Skip Navigation

Normalization of Cerebrospinal Fluid Abnormalities after Neurosyphilis Therapy: Does HIV Status Matter?

  1. Christina M. Marra1,2,
  2. Clare L. Maxwell1,
  3. Lauren Tantalo1,
  4. Molly Eaton3,
  5. Anne M. Rompalo4,
  6. Charles Raines4,
  7. Bradley P. Stoner5,
  8. James J. Corbett6,
  9. Michael Augenbraun7,
  10. Mark Zajackowski8,
  11. Romina Kee8, and
  12. Sheila A. Lukehart2
  1. 1Departments of Neurology and University of Washington School of Medicine, Seattle, Washington
  2. 2Departments of Medicine, Division of Infectious Diseases, University of Washington School of Medicine, Seattle, Washington
  3. 3Department of Medicine, Emory University, Atlanta, Georgia
  4. 4Department of Medicine, Johns Hopkins University, Baltimore, Maryland
  5. 5Department of Medicine, Washington University, St. Louis, Missouri
  6. 6Department of Neurology, University of Mississippi, Jackson, Mississippi
  7. 7Department of Medicine, State University of New York—Downstate, Brooklyn, New York
  8. 8Chicago Department of Health, Chicago, Illinois
  1. Reprints or correspondence: Dr. Christina M. Marra, Harborview Medical Center Box 359775, 325 9th Ave., Seattle, WA 98104-2499 (cmarra{at}u.washington.edu).

  1. Presented in part: 55th Annual Meeting of the American Academy of Neurology, 29 March through 5 April 2003, Honolulu, Hawaii (abstract P03.017).

 
Next Section

Abstract

To identify factors that affect normalization of laboratory measures after treatment for neurosyphilis, 59 subjects with neurosyphilis underwent repeated lumbar punctures and venipunctures after completion of therapy. The median duration of follow-up was 6.9 months. Stepwise Cox regression models were used to determine the influence of clinical and laboratory features on normalization of cerebrospinal fluid (CSF), white blood cells (WBCs), CSF protein concentration, CSF Venereal Disease Research Laboratory (VDRL) reactivity, and serum rapid plasma reagin (RPR) titer. Human immunodeficiency virus (HIV)—infected subjects were 2.5 times less likely to normalize CSF-VDRL reactivity than were HIV-uninfected subjects. HIV-infected subjects with peripheral blood CD4+ T cell counts of ⩽200 cells/µL were 3.7 times less likely to normalize CSF-VDRL reactivity than were those with CD4+ T cell counts of >200 cells/µL. CSF WBC count and serum RPR reactivity were more likely to normalize but CSF-VDRL reactivity was less likely to normalize with higher baseline values. Future studies should address whether more intensive therapy for neurosyphilis is warranted in HIV-infected individuals.

The Centers for Disease Control and Prevention (CDC) recommends 10–14 days of high-dose intravenous penicillin G or intramuscular procaine penicillin G plus oral probenecid for treatment of symptomatic or asymptomatic neurosyphilis [1]. These recommendations are based on the ability of such regimens to achieve sufficient penicillin concentrations in CSF to kill Treponema pallidum, the bacterium that causes syphilis, and on clinical experience. However, rigorous clinical trials of the 2 penicillin regimens have not been conducted. A small study suggested that intravenous ceftriaxone may be an acceptable alternative to penicillin in HIV-infected individuals with neurosyphilis and concomitant early syphilis [2], but these results have not been replicated in a large trial.

The success of neurosyphilis therapy is judged by resolution or stabilization of clinical abnormalities and by resolution of CSF abnormalities. Specifically, the CDC guidelines state that the WBC count in CSF should decrease at 6 months after completion of treatment, although the extent of this decrease is not specified, and that all CSF abnormalities should resolve by 2 years after completion of treatment [1]. Several reports have described clinical and laboratory-defined treatment failures in HIV-infected patients with neurosyphilis [3–5], but the overall frequency of treatment failure in such individuals is not known. The success of neurosyphilis therapy may be affected by several factors. For example, different treatment regimens may have different efficacies. The stage of syphilis at which neurosyphilis is identified may influence the rate or likelihood of normalization of CSF abnormalities, as is true for serum nontreponemal test titers [6, 7]. Furthermore, such abnormalities may resolve spontaneously early in the course of syphilis [8–12]. HIV infection itself may cause CSF abnormalities, such as mild pleocytosis and elevated protein concentrations, which may persist after neurosyphilis therapy [13]. The goal of this study was to identify factors that influence normalization of laboratory abnormalities after completion of treatment for neurosyphilis.

Previous SectionNext Section

Methods

The study protocol and laboratory methods are described elsewhere [14]. In brief, subjects with syphilis underwent lumbar puncture, in accordance with the 1993 CDC guidelines [15]. These guidelines were used because study entry began in 1996, before the availability of the newer guidelines [1]. Subjects were treated for neurosyphilis by the enrolling physician on the basis of clinical and laboratory findings. Regimens included high-dose intravenous penicillin (12–24 million units [MU] q.d., divided into 6 doses, as recommended in the 1993 CDC guidelines; 18–24 MU q.d., divided into 6 doses, as recommended in the 1998 CDC guidelines; or 18–24 MU q.d., divided into 6 doses or given as a continuous infusion, as recommended in the 2002 guidelines) [1, 15, 16], intramuscular procaine penicillin (PPG) plus oral probenecid (2.4 MU PPG daily plus 500 mg probenecid 4 times a day) [1, 15, 16], or intravenous ceftriaxone (2.0 g q.d.) [1, 2, 15]. The follow-up protocol called for repeated lumbar puncture 3 months after completion of therapy for all subjects and 6 and 12 months after completion of therapy for subjects with persistent CSF abnormalities at the previous assessment; blood samples were obtained 3, 6, and 12 months after completion of therapy. If a subject missed the scheduled visit but returned later, they were allowed to complete the visit late, but subjects were not allowed to complete >4 study visits. In this analysis, neurosyphilis is defined as a CSF WBC count of >20 cells/µL or a reactive CSF Venereal Disease Research Laboratory (CSF-VDRL) test result. Subjects were included in the analysis if they met this definition of neurosyphilis at the entry visit and if they were treated for ⩾10 days with 1 of the 3 treatment regimens described above.

The 59 subjects were enrolled and followed at the following 8 sites: University of Washington (Seattle; 38 subjects); Emory University (Atlanta; 9 subjects); Johns Hopkins University (Baltimore; 6 subjects); Washington University (St. Louis; 3 subjects); State University of New York (Brooklyn, NY; 1 subject); University of Mississippi (Jackson; 1 subject); and the Chicago Department of Health (1 subject). Results of serological tests for HIV-1 and peripheral blood CD4+ T lymphocyte counts were obtained from review of medical records. Only peripheral blood CD4+ T lymphocyte counts determined within 90 days before or after the first the lumbar puncture were used. Treatment for a previous episode of nonneurological syphilis before study entry was recorded, and the interval between the time of the treatment and of the lumbar puncture was calculated. Treatment up to 14 days before the lumbar puncture was considered to have little impact on CSF findings; thus, subjects who were not treated for syphilis before the lumbar puncture and those who were treated within 14 days of the lumbar puncture were grouped together in the “no treatment” group. Written informed consent was obtained from all subjects, and the guidelines for human experimentation of the US Department of Health and Human Services and of the authors' institutions were followed in the conduct of this research.

Associations between categorical variables were assessed using the χ2 test or Fisher's exact test. Associations between continuous variables and categorical variables were assessed using the Mann-Whitney U test. HIV infection causes mild CSF pleocytosis and mild elevations in the CSF protein level [13]; thus, we used a limit of ⩽20 cells/µL as normal for CSF WBC count and ⩽50 mg/dL as normal for CSF protein level, because values greater than these limits are uncommonly due to HIV infection in the absence of syphilis. For the CSF-VDRL and the serum rapid plasma reagin (RPR) test, a 4-fold decrease in titer or reversion to a nonreactive result was defined as a normal response. Time to normalization of each measure was estimated by Kaplan Meier analysis. Stepwise Cox regression models were used to determine the influence of the following factors on the likelihood of normalization of each measure: (1) neurosyphilis treatment regimen (intravenous ceftriaxone, vs. intravenous aqueous penicillin G, vs. intramuscular procaine penicillin with oral probenecid), (2) no prior treatment for nonneurological syphilis (treatment 0–14 days before the lumbar puncture) versus prior nonneurological syphilis treatment (treatment ⩾15 days before the lumbar puncture), (3) syphilis stage (secondary and early latent vs. late latent and syphilis of unknown duration), (4) baseline laboratory values (greater than or less than the median value for those subjects with each abnormality), and (5) HIV infection status. Two-tailed P values of <.05 were considered to be statistically significant. The final hazard ratios (HRs) were estimated from models that considered only the factors found significant in the stepwise procedure.

Previous SectionNext Section

Results

The baseline characteristics of the 59 subjects are shown in table 1. Most subjects were men and were HIV infected. Compared with HIV-uninfected subjects, more HIV-infected subjects had early-stage syphilis (P = .03). The baseline serum RPR titer was significantly higher in HIV-infected subjects; this likely reflects the higher proportion of subjects with early-stage syphilis. Symptoms and signs were categorized into those most consistent with syphilitic meningitis (headache, stiff neck, photophobia, and subjective hearing loss or an abnormal finger friction test result for either ear) and those most consistent with syphilitic ocular disease (subjective decrease in vision, ocular inflammation, or abnormal visual acuity). Twenty-six subjects (44.1%) met the definition of meningitis alone, 12 (20.3%) met the definition of ocular disease alone, and 9 (15.2%) met the definition for both conditions. The median duration of follow-up after completion of treatment was 6.9 months (range, 3.0–28.6 months). Baseline laboratory measures are shown in table 2. The values for CSF WBC, CSF protein concentration, and CSF-VDRL titer were comparable in HIV-uninfected and HIV-infected subjects.

Figure 1
Figure 1

Kaplan Meier plots of normalization of each CSF measure and of serum rapid plasma reagin (RPR) reactivity. See Methods for definitions of normalization. VDRL, Venereal Disease Research Laboratory.

Table 1
Table 1

Baseline characteristics of the 59 subjects.

Table 2
Table 2

CSF and serum measures in HIV-infected and HIV-uninfected subjects before treatment for neurosyphilis.

The median time to normalization of each laboratory measure is shown in table 3. As shown in figure 1, the CSF WBC count normalized in all subjects 12.4 months after completion of treatment, and CSF-VDRL titer normalized by 14.2 months after treatment in 94% of subjects. In contrast, the CSF protein concentration remained elevated 12.7 months after treatment in 25% of subjects, and serum RPR reactivity had not normalized in 11% of subjects 15.8 months after treatment.

Table 3
Table 3

Time to normalization of laboratory measures after treatment for neurosyphilis.

The specific neurosyphilis treatment regimen did not influence normalization of any of the 4 laboratory measures. Factors that were included in the final regression models of normalization of each laboratory measure and their HRs are shown in table 4. Normalization of the CSF WBC count was more likely in subjects with a higher CSF WBC count at baseline (P = .03) and tended to be more likely in those who had not been treated for syphilis before the lumbar puncture (P = .05). Normalization of the CSF protein concentration was more likely in subjects who were not treated for syphilis before the lumbar puncture (P = .006). CSF-VDRL reactivity was less likely to normalize in HIV-infected subjects (P = .03) and in subjects with higher baseline CSF-VDRL titers (P = .01). CSF-VDRL reactivity also tended to be more likely to normalize in subjects with early syphilis (P = .14). Among HIV-infected subjects, those with peripheral blood CD4+ T cell counts of >200 cells/µL were more likely to normalize CSF-VDRL reactivity than were those with peripheral blood CD4+ T cell counts of ⩽200 cells/µL (HR, 3.7; 95% CI, 1.2–11.2; P = .02). Serum RPR reactivity was more likely to normalize in subjects with higher baseline serum RPR titers (P = .0001) and in those who were not treated for nonneurological syphilis before the lumbar puncture (P = .007).

Table 4
Table 4

Factors that influence normalization of each laboratory measure in the final Cox regression models.

Previous SectionNext Section

Discussion

Because T. pallidum cannot be cultured in the laboratory, failure to detect the organism in CSF specimens after completion of treatment is not used as a means of determining the efficacy of therapy. Instead, the laboratory assessment of response to treatment is determined by resolution of other CSF abnormalities, including elevated CSF WBC count, elevated CSF protein concentration, and reactive CSF-VDRL tests. Normalization of these measures is taken as a surrogate for clearance of T. pallidum from the CSF. Our results show that CSF abnormalities, particularly CSF pleocytosis and reactive CSF-VDRL, normalize rapidly after receipt of specific treatment for neurosyphilis. In contrast, the CSF protein concentration normalized more slowly and incompletely, despite resolution of the other CSF abnormalities that may be more specific to neurosyphilis. These results suggest that, compared with other CSF measures, it may be more difficult to define an appropriate time or rate of normalization of CSF protein concentration and that persistently elevated CSF protein concentrations may not necessarily be an indication of treatment failure.

We used stepwise Cox regression to analyze the likelihood of normalization of CSF and serum abnormalities after completion of treatment for neurosyphilis. HIV-infected subjects were 2.5 times less likely to normalize CSF-VDRL reactivity, even after taking into account baseline CSF-VDRL titer and stage of syphilis at which neurosyphilis was diagnosed. Among the HIV-infected subjects, those with peripheral blood CD4+ T cell counts of ⩽200 cells/µL were 3.7 times less likely to normalize CSF-VDRL reactivity than were those with CD4 counts of >200 cells/µL. These results suggest that HIV-induced immune impairment contributes to the slower normalization seen in HIV-infected persons and may support the contention advanced by us [5] and by others [17] that clearance of CNS organisms may be impaired by concomitant HIV infection.

The baseline CSF WBC count, CSF-VDRL titer, and serum RPR test titer influenced the likelihood of normalization of each value. For CSF WBC count and serum RPR reactivity, higher baseline values were associated with a greater likelihood of normalization. Previous studies have shown that higher serum nontreponemal antibody titers decrease faster than do lower titers [6, 7], but these studies did not specifically examine the likelihood of normalization as our study did. Subjects with a pretreatment CSF-VDRL titer >1 : 1 (the median value for patients with reactive CSF-VDRL results) were 5 times less likely to normalize CSF-VDRL reactivity than were those with a titer of ⩽1 : 1. The explanation for this finding is speculative. One possibility is that higher CSF-VDRL titer reflects greater concentration of T. pallidum in CSF, leading to greater antigenic stimulation. More time would be required to clear a higher organism burden; thus, the likelihood of normalization of CSF-VDRL reactivity at the end of a given period of observation would be lower. A potential flaw in this argument is that higher peripheral blood VDRL or RPR titers might also reflect greater concentration of T. pallidum in blood, yet these decrease faster when the baseline titer is higher [6, 7].

CSF protein concentrations and serum RPR reactivity were more likely to normalize in untreated subjects than in subjects who were treated with standard therapy for nonneurological syphilis before the lumbar puncture, and CSF WBC counts also tended to be more likely to normalize in subjects who had not previously been treated. Serum RPR normalizes more slowly in individuals with previous syphilis [6], and this observation may explain our finding. However, it is more difficult to explain why previous receipt of treatment for nonneurological syphilis would influence the likelihood or rate of normalization of CSF protein concentration or CSF WBC count. This finding is not associated with higher baseline concentrations, because the baseline CSF protein value did not influence the likelihood of normalization of the CSF protein concentration, and because the baseline CSF WBC count was included in the analysis of normalization of the CSF WBC count. One possibility relates to a theory put forward in the era before penicillin was available, which is that syphilis therapy that treats systemic but not neurologic infection, as intramuscular benzathine penicillin G therapy would do, down-regulates the immune response and augments CNS infection [18, 19]. Thus, the CSF protein concentration and the CSF WBC count may normalize more slowly in subjects treated for syphilis before undergoing a lumbar puncture, because the immune response has been rendered less effective.

We used a definition of neurosyphilis that included either a reactive CSF-VDRL or a CSF WBC count of >20 cells/µL. This cutoff for defining an abnormal CSF WBC count was chosen because it is above the mild pleocytosis due to HIV infection itself [13]. The observation that all CSF WBC counts >20 cells/µL resolved after receipt of neurosyphilis therapy further supports the validity of our definition. Although our prospective study is the largest modern analysis to examine factors that influence normalization of laboratory measures after receipt of neurosyphilis treatment, it has limitations. The number of HIV-uninfected subjects was small, and this limits our ability to detect differences between the HIV-infected and HIV-uninfected persons. Moreover, the number of subjects included in our regression analyses ranged from 23 to 56 persons. Small changes in these numbers could influence which individual factors affect normalization of the laboratory measures. Nonetheless, several of our findings show a consistent pattern across outcomes. Concomitant HIV infection significantly decreased the likelihood of resolution of CSF-VDRL reactivity. The significant association with low peripheral blood CD4+ T cell count confirms the validity of the finding. The baseline values for the laboratory measures affected the likelihood of normalization of all measures except CSF protein concentration. Moreover, our findings regarding greater likelihood of normalization of serum RPR reactivity with a higher baseline titer and the lesser likelihood of normalization in patients who had previously been treated for nonneurological syphilis are consistent with the results of other studies [6, 7].

The practical implications of our findings for the clinician are that CSF WBC count and CSF-VDRL titer normalize in virtually all patients with neurosyphilis within 1 year after completion of therapy. CSF WBC counts are more likely to normalize in patients with a higher baseline value, and CSF-VDRL reactivity is less likely to normalize in patients with a higher baseline titer. The CSF protein concentration is slow to normalize and may remain abnormal when other measures have become normal, suggesting that it may not be a good indicator of the success or failure of therapy. Concomitant HIV infection decreases the likelihood of normalization of CSF-VDRL, and this effect is particularly evident in patients with significant immunosuppression, as evidenced by a peripheral blood CD4+ T cell count of <200 cells/µL. Although we do not know whether decreased likelihood of normalization during the observation period is equivalent to treatment failure, the difference between HIV-infected and HIV—uninfected subjects is concerning. These results raise the question of whether HIV-infected individuals, particularly those with low peripheral blood CD4+ T cell counts, might require more-intensive therapy for neurosyphilis than do HIV-uninfected subjects. Future research should address this question.

Previous SectionNext Section

Footnotes

  • Financial support: National Institutes of Health (NIH)/National Institute of Neurological Disorders and Stroke (NINDS) (grant 34235 to C.M.M.) and NIH/NINDS (grant 38663 to C.M.M.).

  • Received August 29, 2003.
  • Revision received October 30, 2003.
  • Accepted March 16, 2004.
Previous Section
 

References

  1. ↵
    1. Centers for Disease Control Prevention
    . Sexually transmitted diseases treatment guidelines 2002. MMWR Morb Mortal Wkly Rep 2002;51:18-30.
    Medline
  2. ↵
    1. Marra CM,
    2. Boutin P,
    3. McArthur JC,
    4. et al
    . A pilot study evaluating ceftriaxone and penicillin G as treatment agents for neurosyphilis in human immunodeficiency virus-infected individuals. Clin Infect Dis 2000;30:540-4.
    Abstract/FREE Full Text
  3. ↵
    1. Malone JL,
    2. Wallace MR,
    3. Hendrick BB,
    4. et al
    . Syphilis and neurosyphilis in a human immunodeficiency virus type-1 seropositive population: evidence for frequent serologic relapse after therapy. Am J Med 1995;99:55-63.
    CrossRefMedlineWeb of Science
    1. Gordon SM,
    2. Eaton ME,
    3. George R,
    4. et al
    . The response of symptomatic neurosyphilis to high-dose intravenous penicillin G in patients with human immunodeficiency virus infection. N Engl J Med 1994;331:1469-73.
    CrossRefMedlineWeb of Science
  5. ↵
    1. Marra CM,
    2. Longstreth WT Jr,
    3. Maxwell CL,
    4. Lukehart SA
    . Resolution of serum and cerebrospinal fluid abnormalities after treatment of neurosyphilis: influence of concomitant human immunodeficiency virus infection. Sex Transm Dis 1996;23:184-9.
    MedlineWeb of Science
  6. ↵
    1. Romanowski B,
    2. Sutherland R,
    3. Fick GH,
    4. Mooney D,
    5. Love EJ
    . Serologic response to treatment of infectious syphilis [see comments]. Ann Intern Med 1991;114:1005-9.
    Abstract/FREE Full Text
  7. ↵
    1. Brown ST,
    2. Zaidi A,
    3. Larsen SA,
    4. Reynolds GH
    . Serological response to syphilis treatment: a new analysis of old data. JAMA 1985;253:1296-9.
    Abstract/FREE Full Text
  8. ↵
    1. Moore J
    . Studies in asymptomatic neurosyphilis. II. The classification, treatment, and prognosis of early asymptomatic neurosyphilis. Bull Johns Hopkins Hosp 1922;33:231-46.
    Web of Science
    1. Ravaut P
    . Le liquide céphalo-rachidien des syphiliques en période secondaire. Ann Dermatol Syphiligr 1903;4:537-54.
    1. Wile UJ,
    2. Stokes JH
    . A study of the spinal fluid with reference to involvement of the nervous system in secondary syphilis. J Cutan Dis 1914;32:607-23.
    1. Fildes P,
    2. Parnell RJG,
    3. Maitland HB
    . The occurrence of unsuspected involvement of the central nervous system in unselected cases of syphilis. Brain 1918;41:255-301.
    FREE Full Text
  12. ↵
    1. Mills CH
    . Routine examination of the cerebrospinal fluid in syphilis: its value in regard to more accurate knowledge, prognosis and treatment. BMJ 1927;2:527-32.
    FREE Full Text
  13. ↵
    1. Collier AC,
    2. Marra C,
    3. Coombs RW,
    4. et al
    . Central nervous system manifestations in human immunodeficiency virus infection without AIDS. J Acquir Immune Defic Syndr 1992;5:229-41.
    MedlineWeb of Science
  14. ↵
    1. Marra CM,
    2. Maxwell CL,
    3. Smith SL,
    4. et al
    . Cerebrospinal fluid abnormalities in patients with syphilis: association with clinical and laboratory features. J Infect Dis 2004;189:369-76.
    Abstract/FREE Full Text
  15. ↵
    1. Centers for Disease Control Prevention
    . 1993 Sexually transmitted diseases treatment guidelines. MMWR Recomm Rep 1993;42:1-102.
    Medline
  16. ↵
    1. Centers for Disease Control Prevention
    . 1998 Guidelines for treatment of sexually transmitted diseases. MMWR Recomm Rep 1998;47(RR-1):1-111.
    Medline
  17. ↵
    1. Musher DM
    . Syphilis, neurosyphilis, penicillin, and AIDS. J Infect Dis 1991;163:1201-6.
    Abstract/FREE Full Text
  18. ↵
    1. Moore JE
    . The relation of neurorecurrences to late syphilis: a clinical study of eighty-one cases. Arch Neurol Psychiatr 1929;21:117-36.
  19. ↵
    1. Merritt HH,
    2. Moore M
    . Acute syphilitic meningitis. Medicine (Baltimore) 1935;14:119-83.
    CrossRef
| Table of Contents

This Article

  1. Clin Infect Dis. (2004) 38 (7): 1001-1006. doi: 10.1086/382532
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most