# paper.pdf ## Metadata - PDFFormatVersion=1.6 - Language=en - IsLinearized=true - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - CreationDate=D:20260713214946+05'30' - Creator=Servigistics Arbortext Advanced Print Publisher 11.1.4546/W-x64 - Custom.EPSprocessor=PStill version 1.84.42 - ModDate=D:20260713214946+05'30' - Producer=PDFlib+PDI 9.0.7p3 (C++/Win64) - xmpmm:instanceid=uuid:DE1D7CE6-A40F-5178-76FD-28230BC33253 - xmpmm:documentid=uuid:62809F96-B560-9C7A-E16B-5D86247945F6 - xmprights:marked=True - pdf:keywords=ceftriaxone; HIV; limited-resource settings; neurosyphilis; penicillin - dc:format=application/pdf - dc:identifier=10.1093/ofid/ofag387 - dc:lang=en - dc:publisher=Oxford University Press - dc:title=Penicillin or Ceftriaxone for Neurosyphilis in People With HIV: A Retrospective, Propensity Score–Matched Multicenter Study in Mexico - dc:creator=Rodrigo Ville Benavides,Héctor Rivera-Villegas,Obed Ocampo-Valdez,Rafael Franco-Cendejas,Jaime Federico Andrade-Villanueva,Luz Alicia González-Hernández,Pedro Martínez-Ayala,Andrea Cárdenas-Ortega,Paulina Carreño-Pérez,Jezer Lezama-Mora,Juan Carlos Rodríguez-Aldama,Edgar Pérez-Barragán,Pamela Alatorre-Fernández,Estefanía Sienra-Iracheta,Ana Patricia Rodríguez-Zulueta,Norma Eréndira Rivera-Martínez,Brenda Crabtree-Ramírez - dc:description=DOI: 10.1093/ofid/ofag387; Open Forum Infectious Diseases, 13, 7, 2026-7-13.; Abstract: Background: Neurosyphilis, ocular syphilis, and otosyphilis occur more frequently among people with HIV-1 and are associated with higher morbidity and treatment failure. Intravenous penicillin G remains the standard therapy but requires hospitalization or prolonged intravenous access, creating barriers that are amplified in resource-limited settings facing recurrent penicillin shortages. Ceftriaxone is a potential alternative; however, comparative data in people with HIV are limited. Methods: We conducted a multicenter retrospective cohort study of adults with HIV at 5 public hospitals and 2 outpatient HIV clinics in Mexico (2015–2024). Participants had confirmed neurosyphilis, ocular syphilis, or otosyphilis. Participants received intravenous penicillin G (18–24 million IU/day) or ceftriaxone (1–2 g/day) for 10–14 days. The primary outcome was early serological response (ESR), defined as a ≥4-fold nontreponemal titer decline or seroreversion at 6 months. We performed 1:1 propensity score matching (39 pairs, n = 78), followed by logistic regression in the matched cohort. Results: Of 143 participants, 104 (73%) received intravenous penicillin G and 39 (27%) ceftriaxone. In the matched cohort, ESR occurred in 74.4% of penicillin-treated participants and 71.8% of ceftriaxone-treated participants (odds ratio 0.88, 95% confidence interval .32–2.40; P = .80). The results were consistent across all prespecified sensitivity analyses. Conclusions: In this multicenter cohort of people with HIV diagnosed with neurosyphilis, ocular syphilis, or otosyphilis, ceftriaxone showed comparable ESR to intravenous penicillin G. These findings support ceftriaxone as a reasonable alternative when standard therapy is limited by penicillin shortages, limited inpatient capacity, or financial constraints. - dc:subject=ceftriaxone,HIV,limited-resource settings,neurosyphilis,penicillin - dc:rights=© The Author(s) 2026. Published by Oxford University Press on behalf of Infectious Diseases Society of America. - pdf:producer=PDFlib+PDI 9.0.7p3 (C++/Win64) - prism:aggregationtype=Journal - prism:publicationname=Open Forum Infectious Diseases - prism:copyright=© The Author(s) 2026. Published by Oxford University Press on behalf of Infectious Diseases Society of America. - prism:elssn=2328-8957 - prism:volume=13 - prism:number=7 - prism:startingpage=1 - prism:endingpage=9 - prism:doi=10.1093/ofid/ofag387 - prism:url=https://doi.org/10.1093/ofid/ofag387 - pdfaextension:schemas=PDF/UA Universal Accessibility Schemahttp://www.aiim.org/pdfua/ns/id/pdfuaidpartIntegerinternalIndicates, which part of ISO 14289 standard is followed - pdfuaid:part=1 - xmp:createdate=2026-07-13T21:49:46+05:30 - xmp:modifydate=2026-07-13T21:49:46+05:30 - xmp:creatortool=Servigistics Arbortext Advanced Print Publisher 11.1.4546/W-x64 - xmp:metadatadate=2026-07-13T21:49:46+05:30 - pdfx:epsprocessor=PStill version 1.84.42 ## Contents ### Page 1 # Penicillin or Ceftriaxone for Neurosyphilis in People With HIV: A Retrospective, Propensity Score–Matched Multicenter Study in Mexico Rodrigo Ville Benavides,, Héctor Rivera-Villegas,, Obed Ocampo-Valdez,, Rafael Franco-Cendejas,, Jaime Federico Andrade-Villanueva,,Luz Alicia González-Hernández,, Pedro Martínez-Ayala,, Andrea Cárdenas-Ortega,, Paulina Carreño-Pérez,, Jezer Lezama-Mora,Juan Carlos Rodríguez-Aldama,, Edgar Pérez-Barragán,, Pamela Alatorre-Fernández,, Estefanía Sienra-Iracheta,,Ana Patricia Rodríguez-Zulueta,, Norma Eréndira Rivera-Martínez,, and Brenda Crabtree-Ramírez; on behalf of the Alternative Neurosyphilis Treatments Study (ANTS) Group Departamento de Infectología, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City, Mexico, Subdirección de Investigación Biomédica, Instituto Nacional de Rehabilitación Luis Guillermo Ibarra Ibarra, México City, Mexico, Unidad de VIH, Hospital Civil de Guadalajara “Fray Antonio Alcalde,” Centro Universitario de Ciencias de la Salud, Universidad de Guadalajara, Guadalajara, Mexico, Clínica de Inmunocompromiso por Enfermedades Infecciosas (CIENI), Instituto Nacional de Enfermedades Respiratorias Ismael Cosío Villegas, Mexico City, Mexico, Clínica de Sífilis, Clínica Especializada Condesa Cuauhtémoc, Mexico City, Mexico, Departamento de Infectología, Clínica Especializada Condesa Iztapalapa, Mexico City, Mexico, Departamento de Infectología, Instituto Nacional de Cancerología, Mexico City, Mexico, Departamento de Infectología, Hospital General Manuel Gea González, Mexico City, Mexico, and Departamento de Infectología, Hospital Regional de Alta Especialidad de Oaxaca (HRAEO) IMSS Bienestar, Oaxaca de Juárez, Mexico Background. Neurosyphilis, ocular syphilis, and otosyphilis occur more frequently among people with HIV-1 and are associated with higher morbidity and treatment failure. Intravenous penicillin G remains the standard therapy but requires hospitalization or prolonged intravenous access, creating barriers that are amplified in resource-limited settings facing recurrent penicillin shortages. Ceftriaxone is a potential alternative; however, comparative data in people with HIV are limited. Methods. We conducted a multicenter retrospective cohort study of adults with HIV at 5 public hospitals and 2 outpatient HIV clinics in Mexico (2015–2024). Participants had confirmed neurosyphilis, ocular syphilis, or otosyphilis. Participants received intravenous penicillin G (18–24 million IU/day) or ceftriaxone (1–2 g/day) for 10–14 days. The primary outcome was early serological response (ESR), defined as a ≥4-fold nontreponemal titer decline or seroreversion at 6 months. We performed 1:1 propensity score matching (39 pairs, n = 78), followed by logistic regression in the matched cohort. Results. Of 143 participants, 104 (73%) received intravenous penicillin G and 39 (27%) ceftriaxone. In the matched cohort, ESR occurred in 74.4% of penicillin-treated participants and 71.8% of ceftriaxone-treated participants (odds ratio 0.88, 95% confidence interval .32–2.40; P = .80). The results were consistent across all prespecified sensitivity analyses. Conclusions. In this multicenter cohort of people with HIV diagnosed with neurosyphilis, ocular syphilis, or otosyphilis, ceftriaxone showed comparable ESR to intravenous penicillin G. These findings support ceftriaxone as a reasonable alternative when standard therapy is limited by penicillin shortages, limited inpatient capacity, or financial constraints. Received 09 March 2026; accepted 18 June 2026; published online 13 July 2026 Correspondence: Brenda Crabtree-Ramírez, MD, Departamento de Infectología, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Vasco de Quiroga No. 15, Col. Belisario Domínguez Sección XVI, Alcaldía Tlalpan, Mexico City 14080, Mexico (). Open Forum Infectious Diseases®© The Author(s) 2026. Published by Oxford University Press on behalf of Infectious Diseases Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact reprints@oup.-com for reprints and translation rights for reprints. All other permissions can be obtained through our RightsLink service via the Permissions link on the article page on our site—for fur-ther information please contact journals.permissions@oup.com. Open Forum Infectious DiseasesM A J O R A R T I C L E ### Page 2 Graphical abstract ![For image description, please refer to the figure legend and surrounding text.]() Keywords. ceftriaxone; HIV; limited-resource settings; neurosyphilis; penicillin. Syphilis is a sexually transmitted infection caused by the spiro-chete Treponema pallidum, an obligate human pathogen that disseminates hematogenously soon after infection and pro-gresses through primary, secondary, latent, and tertiary stages, with diverse mucocutaneous, systemic, ophthalmologic, oto-logic, and neurological manifestations. The central nervous sys-tem (CNS) can be invaded at any stage; however, the organism is frequently cleared spontaneously, and CNS invasion alone does not constitute neurosyphilis, which requires compatible clinical features, supported by serological and, when available, cerebrospinal fluid (CSF) findings. Neurosyphilis encom-passes a spectrum of CNS manifestations associated with sub-stantial morbidity, including irreversible neurological and sensory deficits [,]. The global burden of syphilis has risen substantially, with annual incident cases increasing from 7.1 million in 2020 to 8.0 million in 2022 []. Treponema pallidum CNS invasion has been documented in up to 30% of individuals with syphilis, while clinical neurosyphilis ac-counts for ∼4%–10% of all reported cases [,]. People with HIV-1 (PWH) are disproportionately affected, with higher rates of neurologic, ophthalmic, and otologic involve-ment, and an increased risk of treatment failure, recurrence, and reinfection [,]. International guidelines group neurosyphilis, ocular syphilis, and otosyphilis together as conditions sharing a common treatment approach, recommending high-dose intravenous pen-icillin G as the standard therapy for all three []. However, this regimen requires prolonged inpatient care or continuous intrave-nous access, posing logistical, and financial barriers for patients and health systems []. Recurrent aqueous penicillin short-ages, limited inpatient capacity, and high out-of-pocket costs fur-ther constrain access to standard neurosyphilis treatment, particularly in resource-limited settings [,]. Ceftriaxone has emerged as a potential alternative owing to its prolonged half-life, adequate treponemicidal CSF concen-trations, and once-daily dosing, which may facilitate outpatient administration []. Observational studies suggest compa-rable outcomes between ceftriaxone and intravenous penicillin; however, most were conducted in high-resource settings, in-cluded a limited number of PWH, focused on uveitis or lacked adjustment for confounding [,,]. Robust compara-tive evidence focusing specifically on PWH remains scarce, par-ticularly in resource-limited, real-world settings []. To address this gap, we conducted a multicenter retrospec-tive cohort study of PWH diagnosed with neurosyphilis, ocu-lar syphilis, or otosyphilis, at public hospitals and outpatient HIV clinics in Mexico. Using propensity score matching (PSM), we compared 6-month early serological response (ESR) between PWH treated with ceftriaxone or intravenous penicillin G. ### Page 3 ## METHODS We conducted a multicenter, retrospective cohort study of adults with HIV treated for neurosyphilis, ocular syphilis, or otosyphilis at 5 public hospitals and 2 outpatient HIV clinics in 3 cities: Mexico City, Guadalajara, and Oaxaca, between January 2015 and January 2024. ### Study Population and Inclusion Criteria We included adults (≥18 years) with HIV with confirmed syph-ilis, defined by both a baseline serum nontreponemal titer (rap-id plasma reagin [RPR] or Venereal Disease Research Laboratory [VDRL]) ≥1:4 and a positive Fluorescent Treponemal Antibody Absorption test, together with compat-ible neuropsychiatric, ophthalmologic, or otologic manifesta-tions confirmed by specialist evaluation (ophthalmology, neurology, psychiatry, or otolaryngology, as clinically indicat-ed), with or without CSF evaluation. When CSF data were available, confirmation could also be supported by CSF abnor-malities, including a reactive nontreponemal test, pleocytosis (CSF white cell count ≥5 cells/μL), or CSF protein ≥45 mg/ dL. In the absence of national Mexican criteria for these con-ditions, diagnostic definitions followed CDC guidelines []. Participants received intravenous penicillin G (18–24 mil-lion international units/day for 10–14 days) or ceftriaxone (1–2 g/day intravenously or intramuscularly for 10–14 days) according to clinician judgment. The ceftriaxone dose range re-flects variation in prescribing practice across participating cen-ters during the study period, consistent with published recommendations []. We excluded individuals with alterna-tive diagnoses or incomplete baseline or 6-month follow-up data. ### Data Collection Between January and June 2024, we obtained data from electron-ic and paper medical records and entered them into a secure REDCap database. We extracted anonymized demographic, clinical, HIV-related, laboratory, and CSF variables using a predefined protocol. We extracted only objectively documented variables because provider-rated disease severity was not consis-tently documented across centers. ### Outcomes The primary outcome was ESR, defined as a ≥4-fold decline (or seroreversion) in serum RPR/VDRL titer at 6 months compared with baseline. As a secondary outcome, we evaluated early clinical response (ECR), defined as the resolution of neu-rological, ophthalmologic, or otologic manifestations among symptomatic individuals at baseline. Participants without base-line symptoms were classified as not evaluable for this outcome, rather than as missing. Cerebrospinal fluid normalization was not included as an outcome due to inconsistent follow-up lum-bar puncture availability. ### Statistical Analysis We described baseline characteristics using χ2 or Fisher’s exact test for categorical variables and Student’s t test or the Wilcoxon rank-sum test for continuous variables, as appropri-ate. To address confounding by indication, we estimated pro-pensity scores for receipt of ceftriaxone or intravenous penicillin using a logistic regression model incorporating clin-ically relevant covariates selected a priori based on biological plausibility: age, neuropsychiatric, ophthalmologic, or otologic manifestations, hepatitis B coinfection, hepatitis C coinfection, maximum baseline nontreponemal titer, HIV viral load, and advanced HIV disease (defined as <200 CD4 cells/mm3). Cerebrospinal fluid parameters were not included in the pro-pensity score model, because lumbar puncture was not routine-ly performed at the outpatient HIV clinics where most ceftriaxone-treated participants received care. We then per-formed 1:1 nearest neighbor, without a caliper, targeting the av-erage treatment effect in the treated, yielding 39 pairs (N = 78). We evaluated covariate balance using standardized mean dif-ferences (SMD <0.1 considered adequate). The robustness of the matching specification (0.2-SD caliper) and of residual im-balance (within-matched-cohort regression adjustment) was examined in additional sensitivity analyses (). The primary analysis estimated the odds ratio (OR) for ESR with ceftriaxone versus intravenous penicillin G using logistic regression in the matched cohort, with treatment as the sole predictor. We conducted 3 prespecified sensitivity analyses: (1) matched model adjusted for uncontrolled HIV-1 viremia and advanced HIV disease; (2) matched model adjusted for baseline maximum serum nontreponemal titer; and (3) a fully adjusted multivariable logistic regression model in the full co-hort. We compared model fit using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). In additional analyses, we repeated the comparison using a 12-month serological response (identical definitions), under a neurosyphilis definition in PWH of CSF white cell count ≥20 cells/μL [] or reactive CSF VDRL, and under varying matching specifications (,). We handled missing data by complete-case analysis; 6-month titers were those recorded at the scheduled 6-month follow-up visit, and we assessed attri-tion by comparing participants with and without paired 6-month titers according to treatment, baseline symptoms, age, CD4 count, and HIV-1 RNA. The study was powered to detect a ≥15% absolute difference in ESR (2-sided α = .05, 80% power); however, given the retrospective design, this cal-culation was post hoc and is reported as a characterization of detectable effect size only. Treatment effects are presented as ORs with 95% confidence intervals (CIs) rather than as a formal noninferiority analysis, because no noninferiority margin was prespecified and the matched cohort was underpowered for ### Page 4 such testing. Accordingly, all comparisons are interpreted de-scriptively. All analyses were conducted in R version 4.5.2. ### Patient Consent Statement All institutional review boards approved the protocol and waived informed consent given the retrospective design. The study followed the Declaration of Helsinki standards. ## RESULTS ### Study Population Of 340 individuals assessed for eligibility, 143 met the inclusion criteria and had complete 6-month serological follow-up (). Of these, 104 (73%) received intravenous penicillin G and 39 (27%) received ceftriaxone. Propensity score match-ing yielded 39 matched pairs (n = 78); 65 penicillin-treated par-ticipants remained unmatched. Baseline characteristics differed between groups on presentation, including the distribution of neurological and ophthalmologic manifestations (). ### Attrition Of 242 eligible participants, 99 (10%) lacked paired 6-month nontreponemal titers and were not included in the primary anal-ysis. The proportion without paired titers was similar between treatment groups (penicillin 41% [72/176] vs ceftriaxone 41% [27/66]; P = 1.0), indicating nondifferential loss to follow-up by treatment. Participants without paired titers were more likely to have baseline symptoms (79% vs 59%; P = .002) but did not differ in age, CD4 count, or HIV-1 RNA, consistent with differ-ences in care pathways and follow-up capture (symptomatic or inpatient care vs routine outpatient HIV follow-up) rather than treatment-related effects. Among those without paired ti-ters, 32 had a documented follow-up visit but no recorded paired titer, indicating that part of the attrition reflected incomplete lab-oratory documentation rather than true loss to follow-up. ### Determinants of Treatment Allocation Before matching, participants treated with ceftriaxone had higher baseline nontreponemal titers and more frequent CSF ![Flow diagram titled “Figure1. Study patient flow diagram.” A box at the top left states: “Patients with syphilis assessed for eligibility \(n = 340\).” An arrow points right to a box labeled “Excluded,” followed by a box detailing: “Did not meet inclusion criteria \(n = 98\),” with bullet points: No neurosyphilis \(n = 60\); No HIV \(n = 11\); Other treatments for neurosyphilis \(n = 27\). A downward arrow from the initial box leads to “Patients included in the study \(n = 242\).” From this box, an arrow points right to “Excluded from analysis \(n = 99\),” with one bullet point “No paired 6-month serological follow-up \(n = 99\).”]() ### Page 5 abnormalities, whereas ocular and neuropsychiatric involve-ment were more common among those receiving intravenous penicillin G (). Propensity score distributions showed adequate overlap be-tween treatment groups in the unmatched cohort, supporting the use of matching (). Following PSM, baseline covariates achieved acceptable balance, with abso-lute SMD values below the recommended 0.1 (). ### Primary Outcome: Early Serological Response In the primary propensity score–matched model, ESR occurred in 74.4% of penicillin-treated participants and 71.8% of ceftriaxone- treated participants, OR 0.88 (95% CI .32–2.40; P = .80;). ### Sensitivity Analysis Results were consistent across all prespecified sensitivity analy-ses; all point estimates were directionally similar to the primary analysis and no comparison reached statistical significance (P > .05;). The matched model adjusted for HIV dis-ease markers showed a directionally similar estimate; the likelihood-ratio test indicated improved model deviance rela-tive to the primary matched model (P = .049) though this did not alter the treatment effect. The matched model incorporat-ing baseline nontreponemal titer demonstrated the best fit among matched models (AIC: 82.1; BIC: 89.2) and produced estimates consistent with the primary analysis. The fully adjusted unmatched multivariable model showed substantially higher information criteria values (AIC: 162.5; BIC: 198.0), re-flecting the larger dataset rather than poorer model quality, and yielded a consistent treatment effect estimate (OR 1.06, 95% CI .39–2.97). Covariate-specific estimates from the unmatched multivariable model are shown in. Missing paired 6-month VDRL titers were similar between treatment groups: 67/171 (39%) in the penicillin group and 26/65 (40%) in the cef-triaxone group (P = 1.0). Participants without paired titers were more likely to have baseline symptoms than those with paired titers (80% vs 59%; P = .002), but did not differ by age, CD4 count, or HIV RNA. In a sensitivity analysis applying a 12-month response definition to a partially overlapping cohort of 85 participants, ESR occurred in 82.5% (47/57) of penicillin- treated participants and 75.0% (21/28) of ceftriaxone-treated participants before matching (OR 0.64, 95% CI .19–2.28;P = .57). After 1:1 PSM (28 pairs), response rates were identical between arms (21/28 [75%] each; OR 1.00, 95% CI .29–3.41;P = 1.0), consistent with the 6-month primary analysis (). In a sensitivity analysis restricted to participants meeting a stricter CSF definition of neurosyphilis (CSF white cell count ≥20 cells/μL or reactive CSF VDRL; n= 78), propensity-score matching yielded 34 pairs; ESR oc-curred in 25/34 (73.5%) penicillin-treated participants and 23/34 (67.6%) ceftriaxone-treated participants (OR 0.75, 95% CI .26–2.14; P = .60), consistent with the primary analysis (OR 0.88;). ### Page 6 ![Bar chart titled “Early serological response at 6 m by treatment group.” The y-axis shows proportion \(95% confidence interval \[CI\]\) from 0% to 100%, and the x-axis lists two treatment groups: penicillin and ceftriaxone. The penicillin group has a response rate of 74.4% with a 95% CI of 58.9%-85.4%. The ceftriaxone group has a response rate of 71.8% with a 95% CI of 56.2%-83.5%. Error bars indicate the 95% CIs for each group. At the top of the figure, the comparison reports an odds ratio \(OR\) of 0.88 \(95% CI .32-2.40\) with P = .80, indicating no statistically significant difference between treatments.]() ![Forest plot titled “Effect of ceftriaxone versus penicillin on early serological response,” with subtitle “Primary and sensitivity analyses.” The x-axis shows odds ratios on a logarithmic scale ranging approximately from 0.10 to 8.00, with a vertical dashed reference line at 1.0 \(no effect\). Four analyses are displayed as horizontal lines with central point estimates and 95= confidence intervals: Primary: PS-matched-OR 0.88 \(95= CI .32-2.40\); Sensitivity 1: matched + HIV variables-OR 0.73 \(95= CI .24-2.13\); Sensitivity 2: matched + max RPR/VDRL titer-OR 0.92 \(95= CI .31-2.77\); Sensitivity 3: full multivariable \(unmatched\)-OR 1.06 \(95= CI .39-2.97\). All confidence intervals cross 1.0, indicating no statistically significant difference between ceftriaxone and penicillin in any analysis.]() ### Page 7 ### Secondary Outcome: Early Clinical Response Early clinical response applied only to symptomatic partici-pants at baseline (85/143). Six-month symptom status was doc-umented in 74/75 penicillin-treated participants and 10/10 ceftriaxone-treated symptomatic participants. The remaining 58 participants were asymptomatic at baseline and were there-fore not candidates for this assessment, rather than lost to follow-up. Among evaluable participants (n = 84), ECR oc-curred in 55/74 (74.3%) of penicillin-treated participants and 9/10 (90.0%) of ceftriaxone-treated participants (OR 3.08, 95% CI .38–143.25; Fisher’s exact P = .44). At 6 months, CSF follow-up was available for 25/143 partic-ipants (17%), and only 11 had sufficient data for complete bio-chemical response assessment. Among these 11 participants, agreement between serological and CSF response was limited (5/11, 45%); discordance was mainly driven by 6 participants who met serological response criteria but did not achieve CSF normalization based on pleocytosis, protein concentra-tion, and/or CSF serology. ## DISCUSSION In this multicenter retrospective cohort of PWH diagnosed with neurosyphilis, ocular syphilis, or otosyphilis, we found compara-ble rates of ESR at 6 months between participants treated with ceftriaxone and those treated with intravenous penicillin G. After PSM treatment effect estimates remained stable across sensitivity analyses, with no clinically meaningful differences in ESR and ECR. These findings align with previous observation-al studies suggesting comparable effectiveness of ceftriaxone and intravenous penicillin G for neurosyphilis []. However, much of the existing literature has been limited by small sample sizes, heterogeneous populations, combined outcome definitions, or insufficient adjustment for confounding [,,,]. By focusing exclusively on PWH and applying a matched design, our study strengthens the evidence base for cef-triaxone in a population characterized by higher neurologic in-volvement, reinfection, and substantial barriers to care. Our cohort also carried a substantial burden of advanced HIV dis-ease, ∼40% with CD4 < 200 cells/mm3 and about half with HIV-1 RNA >1000 copies/mL, reflecting late presentation com-mon in Latin America. This contrasts with the comparative ceftriaxone-versus-penicillin literature, conducted predomi-nantly in higher-resource settings with at most moderate immu-nosuppression and generally not assessing immunosuppression as an effect modifier [,]. Reassuringly, our estimate was un-changed after adjustment for uncontrolled viremia and ad-vanced HIV disease (OR 0.73, 95% CI .24–2.13), suggesting comparable ESR even in more immunocompromised individu-als, a group underrepresented in prior studies []. ![Forest plot titled “Multivariable analysis of serological response.” The x-axis shows odds ratios on a logarithmic scale from approximately 0 to 10, with a vertical dashed reference line at 1.0 \(no effect\). Each row displays a covariate with a point estimate and horizontal 95% confidence interval. Variables and results: Ceftriaxone \(vs penicillin\): OR 1.06 \(95% CI .39-2.97\); Ocular symptoms: OR 1.07 \(95% CI .43-2.65\); Neurologic symptoms: OR 0.98 \(95% CI .29-3.50\); Otologic symptoms: OR 0.54 \(95% CI .09-3.03\); Max nontreponemal titer \(baseline\): OR 1.01 \(95% CI 1.00-1.01\); HIV uncontrolled: OR 2.09 \(95% CI .68-7.06\); Advanced HIV disease: OR 0.41 \(95% CI .12-1.27\); Age \(per year\): OR 1.02 \(95% CI .98-1.06\). All confidence intervals cross 1.0 except the baseline nontreponemal titer, which is centered very close to 1.0, indicating no statistically significant associations for most variables.]() ### Page 8 Intravenous penicillin often requires hospitalization or prolonged intravenous access, whereas ceftriaxone’s pro-longed half-life and once-daily dosing facilitates outpatient administration. In our cohort, 95% of penicillin-treated par-ticipants required hospitalization, compared with 23% of those receiving ceftriaxone, illustrating a practical advantage of outpatient ceftriaxone administration where inpatient ca-pacity is limited. Additionally, ceftriaxone is well tolerated, can be given intramuscularly, and offers an option for penicillin-allergic patients []. Intramuscular procaine penicillin G with oral probenecid (a guideline-recommended regimen permitting outpatient peni-cillin administration) has been discontinued or is recurrently unavailable in many settings, removing one of the few nonhos-pitalized penicillin options and further supporting ceftriaxone as a practical alternative. Oral regimens including doxycycline, amoxicillin–probenecid, and linezolid are under investigation but remain outside current guideline recommendations pend-ing further evidence [,]. Finally, the recent identifica-tion of mutations associated with partial β-lactam resistance and macrolides in contemporary T. pallidum strains further supports interest in alternative treatment options [,], though clinical relevance of these findings requires prospective validation. The comparable ESR observed with ceftriaxone and intrave-nous penicillin G in this cohort, together with ceftriaxone’s practical advantages, is clinically relevant. Furthermore, higher baseline nontreponemal titers were weakly associated with ESR in the multivariable model (OR 1.006 per unit increase, 95% CI 1.002–1.012), suggesting that serological burden at diagnosis may modestly influence treatment response; these findings warrant confirmation in larger prospective studies. The ob-served imbalances in ocular disease and CSF abnormalities re-flect treatment selection by clinical phenotype and care setting rather than differential ascertainment, as ophthalmologic and CSF evaluation were performed in nearly all participants in both groups, and CSF abnormality rates were similar across phenotypes among those examined. Ocular syphilis was prefer-entially treated with inpatient intravenous penicillin G, where-as nonocular, CSF-defined neurosyphilis was more often managed with outpatient ceftriaxone. Although PSM balanced clinical presentation, residual confounding from unmeasured differences in clinical decision-making, including clinician- perceived disease severity and referral patterns, cannot be excluded. This study has several limitations. First, unmeasured con-founding cannot be excluded despite successful matching; nota-bly, CSF parameters, which differed substantially between groups, could not be incorporated into the propensity score model due to structural absence of lumbar puncture data at out-patient centers. Second, ECR analysis was limited by substantial missing follow-up data, precluding reliable comparisons; the small number of symptomatic ceftriaxone-treated participants yielded an imprecise estimate, rendering this exploratory com-parison hypothesis-generating only. Third, we were unable to distinguish treatment failure from reinfection in a population with continuous exposure risk. Fourth, participating centers were predominantly urban and specialized, limiting generaliz-ability to primary care or rural settings. Fifth, among the small subset with 6-month CSF follow-up, serological response ap-peared to overestimate improvement relative to CSF markers; however, the limited availability of repeat CSF assessments pre-cluded its use for comparative effectiveness conclusions. Sixth, the retrospective design also precluded reliable syndromic stag-ing of neurosyphilis; participants were grouped by affected sys-tem (neurological/psychiatric, ophthalmologic, or otologic) rather than by disease stage. Importantly, our study was not de-signed as a noninferiority study: no margin was prespecified, and the matched cohort was underpowered for formal noninferiority testing. Accordingly, although ESR was comparable between groups, our findings are consistent with, but cannot formally es-tablish, noninferiority of ceftriaxone. Finally, only early out-comes were assessed; longer-term standardized serological and clinical outcomes warrant further study. Despite these limitations, ceftriaxone represents a reasonable alternative to intravenous penicillin G for PWH with neurosy-philis when standard therapy is not feasible. Future prospective studies incorporating patient-centered outcomes, standardized clinical monitoring and CSF assessments with extended follow- up are needed to define the optimal role of alternative regimens across diverse care settings and populations. ## CONCLUSIONS In this multicenter cohort of PWH diagnosed with neurosyphi-lis, ocular syphilis, or otosyphilis, ceftriaxone showed compara-ble ESR to intravenous penicillin G. These findings support ceftriaxone as a reasonable alternative when standard therapy is not feasible due to penicillin shortages, limited inpatient ca-pacity, or the financial and logistical constraints of prolonged intravenous administration, particularly in resource-limited settings. ## Supplementary Data Supplementary materials are available atonline. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corre-sponding author. ## Notes Acknowledgments. The authors thank Dr Yanink Caro Vega for her re-view of the statistical methodology and results. Author Contributions. R. V. B.: conceptualization, project administra-tion, investigation, methodology, formal analysis, visualization, validation, writing—original draft, review, and editing. H. R.-V.: project administration, data curation, investigation, methodology, validation, writing—original ### Page 9 draft, review, and editing. O. O.-V.: data curation, investigation, visualiza-tion, software, writing—review and editing. R. F.-C.: project administration, investigation, writing—review and editing. L. A. G.-H.: methodology, inves-tigation, validation, writing—review and editing. J. F. A.-V.: writing—review & editing. P. M.-A.: investigation, methodology. A. C.-O.: investigation, writ-ing—original manuscript, review, and editing. P. C.-P.: investigation, writing —original draft, review, and editing. J. L.-M.: investigation, writing—original draft, review, and editing. J. C. R.-A.: investigation, writing—original draft, review, and editing. E. P.-B.: investigation, writing—original draft, review, and editing. P. A.-F.: investigation, writing—original draft, review, and editing. E. S.-I.: investigation, writing—review and editing. N. E. R.-M.: in-vestigation, writing—review and editing. B. C.-R.: conceptualization, investi-gation, methodology, validation, writing—original draft, review, and editing, supervision, resources, funding acquisition. Data availability. The data underlying this article will be shared on rea-sonable request to the corresponding author. Use of artificial intelligence. All manuscript contents, including the study design, data analysis, interpretation of results, and scientific conclu-sions, were created entirely by the authors. The authors acknowledge the use of Anthropic’s Claude () to assist with grammar correc-tion, language editing and translation, and final preparation for submis-sion. Claude was also used to refine R code for figure improvement. All artificial intelligence–assisted content was reviewed and approved by the authors, who take full responsibility for the accuracy and integrity of the work. Financial support. No funding was received for this work. Potential conflicts of interest. All authors: No reported conflicts. All au-thors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. ## References Hamill MM, Ghanem KG, Tuddenham S. State-of-the-art review: neurosyphilis. Clin Infect Dis 2024; 78:e57–68. Wu S, Ye F, Wang Y, Li D. Neurosyphilis: insights into its pathogenesis, suscept-ibility, diagnosis, treatment, and prevention. Front Neurol 2023; 14:1340321. World Health Organization. Implementing the global health sector strategies on HIV, viral hepatitis and sexually transmitted infections, 2022–2030. Geneva;2024. Available at:. Wu Y, Lu L, Song X, et al. Clinical and immunological characteristics of HIV/ syphilis co-infected patients following long-term antiretroviral treatment. Front Public Health 2024; 11:1327896. Shorer EF, Zauchenberger CZ, Govender S, et al. Neurological manifestations of syphilis-HIV coinfection in South Africa. J Neurol Sci 2023; 455: 122798. Centro Nacional para la Prevención y Control del VIH y el Sida. Guía de Manejo Antirretroviral de las Personas con VIH 2025. Gobierno de México. Available at: Accessed 15 Jan 2026. Janier M, Unemo M, Dupin N, Tiplica GS, Poto M, Patel R. 2020 European guide-line on the management of syphilis Conflicts of interest. 2020. Available at:. Accessed 15 Jan 2026. Kingston M, Apea V, Evans C, et al. BASHH UK guidelines for the management of syphilis 2024. Int J STD AIDS 2024; 35:1142–60. Centers for Disease Control and Prevention. Neurosyphilis, Ocular Syphilis, and Otosyphilis—STI Treatment Guidelines. Sexually Transmitted Infections Treatment Guidelines, 2021. Available at: Accessed 15 Jan 2026. Villarreal DD, Babalola CM. Expanding horizons in syphilis treatment: challeng-es, advances, and opportunities for alternative antibiotics. Curr HIV/AIDS Rep2025; 22:22. Callado GY, Pardo I, Gutfreund MC, et al. Insights into treatment alternatives for neurosyphilis: systematic literature review and meta-analysis. Sex Transm Dis2024; 51:641–7. Buitrago-Garcia D, Martí-Carvajal AJ, Jimenez A, Conterno LO, Pardo R. Antibiotic therapy for adults with neurosyphilis. Cochrane Database Syst Rev2019; 5:CD011399. Liu HY, Han Y, Chen XS, et al. Comparison of efficacy of treatments for early syphilis: a systematic review and network meta-analysis of randomized controlled trials and observational studies. PLoS One 2017; 12:e0180001. Rodriguez-Aguilar R, Marmolejo-Saucedo JA, Zavala-Landin A, Rodriguez- Aguilar M, Marmolejo-Saucedo L. Out of pocket and catastrophic health spend-ing in Mexico in the face of the COVID-19 pandemic. EAI Endorsed Trans Pervasive Health Technol 2023; 9:e3583. Lazzarin S, Giacomelli A, Moschese D, Antinori S, Gervasoni C. Drug shortages highlight the need for expanded treatment options for syphilis. Clin Infect Dis2023; 77:1356–. Shann S, Wilson J. Treatment of neurosyphilis with ceftriaxone. Sex Transm Infect 2003; 79:415–6. Bettuzzi T, Jourdes A, Robineau O, et al. Ceftriaxone compared with benzylpeni-cillin in the treatment of neurosyphilis in France: a retrospective multicentre study. Lancet Infect Dis 2021; 21:1441–7. Jourdes A, Bettuzzi T, Robineau O, et al. La ceftriaxone est non-inférieure à la pénicilline G dans le traitement de la neurosyphilis. Med Mal Infect 2020; 50: S133–4. Wang S, Gu W, Cao Y, et al. Comparison of the clinical efficacy of penicillin and ceftriaxone sodium in the treatment of neurosyphilis with psychiatric symptoms HHS public access. World J Psychiatry Ment Health Res 2022; 5:1030. Agostini FA, Queiroz RP, Azevedo DOM, Henriques JF, Campos WR, Vasconcelos-Santos DV. Intravenous ceftriaxone for syphilitic uveitis. Ocul Immunol Inflamm 2018; 26:1059–65. Marra CM, Boutin P, McArthur JC, 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. Gu X, Lu H, Yang Y, et al. Could ceftriaxone be a viable alternative to penicillin for the treatment of ocular syphilis? Antimicrob Agents Chemother 2024; 68: e0008024. Gomes e Silva A, Alcântara Pedroso L, Veiga Costa F, et al. Intravenous ceftriax-one on an outpatient basis in the treatment of ocular syphilis. Ocul Immunol Inflamm 2024; 32:1284–6. Xu N, Yuan JG, Dai QJ, et al. Syphilitic uveitis in HIV-positive patients: report of a case series, treatment outcomes, and comprehensive review of the literature. Int J Ophthalmol 2023; 16:1250–9. Marra CM, Maxwell CL, Smith SL, et al. Cerebrospinal fluid abnormalities in pa-tients with syphilis: association with clinical and laboratory features. J Infect Dis2004; 189:369–76. Katanami Y, Hashimoto T, Takaya S, et al. Amoxicillin and ceftriaxone as treat-ment alternatives to penicillin for maternal syphilis. Emerg Infect Dis 2017; 23: 827–9. Chapman E, Sued O, Barreto JOM, et al. Research priorities for advanced HIV disease in Latin America and the Caribbean region: a modified Delphi study. J Int AIDS Soc 2026; 29:e70074. Villani P, Regazzi MB, Marubbi F, et al. Cerebrospinal fluid linezolid concentra-tions in postneurosurgical central nervous system infections. Antimicrob Agents Chemother 2002; 46:936–7. Nau R, Sörgel F, Eiffert H. Penetration of drugs through the blood-cerebrospinal fluid/blood-brain barrier for treatment of central nervous system infections. Clin Microbiol Rev 2010; 23:858–83. Bao YK, Hwang J, Long C, et al. Outpatient oral doxycycline therapy for ocular syphilis. JAMA Netw Open 2024; 7:e2449364. O’Mahony C, Bradley MG, McKeown J, Arya OP. Treponemicidal levels of amox-icillin can be achieved in cerebrospinal fluid following oral treatment with only 4 g amoxicillin and 2 g probenecid daily in late stage syphilis. Int J STD AIDS 2012; 23:758. Haynes AM, Giacani L, Mayans MV, et al. Efficacy of linezolid on Treponema pal-lidum, the syphilis agent: a preclinical study. EBioMedicine 2021; 65:103281. Tramont EC, Morrison RE, Harrison SM. Oral amoxycillin, an alternative treat-ment for neurosyphilis. Sex Transm Infect 1985; 61:359–62. Ubals M, Nadal-Baron P, Arando M, et al. Oral linezolid compared with benza-thine penicillin G for treatment of early syphilis in adults (Trep-AB study) in Spain: a prospective, open-label, non-inferiority, randomised controlled trial. Lancet Infect Dis 2024; 24:404–16. Girometti N, Junejo MH, Nugent D, et al. Clinical and serological outcomes in patients treated with oral doxycycline for early neurosyphilis. J Antimicrob Chemother 2021; 76:1916–9. Beale MA, Marks M, Sahi SK, et al. Genomic epidemiology of syphilis reveals in-dependent emergence of macrolide resistance across multiple circulating lineages. Nat Commun 2019; 10:3255. Stamm LV. Global challenge of antibiotic-resistant Treponema pallidum. Antimicrob Agents Chemother 2010; 54:583–9.