF mixing with the sorbent for the FTA-Abs test, but not for the TPHA test, and by the partial removal of antibodies (cross-reactive antibodies to nonpathogenic treponemes). These findings highlight the importance of antibody level in evaluating CSF and reveal some regularities in the CSF test result discrepancies. Our study is limited by at least a 1-year history of syphilis in all but 1 patient. There is a possibility that during the early stages of syphilis, the associations may manifest differently than in the late stages of the disease, and this requires additional studies. Another limitation is the small number of patients in groups 4 and 2. This problem is caused by the rare occurrence of BCB dysfunction in patients with syphilis. Therefore, our analyses may not have statistical power to detect some findings. We do not regard the fact that 98% of our patients are HIV negative as a limitation because this raises our confidence that the detected CSF abnormalities have been caused by syphilitic infection. In terms of practice, our study argues that a large amount of treponemal antibodies in CSF is more likely a consequence of CNS pathology, rather than just a mass of blood-derived antibodies. We suggest using the criterion of a CSF-TPHA titer greater than 320 in untreated patients with syphilis, in whom neurosyphilis is suspected, but CSF-VDRL is nonreactive. The applicability of this criterion for making a diagnosis of neurosyphilis in patients with the history of syphilis treatment is doubtful because it is unknown whether this finding indicates inadequate treatment or it is a scar syndrome in case of cured infection. A search for CSF biomarkers associated with an active phase of intrathecal immune response is necessary. The recently proposed CXCL13 level15 seems to meet these requirements. However, a CSF-TPHA titer is much more accurate for ruling in, rather than ruling out neurosyphilis. In our opinion, a low CSF treponemal antibody level or even a nonreactive result of treponemal tests does not necessarily exclude neurosyphilis because the features of CNS inflammatory focus, especially its location in relation to the CSF spaces, can significantly affect the CSF antibody concentration. A recent publication has proved that there is a possibility of spirochetal infection of the CNS without laboratory signs of intrathecal antibody synthesis.16 Overall, our study has shown that all currently used criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity and support the
922
hypothesis that high CSF treponemal-specific antibody titers are a consequence of inflammatory pathology of the CNS. REFERENCES 1. Workowski KA, Berman S. Centers for Disease Control and Prevention (CDC). Sexual transmitted disease treatment guidelines, 2010. MMWR Recomm Rep 2010; 59:1Y11. 2. French P, Gomberg M, Janier M, et al. IUSTI: 2008 European guidelines on the management of syphilis. Int J STD AIDS 2009; 20:300Y309. 3. Jaffe HW, Larsen SA, Peters M, et al. Tests for treponemal antibody in CSF. Arch Intern Med 1978; 138:252Y255. 4. Luger AF, Schmidt BL, Kaulich M. Significance of laboratory findings for the diagnosis of neurosyphilis. Int J STD AIDS 2000; 11:224Y34. 5. Schmidt BL, Van Voorst Vader PC. Laboratory diagnosis of neurosyphilis in Europe. Presented at: IUSTI/WHO Conference on STI. Europe Syphilis Guideline Expert Workshop; 2004; Mykonos. 6. Harding A, Ghanem K. The performance of cerebrospinal fluid treponemal-specific antibody tests in neurosyphilis: A systematic review. Sex Transm Dis 2012; 39:291Y97. 7. Prange HW. Neurosyphilis. In: Gross G, Tyring S, eds. Sexually Transmitted Infections and Sexually Transmitted Diseases. Berlin: Springer; 2011:163Y171. 8. Reiber H, Peter JB. Cerebrospinal fluid analysisVDisease related data patterns and evaluation programs. J Neurol Sci 2001; 184:101Y122. 9. Vartal F, Vandvik B, Michaelsen TE, et al. Neurosyphilis: Intrathecal synthesis of oligoclonal antibodies to Treponema pallidum. Ann Neurol 1982; 11:35Y40. 10. Hische E, Tutuarima J, Wolters E, et al. Cerebrospinal fluid IgG and IgM indexes as indicators of active neurosyphilis. Clin Chem 1988; 34:665Y667. 11. Lo¨whagen GB, Andersson M, Blomstrand C, et al. Central nervous system involvement in early syphilis. Part 1. Intrathecal immunoglobulin production. Acta Derm Venereol 1983; 63:409Y417. 12. Brettschneider J, Claus A, Kassubek J, et al. Isolated bloodYcerebrospinal fluid barrier dysfunction: Prevalence and associated diseases. J Neurol 2005; 252:1067Y1073. 13. Deisenhammera F, Bartosb A, Egga R, et al. Guidelines on routine cerebrospinal fluid analysis. Report from an EFNS task. Eur J Neurol 2006; 13:913Y922. 14. Jarius S, Eichhorn P, Wildemann B, et al. Usefulness of antibody index assessment in cerebrospinal fluid from patients negative for total-IgG oligoclonal bands. Fluids Barriers CNS 2012; 9:14. 15. Marra CM, Tantalo LC, Sahi SK, et al. CXCL13 as a cerebrospinal fluid marker for neurosyphilis in HIV-infected patients with syphilis. Sex Transm Dis 2010; 37:283Y287. 16. Vrethem M, Widhe M, Ernerudh J, et al. Clinical, diagnostic and immunological characteristics of patients with possible neuroborreliosis without intrathecal Ig-synthesis against Borrelia antigen in the cerebrospinal fluid. Neurol Int 2011; 3:e2.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Criteria for the Diagnosis of Neurosyphilis in Cerebrospinal Fluid: Relationships With Intrathecal Immunoglobulin Synthesis and BloodYCerebrospinal Fluid Barrier Dysfunction Nadezhda Levchik, MD, PhD,* Marina Ponomareva, MS,* Vera Surganova, MD, PhD,Þ Natalia Zilberberg, MD, PhD,þ and Nikolai Kungurov, MD, PhDþ
Background: The origin of cerebrospinal fluid (CSF) syphilis antibodies (intrathecal or blood-derived) is in doubt. Little is known about CSF test behavior under the condition of physiological or disturbed functioning of blood-CSF barrier (BCB) and intrathecal immunoglobulin (Ig) production. Methods: We collected 126 serum/CSF pairs from patients with serological evidence of syphilis. We explored the relationships between the established facts of intrathecal Ig synthesis and/or BCB dysfunction and the results of CSF diagnostic tests: the Treponema pallidum hemagglutination (TPHA) test, the fluorescent treponemal antibody absorption (FTA-Abs) test, the Venereal Disease Research Laboratory (VDRL) test, and white blood cell counts. We checked the criteria used either to support or refute the diagnosis of neurosyphilis. Results: Reactive CSF-VDRL tests, elevated CSFYwhite blood cell counts, and elevated CSF-TPHA titers/indices were associated with the signs of intrathecal Ig synthesis, whereas nonreactive CSFYfluorescent treponemal antibody absorption, nonreactive CSF-TPHA tests, and CSF-TPHA titers from 1:4 to 1:160 were associated with cases where the intrathecal synthesis was not detected. There were some peculiarities of the tests toward BCB dysfunction. Most of reactive CSF-VDRL test samples and CSF samples with pleocytosis were also meeting at least 1 of the CSF-TPHA titer/indicesbased criteria. T. pallidum hemagglutination indices were in no better conformity with the facts of intrathecal immune response than CSFTPHA titers. Conclusions: Our findings have shown that all the examined criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity and support the hypothesis that high CSF treponemal-specific antibody titers are a consequence of inflammatory pathology of the central nervous system.
T
he laboratory verification of the central nervous system (CNS) involvement in syphilis constitutes a problem. The purpose of cerebrospinal fluid (CSF) analysis in the neurosyphilis
From the Departments of *Laboratory Medicine, †Venereology, and ‡Science, Urals Institute of Dermatovenereology and Immunopathology, Yekaterinburg, Russia
Conflict of interest: None declared. Funding: None declared. Correspondence: Nadezhda Levchik, MD, PhD, Department of Laboratory Medicine, Urals Institute of Dermatovenereology and Immunopathology, Shcherbakova 8, 620076 Yekaterinburg, Russia. E-mail: [email protected]. Received for publication April 26, 2013, and accepted September 20, 2013. DOI: 10.1097/OLQ.0000000000000049 Copyright * 2013 American Sexually Transmitted Diseases Association All rights reserved.
Sexually Transmitted Diseases
&
diagnosis is to detect pathogen-driven intrathecal immune response. Cerebrospinal fluid pleocytosis is a standard marker of active inflammation within CNS, but it is nonspecific and low sensitive. The other diagnostic tools are CSF nontreponemal and treponemal tests that identify antiphospholipid and antitreponemal antibodies, respectively. Cerebrospinal fluid syphilitic antibodies may have a dual origin. On the one hand, small amounts of immunoglobulins (Igs) are constantly supplied by the blood; on the other hand, Igs may be synthesized by plasma cells of perivascular inflammatory infiltrates in response to CNS infection. Only the latter ones have a diagnostic value. However, it is theoretically possible that an increase in blood-derived Ig concentration in CSF will lead to misdiagnosing intrathecal synthesis. The increase may be caused by high blood antibody level and/or the so-called blood-CSF barrier (BCB) dysfunction (a disorder of blood-CNS barrier integrity and/or impairment of CSF circulation). It is assumed that CSF antiphospholipid antibodies have intrathecal origin because their blood level is low. Reactive CSF nontreponemal tests are considered highly specific, and one of them, the Venereal Disease Research Laboratory (VDRL) test, is currently the ‘‘gold standard’’ for neurosyphilis.1,2 However, the gold standard is not perfect because a significant part of patients, who were clinically defined as having neurosyphilis, had a nonreactive CSF-VDRL test. The intrathecal origin of CSF antitreponemal antibodies is questionable because their blood level is considerable, and it has been shown that there is a relationship between the CSF test reactivity and a blood antibody titer.3 Some have suggested that the level of CSF antitreponemal antibodies is crucial to the distinction between neurosyphilis and nonneurosyphilis. A CSF titer 1:320 in the Treponema pallidum hemagglutination (TPHA) test is recommended as a quantitative cutoff.4 Another variant of this approach is to use the indices, such as the TPHA index and the intrathecally produced T. pallidum antibody (ITpA) index.4 The indices not only take into account the CSF-TPHA titer but also seem to minimize the influence of BCB dysfunction and blood antibody concentrations and are regarded as indicators of specific intrathecal synthesis. These TPHA titer/indices tests are more sensitive but not widely approved and are declared as requiring further validation.5 Currently, only nonreactive results of CSF treponemal tests are widely accepted criteria for excluding neurosyphilis.1,2 However, a recent study has shown a possibility of neurosyphilis with nonreactive CSF treponemal tests.6 To better understand the influence of CNS disturbance on CSF status in syphilis, we explored the relationships between the CSF test results and the established facts of intrathecal Ig synthesis and/or BCB dysfunction in patients with serological evidence of syphilis. We also verified the criteria currently mentioned in the guidelines.1,2,7
Volume 40, Number 12, December 2013
917
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
MATERIALS AND METHODS
showed evidence of both intrathecal Ig synthesis and BCB dysfunction (n = 16).
Data Collection We collected paired CSF and blood samples from 140 consecutive patients with serological evidence of syphilis that were undergoing diagnostic lumbar puncture in the venereology department at the Urals Institute of Dermatovenereology and Immunopathology in Yekaterinburg, Russia, from May 2006 through November 2011. The patients were deemed by the managing clinicians as possibly having neurosyphilis or were followed up after the treatment of neurosyphilis. The patients were referred from outpatient sexually transmitted diseases clinics, neurology units of inpatient hospitals, and psychiatry inpatient hospitals. We focused our attention on laboratory aspects and so did not limit our study to particular clinical forms or previous treatment of syphilis/neurosyphilis. The local ethics committee approved our study protocol. After giving their written informed consent, the patients underwent neurological examination, lumbar puncture, venipuncture, and subsequent standard diagnostic procedure. After being diagnosed, patients were treated and then were referred to outpatient sexually transmitted diseases clinics for monitoring of follow-up. Each pair of CSF and blood samples was collected within a single day. We used a portion of each sample for standard neurosyphilis diagnostic procedures. Then we aliquoted the remainders and stored them at j80-C for analysis. After collecting all the samples, we submitted them to an in-depth laboratory study. The clinical diagnosis of the samples was unknown to the operators of the laboratory assays. We obtained the patients’ characteristics from standardized medical records when laboratory analyses had been completed. We excluded 14 pairs of samples from the analysis for any one of the following technical reasons: blood contamination of CSF samples (6 subjects), blood and CSF samples collection on different days (5 subjects), and insufficient CSF sample quantity (3 subjects). The final number of patients in the study group was 126.
Diagnostic CSF Tests The diagnostic CSF examination included the CSFYfluorescent treponemal antibody absorption (FTA-Abs), CSF-TPHA, CSF-VDRL, and CSFYwhite blood cell (WBC) tests. We diluted CSF 1:5 with sorbent for the CSFYFTA-Abs test and considered the CSF sample as ‘‘nonreactive’’ when there was no or minimal reaction. We considered the CSF-TPHA reactive at dilutions of 1:4 or greater. We calculated the TPHA index by dividing the CSF-TPHA titer by the albumin quotient (CSF albumin 103/serum albumin). We calculated the ITpA index as the ratio of ratios: (CSF-TPHA/serum TPHA) to (total CSF-IgG/total serum IgG). We considered the CSF sample as ‘‘reactive’’ for the CSF-VDRL test if it was weakly reactive or reactive. We evaluated the 126 CSF samples as meeting or not meeting the ‘‘rule out’’ or ‘‘rule in’’ criteria for the diagnosis of neurosyphilis in CSF. By the rule-in criteria, we meant the criteria used in support of the diagnosis of neurosyphilis: a reactive CSF-VDRL test,1,2 a CSF-WBC count greater than 5 cells/KL plus a reactive CSF-TPHA and/or CSFYFTA-Abs tests,1,2 a reactive CSF-TPHA test at a titer greater than 1:320,2 a TPHA index greater than 70.0,2 and an ITpA index greater than 3.0.7 By the rule-out criteria, we meant the criteria used to refute the diagnosis of neurosyphilis: a nonreactive CSFYFTA-Abs test,1,2 a nonreactive CSF-TPHA test,2 and a reactive CSF-TPHA test at a titer lower than 1:320.2
Statistical Methods We used dichotomous variables (meeting/not meeting the criteria) for our analysis. We calculated the proportions of CSF samples meeting each of the criteria within each of the 4 groups and then analyzed the differences in proportions among the groups using W2 test, corrected for multiple comparisons using the BonferroniHolm method. We used MedCalc version 12.2 (MedCalc Software bvba, Belgium) to compute all our statistical tests.
RESULTS Detection of Intrathecal Ig Synthesis and BCB Dysfunction We determined serum and CSF albumin concentrations by nephelometric methods. We used enzyme-linked immunosorbent assays to quantitate the total IgA, IgM, and IgG. We analyzed the CSF and serum albumin and Ig concentrations within the same analytical series to improve the accuracy of calculating CSF/serum quotients. All tests were performed in duplicate. In those cases, where the values of secondary tests differed by more than 15%, a third study was conducted and the arithmetic mean value was used for the analysis. We estimated the BCB status with the CSF/serum albumin quotient (QAlb) using the age-related reference range [QAlb = (4 + age/15) 10j3] as the cutoff.8 Values above the cutoff were an indication of the BCB dysfunction. We performed a quantitative assessment of intrathecal Ig synthesis using Reiber’s hyperbolic formula and calculating the intrathecal Ig fraction.8 We diagnosed intrathecal Ig synthesis if the intrathecal fraction of any Ig class (IgG and/or IgM and/or IgA) was greater than 0 mg/L. We used intrathecal Ig synthesis and BCB dysfunction to classify the 126 CSF samples into 4 groups: group 1, which showed no evidence of either BCB dysfunction or intrathecal Ig synthesis (n = 51); group 2, which showed evidence of BCB dysfunction solely (n = 20); group 3, which showed evidence of intrathecal Ig synthesis solely (n = 39); and group 4, which
918
Patient Characteristics We examined CSF and blood samples from 126 patients with suspected neurosyphilis. All patients were reactive in 2 serum treponemal tests, and 111 had reactive serum nontreponemal tests. All but 2 patients were HIV negative. Twentytwo of our patients with untreated syphilis had the late-stage disease or the disease of unknown duration, and 1 had secondary syphilis. Eighty-eight patients had the history of previous treatment of nonneurological syphilis; 47 of them had more than 1 episode of treatment or retreatment. Patients received parenteral penicillin (long, intermediate, or short acting) and ceftriaxone. Information on serological response to prior treatment was available for 22 patients, of whom 1 had a 4-fold rise and 14 did not have a 4-fold drop in the titer over the recommended period of follow-up, and 7 had an adequate serological decline, but the titers did not revert to negative. Sixty-six patients did not have documented information about nontreponemal test titers evolution; 9 of them showed high titers (Q1:32), 31 showed titers from 1:4 to 1:16, 13 showed titers less or equal to 1:2, and 13 were nonreactive. Fifteen patients had a follow-up visit after neurosyphilis treatment, and 3 of them were examined before and after treatment of neurosyphilis. Eighty-five patients had any neurological and/or ophthalmologic and/or psychiatric symptoms or signs. Table 1 shows how patients with different demographic and clinical characteristics and syphilis histories were distributed among the 4 groups.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Criteria for the Diagnosis of Neurosyphilis in CSF
TABLE 1.
Distribution of 126 Patients With Different Characteristics Among 4 Groups
Characteristic
Group 1 (n = 51) Group 2 (n = 20) Group 3 (n = 39) Group 4 (n = 16)
Sex Male Female Age, M (R), y Treatment before study entry No history of syphilis treatment* Had history of nonneurological syphilis treatment Follow-up visit after neurosyphilis treatment (6 moY5 y) Clinical abnormalities Vision loss Hearing loss Headache seizures Cerebrovascular accident Neuropsychiatric disorders Cranial nerve disorders Motor function disorders Sensory disorders Coordination disorders Gait abnormalities Others diseases Lyme disease history or seroposivity Tuberculosis history HIV infection
16 (31) 35 (69) 37.1 (10.6)
15 (75) 5 (25) 42.0 (8.4)
27 (69) 12 (31) 42.7 (8.7)
15 (94) 1 (6) 45.4 (7.8)
5 (10) 45 (88) 1 (2)
2 (10) 18 (90) 0 (0)
7 (18) 19 (49) 13 (33)
9 (56) 6 (38) 1 (6)
8 (16) 3 (6) 9 (18) 1 (2) 3 (6) 7 (14) 16 (31) 8 (16) 0 (0) 4 (8) 1 (2)
1 (5) 0 (0) 4 (20) 0 (0) 0 (0) 1 (5) 5 (25) 3 (15) 0 (0) 3 (15) 0 (0)
6 (15) 1 (3) 7 (18) 0 (0) 1 (3) 12 (31) 16 (41) 9 (23) 3 (8) 1 (3) 1 (3)
4 (25) 2 (13) 1 (6) 0 (0) 2 (13) 9 (56) 11 (69) 4 (25) 2 (13) 0 (0) 1 (6)
5 (10) 1 (2) 0 (0)
1 (6) 0 (0) 0 (0)
1 (3) 1 (3) 1 (3)
0 (0) 1 (6) 1 (6)
Definition of groups: group 1 represents patients with no evidence of either BCB dysfunction or intrathecal Ig synthesis; group 2, with evidence of BCB dysfunction solely; group 3, with evidence of intrathecal Ig synthesis solely; group 4, with evidence of both intrathecal Ig synthesis and BCB dysfunction. The indicator of BCB dysfunction was elevated QAlb (age adjustment according to QAlb = (4 + age/15) 10j3, whereas QAlb is the CSF/serum albumin quotient); the indicator of intrathecal synthesis was a minimally detectable (90 mg/L) level of Reiber intrathecal fraction of any Ig class (IgG and/or IgM and/or IgA). *One patient with secondary syphilis and 22 patients with late syphilis (including syphilis of unknown duration).
Distribution of CSF Test Results by Groups To find which CSF test results were most typical for the groups, we counted the number of CSF samples meeting each of the criteria within each of the 4 groups and compared the proportions among the groups (Table 2). The reactive CSF-VDRL test was significantly more likely in groups 3 and 4, and there was a trend toward a higher proportion in group 4 compared with group 3, but the difference was not statistically significant. Cerebrospinal fluid pleocytosis was more common in group 4. The CSF-TPHA titer greater than 1:320 and the TPHA index greater than 70.0 were significantly more likely in groups 3 and 4. The ITpA index greater than 3.0 was often
TABLE 2.
found not only in groups 3 and 4 but also within group 2. Nonreactive CSFYFTA-Abs test was significantly more likely within groups 1 and 2. Nonreactive CSF-TPHA test samples were relatively common only within group 1. Reactive CSFTPHA samples at a titer below 1:320 were significantly more likely within group 2.
Combinations of CSF Test Results To compare the results that were obtained by applying different criteria to the same sample, we created a complete list of combinations of CSF test results that meeting (or not meeting) the rule-in and the rule-out criteria (Table 3).
Results of Laboratory Investigations
No. CSF Samples With the Indicated Diagnostic Test Result
CSF Test Results Reactive CSF-VDRL CSF-WBC 95/KL* CSF-TPHA titer Q1:640 TPHA index 970 ITpA index 93 Nonreactive CSFYFTA-Abs Nonreactive CSF-TPHA CSF-TPHA titer 1:4Y1:160
Group 1 Group 2 Group 3 Group 4 (n = 51) (n = 20) (n = 39) (n = 16) 0 (0) 0 (0) 3 (6) 4 (8) 14 (27) 40 (78) 22 (43) 24 (47)
1 (5) 2 (10) 1 (5) 1 (5) 14 (70) 12 (60) 2 (10) 16 (80)
16 (41) 10 (26) 29 (74) 30 (77) 26 (67) 3 (8) 0 (0) 7 (18)
10 (63) 10 (63) 13 (81) 13 (81) 11 (69) 0 (0) 0 (0) 2 (13)
Statistical Tests Bonferroni-HolmYCorrected Post Hoc Tests (P value) W2
P
42.0 35.9 67.1 66.6 20.1 60.2 33.2 27.5
G0.001 G0.001 G0.001 G0.001 0.002 G0.001 G0.001 G0.001
1 vs. 2 1 vs. 3 1 vs. 4 2 vs. 3 2 vs. 4 3 vs. 4 NS NS NS NS 0.009 NS 0.02 0.048
G0.001 0.001 G0.001 G0.001 0.001 G0.001 G0.001 0.026
G0.001 G0.001 G0.001 G0.001 0.03 G0.001 G0.001 0.036
0.015 NS G0.001 G0.001 NS G0.001 G0.001 G0.001
G0.001 0.006 G0.001 G0.001 NS G0.001 G0.001 G0.001
NS 0.043 NS NS NS NS NS NS
Definition of groups: see Table 1. *In combination with reactive CSF-TPHA and/or CSFYFTA-Abs.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
919
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
TABLE 3.
Combinations of CSF Test Results Meeting Different Criteria Used to Support (rule in) or Refute (rule out) the Diagnosis of Neurosyphilis
No. CSF Samples With the Indicated Combination Group 1 (n = 51)
Group 2 (n = 20)
Group 3 (n = 39)
Group 4 (n = 16)
Total (n = 126)
ITpA index 93 V
0 (0)
0 (0)
1 (3)
3 (19)
4 (3)
0 (0)
0 (0)
3 (8)
3 (19)
6 (5)
ITpA index 93 V
0 (0)
0 (0)
8 (21)
2 (13)
10 (8)
0 (0)
0 (0)
2 (5)
0 (0)
2 (2)
ITpA index 93 V
0 (0)
1 (5)
1 (3)
2 (13)
4 (3)
0 (0)
0 (0)
1 (3)
0 (0)
1 (1)
ITpA index 93 V
0 (0)
0 (0)
0 (0)
2 (13)
2 (2)
0 (0)
0 (0)
3 (8)
1 (6)
4 (3)
ITpA index 93 V
0 (0)
2 (10)
1 (3)
0 (0)
3 (2)
0 (0)
0 (0)
2 (5)
1 (6)
3 (2)
3 (6)
1 (5)
11 (28)
2 (13)
17 (14)
0 (0)
0 (0)
1 (3)
0 (0)
1 (1)
1 (2)
0 (0)
2 (5)
0 (0)
3 (2)
10 (20)
10 (50)
2 (5)
0 (0)
22 (17)
14 (27) 37 (73)
14 (70) 6 (30)
38 (97) 1 (3)
22 (43)
2 (10)
0 (0)
0 (0)
24 (19)
Combinations of CSF Test Results Rule-in criteria Reactive CSF-WBC CSF-VDRL 95/KL Reactive CSF-WBC CSF-VDRL 95/KL Reactive V CSF-VDRL Reactive V CSF-VDRL Reactive V CSF-VDRL Reactive V CSF-VDRL V CSF-WBC 95/KL V CSF-WBC 95/KL V CSF-WBC 95/KL V CSF-WBC 95/KL V V V
V
V
V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640 CSF-TPHA Q1:640 CSF-TPHA Q1:640 V
TPHA index 970 TPHA index 970 TPHA index 970 TPHA index 970 V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640 V
TPHA index 970 TPHA index 970 V
V
V
CSF-TPHA Q1:640 CSF-TPHA Q1:640
TPHA index 970 V
V
TPHA index 970 V
ITpA index 93 ITpA index 93 ITpA index 93 ITpA index 93
Total Results not meeting the rule-in criteria
16 (100) 0 (0)
82 (65) 44 (35)
Rule-out criteria Nonreactive CSFYFTAAbs Nonreactive CSFYFTAAbs V
Nonreactive CSF-TPHA
Nonreactive CSFYFTAAbs
V
V
CSF-TPHA 1:4Y1:160
16 (31)
10 (50)
2 (5)
0 (0)
28 (22)
V
CSF-TPHA 1:4Y1:160 V
8 (16)
6 (30)
5 (13)
2 (13)
21 (17)
1 (2)
0 (0)
1 (3)
0 (0)
2 (2)
47 (92) 4 (8)
18 (90) 2 (10)
8 (21) 31 (79)
2 (13) 14 (88)
75 (60) 51 (40)
Total Results not meeting the rule-out criteria Definition of groups: see Table 1. ‘‘V’’, results not meeting the criteria.
All but 1 (96%) reactive CSF-VDRL test samples and all but 3 (86%) CSF samples with pleocytosis were also meeting at least 1 of the TPHA titer/indices-based criteria. Group 3 had a higher proportion of samples with elevated CSF-TPHA titers and/or indices and nonreactive CSF-VDRL tests than did group 4. All samples with elevated CSF-TPHA titers had an elevated TPHA index and/or ITpA index. Elevated TPHA indices at low CSF-TPHA titers were found in 1 sample in group 1 and 2 samples in group 3. Elevated ITpA indices at low CSF-TPHA titers were found significantly more often: in 11 samples in group 1, in 13 samples in group 2, and in 6 samples in group 3. Samples, nonreactive in both the CSFYFTA-Abs and CSF-TPHA tests, referred primarily to group 1 (n = 22) and in
920
less extent to group 2 (n = 2). Twenty-eight samples, nonreactive in the CSFYFTA-Abs, had low titers in the CSF-TPHA test. There were no samples, reactive in the CSFYFTA-Abs test and nonreactive in the CSF-TPHA test. All samples in group 4 (100%) and all but 1 in group 3 (97%) met at least 1 of the rule-in criteria. All but 4 samples within group 1 (92%) and all but 2 (90%) within group 2 met at least 1 of the rule-out criteria.
CSF Data of 3 Patients Before and After Treatment of Neurosyphilis Of the 15 patients who were punctured for neurosyphilis therapy control, 3 were included in the study both after and
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Criteria for the Diagnosis of Neurosyphilis in CSF
TABLE 4.
CSF Data of 3 Patients With Neurosyphilis Before and After Treatment*
Patient A CSF Data Group CSF-WBC/KL CSF-VDRL CSF-TPHA titer TPHA index ITpA index CSFYFTA-Abs
Patient B
Patient C
Before
After + 29 mo
Before
After + 8 mo
Before
After + 27 mo
3 62 Reactive 1:640 119 1 Reactive 4+
3 2 Nonreactive 1:320 107 7 Reactive 4+
4 22 Reactive 1:10,240 875 0.3 Reactive 4+
4 3 Reactive 1:5120 391 12 Reactive 3+
4 34 Reactive 1:163,840 12,412 8 Reactive 4+
3 2 Nonreactive 1:40,960 6606 161 Reactive 2+
*Patients were treated with intravenous benzyl penicillin G (10 million units twice daily for 28 days).
before their treatment of neurosyphilis. The CSF data on these patients are shown in Table 4. All the 3 patients had more than a 4-fold decline in a serum RPR titer at a follow-up visit. There was positive dynamics of all CSF tests except ITpA indices.
DISCUSSION The diagnosis of the CNS involvement is difficult when a patient with suspected neurosyphilis has no typical clinical manifestations. There is no definitive laboratory test to diagnose neurosyphilis. The CSF-VDRL test is highly specific but insensitive; the CSF-WBC test is both nonspecific and insensitive. These tests are accepted to rule in the diagnosis. The CSF treponemal-specific tests are highly sensitive and used mostly to rule out the diagnosis. There is no consensus on the use of TPHA titer/indices-based criteria for diagnosing neurosyphilis. Current Centers for Disease Control and Prevention standards do not favor the use of TPHA titer/indices for the diagnosis of patients with suspected neurosyphilis. Although some published research9 has shown that treponemal antibodies may be synthesized intrathecally, there is insufficient evidence that these tests always reflect the immune response in the CNS and that peripheral antibodies do not imitate positive results. Our study has been designed to explore the accuracy of current criteria used to diagnose neurosyphilis for assessment of intrathecal immune response. Our findings have shown that all the examined criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity. We have found that elevated CSF-TPHA titers and indices, elevated CSF-WBC counts, and reactive CSF-VDRL tests are associated with the signs of intrathecal Ig synthesis, whereas nonreactive CSFYFTAAbs, nonreactive CSF-TPHA tests, and CSF-TPHA titers from 1:4 to 1:160 are associated with cases showing no intrathecal synthesis. These findings agree with the previous studies that have reported that intrathecal Ig synthesis is a very common laboratory symptom of neurosyphilis10 and that there is a correlation between intrathecal Ig production and the presence of CSF syphilis antibodies and pleocytosis in patients with syphilis.11 Some experts recommend to interpret the combination of intrathecal Ig synthesis, BCB dysfunction, and an increased CSF cell count as an active infection, whereas intrathecal Ig synthesis without BCB dysfunction may be a consequence of past infection.8 Therefore, we have distinguished 2 separate intrathecal Ig synthesis groups. We consider group 4 as a group of high probability of active infection and group 3 as a group with possible past infection. A confirmation of such approach is the fact that the proportion of patients with late syphilis without any history of syphilis treatment was greater within group 4, Sexually Transmitted Diseases
&
whereas the proportion of patients followed up after their treatment of neurosyphilis was greater within group 3. We have found a tendency to higher incidence of the reactive CSF-VDRL test in group 4 compared with group 3, whereas the proportions of elevated CSF-TPHA titers/indices in these groups were almost equal. In most cases, the reactive VDRL test and CSF pleocytosis manifested in the presence of elevated CSF-TPHA titers/indices. Group 3, compared with group 4, had a higher percentage of samples that had elevated CSF-TPHA titers and indices but were nonreactive in the CSF-VDRL test. These findings do not contradict to the previously published studies that have reported that nonreactive treponemal tests predict nonreactive VDRL test6 and that CSF-TPHA titers and indices are more sensitive than the VDRL test.4,5 We presume that the findings are biologically plausible. All forms of neurosyphilis are the result of T. pallidum invasion into the CNS owing to the blood dissemination. Antibody production is an integral part of the immune response against this extracellular pathogen. The immune reaction against T. pallidum proteins is strong, obligatory, and permanent, and theoretically, intrathecal treponemal antibody production should be in all immunocompetent patients with neurosyphilis because it is a response to the invasion. The invasion does not mean the disease is active, but it is a prerequisite. The immune response against lipoid (VDRL) antigens is more associated with the disease activity because such antibodies are typically associated with the tissue damage and the ‘‘scar syndrome’’ is less pronounced than in treponemal antibodies. We have found that the TPHA indices are in no better conformity with the facts of intrathecal immune response than the TPHA titers. The TPHA index demonstrates concordant results with the CSF-TPHA titer criterion in most cases. The ITpA index greater than 3.0 often occurs not only in the presence of intrathecal Ig synthesis but also in its absence and BCB dysfunction. Isolated BCB dysfunction is not a common symptom for infectious diseases of CNS and may occur even in nonneurological patients.12,13 A possible explanation for the finding could be higher sensitivity of ITpA index because the total intrathecal IgG synthesis is inferior to antigen-specific intrathecal IgG synthesis.14 Another explanation could be false identification of intrathecal antibody production by ITpA index in low CSF-TPHA titer cases because of the complexity of this parameter. Changes of ITpA index after neuroshyphilis treatment in our 3 patients have also raised doubts about its accuracy. Further studies are required. We have also found the difference between CSFYFTAAbs and CSF-TPHA tests. A significant number of samples are nonreactive by CSFYFTA-Abs test and reactive by CSF-TPHA at titers from 1:4 to 1:160. The nonreactive CSFYFTA-Abs test
Volume 40, Number 12, December 2013
921
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
Levchik et al.
is more likely among cases with undetected intrathecal synthesis regardless of BCB function, whereas low-titer CSF-TPHA test samples are often found among cases with undetected intrathecal synthesis and impaired BCB. This difference may be explained by the CSF mixing with the sorbent for the FTA-Abs test, but not for the TPHA test, and by the partial removal of antibodies (cross-reactive antibodies to nonpathogenic treponemes). These findings highlight the importance of antibody level in evaluating CSF and reveal some regularities in the CSF test result discrepancies. Our study is limited by at least a 1-year history of syphilis in all but 1 patient. There is a possibility that during the early stages of syphilis, the associations may manifest differently than in the late stages of the disease, and this requires additional studies. Another limitation is the small number of patients in groups 4 and 2. This problem is caused by the rare occurrence of BCB dysfunction in patients with syphilis. Therefore, our analyses may not have statistical power to detect some findings. We do not regard the fact that 98% of our patients are HIV negative as a limitation because this raises our confidence that the detected CSF abnormalities have been caused by syphilitic infection. In terms of practice, our study argues that a large amount of treponemal antibodies in CSF is more likely a consequence of CNS pathology, rather than just a mass of blood-derived antibodies. We suggest using the criterion of a CSF-TPHA titer greater than 320 in untreated patients with syphilis, in whom neurosyphilis is suspected, but CSF-VDRL is nonreactive. The applicability of this criterion for making a diagnosis of neurosyphilis in patients with the history of syphilis treatment is doubtful because it is unknown whether this finding indicates inadequate treatment or it is a scar syndrome in case of cured infection. A search for CSF biomarkers associated with an active phase of intrathecal immune response is necessary. The recently proposed CXCL13 level15 seems to meet these requirements. However, a CSF-TPHA titer is much more accurate for ruling in, rather than ruling out neurosyphilis. In our opinion, a low CSF treponemal antibody level or even a nonreactive result of treponemal tests does not necessarily exclude neurosyphilis because the features of CNS inflammatory focus, especially its location in relation to the CSF spaces, can significantly affect the CSF antibody concentration. A recent publication has proved that there is a possibility of spirochetal infection of the CNS without laboratory signs of intrathecal antibody synthesis.16 Overall, our study has shown that all currently used criteria for the diagnosis of neurosyphilis in CSF are different assessment tools of intrathecal humoral immune activity and support the
922
hypothesis that high CSF treponemal-specific antibody titers are a consequence of inflammatory pathology of the CNS. REFERENCES 1. Workowski KA, Berman S. Centers for Disease Control and Prevention (CDC). Sexual transmitted disease treatment guidelines, 2010. MMWR Recomm Rep 2010; 59:1Y11. 2. French P, Gomberg M, Janier M, et al. IUSTI: 2008 European guidelines on the management of syphilis. Int J STD AIDS 2009; 20:300Y309. 3. Jaffe HW, Larsen SA, Peters M, et al. Tests for treponemal antibody in CSF. Arch Intern Med 1978; 138:252Y255. 4. Luger AF, Schmidt BL, Kaulich M. Significance of laboratory findings for the diagnosis of neurosyphilis. Int J STD AIDS 2000; 11:224Y34. 5. Schmidt BL, Van Voorst Vader PC. Laboratory diagnosis of neurosyphilis in Europe. Presented at: IUSTI/WHO Conference on STI. Europe Syphilis Guideline Expert Workshop; 2004; Mykonos. 6. Harding A, Ghanem K. The performance of cerebrospinal fluid treponemal-specific antibody tests in neurosyphilis: A systematic review. Sex Transm Dis 2012; 39:291Y97. 7. Prange HW. Neurosyphilis. In: Gross G, Tyring S, eds. Sexually Transmitted Infections and Sexually Transmitted Diseases. Berlin: Springer; 2011:163Y171. 8. Reiber H, Peter JB. Cerebrospinal fluid analysisVDisease related data patterns and evaluation programs. J Neurol Sci 2001; 184:101Y122. 9. Vartal F, Vandvik B, Michaelsen TE, et al. Neurosyphilis: Intrathecal synthesis of oligoclonal antibodies to Treponema pallidum. Ann Neurol 1982; 11:35Y40. 10. Hische E, Tutuarima J, Wolters E, et al. Cerebrospinal fluid IgG and IgM indexes as indicators of active neurosyphilis. Clin Chem 1988; 34:665Y667. 11. Lo¨whagen GB, Andersson M, Blomstrand C, et al. Central nervous system involvement in early syphilis. Part 1. Intrathecal immunoglobulin production. Acta Derm Venereol 1983; 63:409Y417. 12. Brettschneider J, Claus A, Kassubek J, et al. Isolated bloodYcerebrospinal fluid barrier dysfunction: Prevalence and associated diseases. J Neurol 2005; 252:1067Y1073. 13. Deisenhammera F, Bartosb A, Egga R, et al. Guidelines on routine cerebrospinal fluid analysis. Report from an EFNS task. Eur J Neurol 2006; 13:913Y922. 14. Jarius S, Eichhorn P, Wildemann B, et al. Usefulness of antibody index assessment in cerebrospinal fluid from patients negative for total-IgG oligoclonal bands. Fluids Barriers CNS 2012; 9:14. 15. Marra CM, Tantalo LC, Sahi SK, et al. CXCL13 as a cerebrospinal fluid marker for neurosyphilis in HIV-infected patients with syphilis. Sex Transm Dis 2010; 37:283Y287. 16. Vrethem M, Widhe M, Ernerudh J, et al. Clinical, diagnostic and immunological characteristics of patients with possible neuroborreliosis without intrathecal Ig-synthesis against Borrelia antigen in the cerebrospinal fluid. Neurol Int 2011; 3:e2.
Sexually Transmitted Diseases
&
Volume 40, Number 12, December 2013
Copyright © 2013 by the American Sexually Transmitted Diseases Association. Unauthorized reproduction of this article is prohibited.
