Lumbar puncture for neurosyphilis investigation in asymptomatic patients with HIV-syphilis coinfection: a cross-sectional study among infectious disease specialists

ABSTRACT BACKGROUND: Syphilis is a major public health issue worldwide. In people living with human immunodeficiency virus (PLHIV), there are higher incidences of both syphilis and neurosyphilis. The criteria for referring PLHIV with syphilis for lumbar puncture is controversial, and the diagnosis of neurosyphilis is challenging. OBJECTIVE: To describe the knowledge, attitudes, and practices of infectious disease specialists and residents in the context of care for asymptomatic HIV-syphilis coinfection using close-ended questions and case vignettes. DESIGN AND SETTING: Cross-sectional study conducted in three public health institutions in São Paulo (SP), Brazil. METHODS: In this cross-sectional study, we invited infectious disease specialists and residents at three academic healthcare institutions to answer a self-completion questionnaire available online or in paper form. RESULTS: Of 98 participants, only 23.5% provided answers that were in line with the current Brazilian recommendation. Most participants believed that the criteria for lumbar puncture should be extended for people living with HIV with low CD4+ cell counts (52.0%); in addition, participants also believed that late latent syphilis (29.6%) and Venereal Disease Research Laboratory (VDRL) titers ≥ 1:32 (22.4%) should be conditions for lumbar puncture in PLHIV with no neurologic symptoms. CONCLUSION: This study highlights heterogeneities in the clinical management of HIV-syphilis coinfection. Most infectious disease specialists still consider syphilis stage, VDRL titers and CD4+ cell counts as important parameters when deciding which patients need lumbar puncture for investigating neurosyphilis.


INTRODUCTION
Syphilis is a major public health problem with increasing occurrence in several countries. In Brazil, data from the Ministry of Health show a three-fold increase in syphilis detection between 2014 and 2018, with incidence rates escalating from 25.1 to 75.8 cases per 100,000 personyears. 1 Similar trends have also been reported in the United States, with a two-fold increase between 2014 and 2018, 2 and in Europe, with greater risk among men who have sex with men. 3 The prevalence of syphilis among people living with human immunodeficiency virus (PLHIV) is higher than in the general population. Studies performed in Brazil suggest that the prevalence of syphilis ranges from 2.7 to 20.5% among PLHIV; 4-6 similarly, syphilis coinfection has been reported in 1%-21% of PLHIV in North America and 2%-43% in Europe. 7 Besides its local manifestations, Treponema pallidum has systemic effects, notably, in the central nervous system. Conclusive diagnostic investigation of neurosyphilis may be challenging in the context of HIV coinfection, since serological and chemocytological abnormalities of the cerebrospinal fluid (CSF) may occur in PLHIV even without neurosyphilis. Moreover, given the high incidence of re-exposure to syphilis, the interpretation of the serological response after treatment may be challenging in this population. [8][9][10] One of the most debated topics in the management of syphilis is the need and timing of CSF examination in HIV-syphilis coinfected patients with no neurologic symptoms. Guidelines and recommendations have been changing regarding this topic. Prior studies recommended a more aggressive approach with lumbar puncture based on CD4 + cell count, Venereal Disease Research Laboratory (VDRL) titers [11][12][13] or syphilis stage. 14,15 However, a less invasive approach suggests performing lumbar puncture based on criteria that are similar to those applied to HIV-uninfected individuals. 16 are an obstacle to effective adherence to guidelines. 18 Other potential barriers include physicians' lack of familiarity, agreement, or motivation for specific guidelines, favoring the persistence of previous practices. External factors including the inability to reconcile patient preferences, lack of time, lack of resources and organizational constraints also play a role in heterogenous practices. 18,19 Based on our routine observation, we hypothesized that some providers may tailor decisions regarding lumbar puncture based on barriers faced to perform the exam (i.e., long waiting time, lack of trained practitioners, lack of an appropriate procedure room) or difficulties to implement neurosyphilis treatment after the diagnosis (i.e., absence of hospital service and long waiting time for hospitalization).
Few studies have investigated the knowledge and attitudes of healthcare providers regarding the management of syphilis-HIV coinfection, [20][21][22][23][24] and studies exploring attitudes on the investigation of asymptomatic neurosyphilis in PLHIV are even more scarce. 25

OBJECTIVE
Our aim was to describe the knowledge, attitudes, and practices of infectious disease specialists in the context of asymptomatic HIV-syphilis coinfection using close-ended questions and case vignettes. We also explored if attitudes and practices of providers who report difficulties for lumbar puncture procedure and/or neurosyphilis in-hospital treatment varied among participants.

METHODS
In this cross-sectional study, we invited infectious disease specialists and residents from three public and academic healthcare institutions in São Paulo, Brazil, to answer a self-completion questionnaire. The institutions were selected based on the antici- Written informed consent was obtained from all participants, and no identifiable information was collected during the study.

Ethical aspects
The study was approved by the Ethics Committee at the coordi-

Participant characteristics
Between December 2019 and September 2020, 98 infectious disease specialists or residents responded to the survey.
The demographics, training, and practice characteristics are described in Table 1 Comparisons of the demographics, training, and practice characteristics according to group categorization are presented in Table 1.

Responses to case vignettes
In the second section of the questionnaire, case vignettes with hypothetical situations addressing neurosyphilis investigation with lumbar punctures and interpretation of CSF laboratory reports were presented to participants, as described in Tables 2 and 3.
The first two vignettes described a PLHIV with early latentstage syphilis and a VDRL titer of 1:128. When the CD4 + cell count was above 350 cells/mm 3 , 21.3% of respondents referred the patient for lumbar puncture; this percentage rose to 65.3% when the CD4 + cell count was below 350 cells/mm 3 .
The third vignette described a patient with early latent syphilis with a CD4 + cell count above 350 cell/mm 3 and a VDRL titer of 1:128 with a four-fold (two dilution) decrease in the titer within 12 months after adequate treatment. According to 67.7% of the respondents, this patient should be referred for lumbar puncture.
The fourth and fifth vignettes presented a patient recently diagnosed with HIV infection, with latent syphilis of unknown duration.
When the vignette described a patient with a CD4 + cell count of 110 cells/mm 3 and a VDRL titer of 1:4, 51.6% of participants referred the patient for lumbar puncture. When the case presented a patient with a CD4 + cell count above 350 cells/mm 3 and VDRL titer of 1:32, the 40.8% of respondents referred the patient to lumbar puncture.
We found no statistically significant differences between Groups 1 and 2 in the answers to case vignettes 1-5 ( Table 2).
In the five vignettes addressing the interpretation of CSF laboratory reports, we presented hypothetical patients with different CD4 + cell counts and chemocytological findings in CSF. For all situations, the treponemal serological test was reactive, while VDRL Table 1. Demographics, training, and practice characteristics of study participants, overall and according to group category * Missing for one participant; IQR, interquartile range; ** Group 1: Participants reporting no difficulties for lumbar puncture or patient hospitalization; *** Group 2: Participants reporting at least some difficulties for lumbar puncture and/or those who perceived patient hospitalization as very difficult.
Again, we found no statistically significant differences between Groups 1 and 2 in responses to case vignettes in this section.

Knowledge and attitudes regarding lumbar puncture criteria and syphilis clinical management
The 2018 Ministry of Health recommendations in Brazil suggest the use of lumbar puncture for neurosyphilis investigation in PLHIV with syphilis coinfection in the following situations: presence of neurological or ophthalmic symptoms, evidence of active tertiary syphilis, and after antibiotic treatment failure, independently of presumed sexual re-exposure. 17  We did not explore whether the responders considered ceftriaxone a reliable first-line treatment.

DISCUSSION
The results of this cross-sectional study highlight heterogene- following a guideline for up to 89% of physicians. 18 We believe that heterogeneities and recent modifications regarding recommendations for lumbar puncture among PLHIV across local and international guidelines are also likely to contribute to this low percentage of correct answers. Adherence to guideline recommendations could also be influenced by environmental-related barriers. 18 16,17,28 The 2020 European guideline on the management of syphilis highlights that robust evidence is lacking, but reiterate that some experts recommend CSF assessment in asymptomatic PLHIV with late syphilis and CD4 + cells ≤ 350/mm 3 and/or a serum VDRL/RPR titer > 1:32. 29 The 2020 German guidelines on the diagnosis and treatment of neurosyphilis consider CD4 + cell counts, HIV treatment status, and VDRL titers in the decision for lumbar puncture among PLHIV with no neurologic symptoms. 30 The incidence of neurosyphilis is demonstrably higher among PLHIV compared to that in the general population. 31,32 Additionally, higher VDRL titers and lower CD4 + cell counts have been associated with the development of neurosyphilis in this population. 33 36 The CSF protein level is neither specific nor sensitive, 37 but it is nevertheless considered for defining neurosyphilis in many published papers 33,34,38 since higher levels can be associated with neurosyphilis with cutoffs that vary from 45 to 50 mg/dL. 13 In our study, elevated cell and protein levels were considered as criteria for neurosyphilis by 59% and 50% of participants, respectively, when CSF VDRL was negative and CSF treponemal was positive.
Neurosyphilis treatment was addressed in one multicenter clinical trial including 36 PLHIV with syphilis coinfection. The authors randomized participants to receive either ceftriaxone 2 g/day or Penicillin G 24 million units/day for 10 days. Only 30 patients were included in the final analysis and the study failed to find differences between groups in the proportions of subjects with improvements in CSF cell count or protein levels. 39 Due to scarcity of data and study limitations, the evidence is insufficient to allow the adoption of ceftriaxone as a first-line treatment for neurosyphilis. 40 In our survey, all respondents accepted penicillin G as the antibiotic of choice and only 44% indicated ceftriaxone as a reliable option.
Our study had some limitations. Only 32.7% of all eligible infectious disease clinicians working in the participating sites responded to the questionnaire, which might have resulted in selection bias.