Polymerase Chain Reaction Testing of Vitreous in Atypical Ocular Syphilis

Abstract

Purpose: To illustrate the diagnostic importance of syphilis PCR testing on intraocular fluid for atypical presentations.

Methods: Retrospective case series.

Results: Two cases of atypical intraocular inflammation were confirmed by PCR testing of vitreous for Treponema pallidum DNA.

Conclusions: Ocular syphilis may be present in atypical fashion, and delayed treatment may lead to irreversible visual loss. Sampling of intraocular fluid for PCR testing may confirm diagnosis and lead to appropriate treatment.

• Atypical
• ocular
• PCR
• syphilis
• uveitis

Syphilis is a sexually transmitted disease caused by the spirochaete Treponema pallidum. Since the late 1990s the incidence of the disease in the United Kingdom has increased substantially, and in 2007 there were 3762 reported new cases, the highest figure since 1950. The majority of this increase has been seen in urban centers, and the effect of this increase on new cases of uveitis in Manchester has been reported.1 A high proportion of cases is seen in men who have sex with men, and HIV co-infection is common. Ocular syphilis often presents with typical features that facilitate diagnosis, but in atypical cases, treatment (which is curative) may be delayed and the visual outcome therefore suboptimal. We present two patients with misleading clinical features, in whom the diagnosis was confirmed with PCR of the vitreous.

Patients

Case 1

A 61-year-old man presented with a 1-week history of floaters in his right eye. He had recently undergone surgery and chemotherapy for esophageal cancer, the last cycle being administered 3 months previously. He described a diffuse chest rash, which his general practitioner was treating with a topical antifungal cream, a productive cough treated with oral antibiotics, and recent lip and gum ulcers. He had no history of or known risk factors for sexually transmitted disease.

On examination his vision was right 0.4 and left −0.18 logMAR. The left eye was normal but the right showed mild nongranulomatous anterior uveitis and substantial vitritis with cell clumping (Figure 1). There was no evidence of retinitis or vasculitis. Fungal endophthalmitis was suspected because of the quality of vitreous infiltrate and in view of his recent history of repeated venous access. He underwent vitrectomy with intravitreal injection of amphotericin, and vitreous specimens were sent for bacterial and fungal culture, and PCR for Candida and Aspergillus. All of these investigations proved to be negative, as did fungal PCR on peripheral blood. Meanwhile, hematological investigations showed normochromic anemia (hemoglobin 10.3 mg/L and red cell count 3.41 × 10¹²/L), mild neutropenia (1.94 × 10⁹/L), and elevated inflammatory markers (ESR 96 mm/1 h, C-reactive protein 20 mg/L). Syphilis serology had also been performed and showed positive treponemal IgG ELISA, positive IgM ELISA, positive Treponema pallidum particle agglutination (TPPA) at 1:>5120 (all strongly suggesting past or present exposure to syphilis), and a positive Rapid Plasma Reagent test at a titer of 1:64 (strongly suggesting active syphilis). The vitreous sample was therefore reexamined by syphilis PCR testing, which also proved to be positive. Lumbar puncture was then performed, revealing positive TPPA at 1:160. He was HIV-negative. He was therefore treated for ocular and neurosyphilis, with the standard UK regime of 17 days of intramuscular procaine penicillin, with probenecid. His rash resolved rapidly and his uveitis resolved. Vitrectomy had permitted clearer funduscopy, and a small area of RPE atrophy, probably indicating settled retinitis, was noted in the right superior fundal periphery. At his last visit his eye was uninflamed with acuity of 0.1 logMAR and post-vitrectomy nuclear sclerosis.

FIGURE 1. (Case 1) Vitreous infiltrate at presentation showing strings of opacities.

Case 2

A 37-year-old HIV-positive male was diagnosed with Mycobacterium avium complex (MAC) pneumonitis and treated with rifabutin and clarithromycin, in addition to highly active anti-retroviral treatment. He later presented with a 1-month history of painless floaters and blurred vision in his left eye. On examination, his vision was right 0.0 and left 0.1 logMAR. Both eyes were white, but the left exhibited substantial (4+) cellular infiltrate, a small (<1 mm) uncoagulated hypopyon, and significant vitritis. No focal retinal or choroidal lesions were seen. It was considered most likely that this unilateral nongranulomatous panuveitis represented rifabutin toxicity. However, in order to exclude viral causes he underwent aqueous sampling for herpesviral PCR, all of which (HSV-1, HSV-2, VZV, CMV) were negative. The rifabutin was discontinued and he was commenced on frequent topical steroids and oral prednisolone 30 mg/day.

Symptoms initially improved, but after 4 weeks he complained again of increased floaters and decreased vision to 0.4 logMAR. On this occasion, small white lesions were seen in the nasal retina and fluorescein angiography revealed significant arteriolar closure (Figure 2a). He was further investigated and he proved to have positive syphilis serology consistent only with past infection, with an RPR titer of 1:4. He underwent further aqueous sampling for both syphilis and herpesviral PCR. The former was negative, but he was PCR-positive for HSV-1. He was therefore admitted and treated with intravenous aciclovir 10 mg/kg TDS for 10 days and discharged on oral aciclovir 800 mg twice daily. His vision improved and the retinal signs improved.

FIGURE 2. (Case 2) Top - White dots overlying widespread ischaemic retinitis. Bottom - A new area of retinitis with overlying white dots strongly suggesting syphilitic retinitis.

Four weeks later, his vision had again worsened to 0.6 logMAR. He had clearly developed a new area of retinitis temporally (Figure 2b) with features suggesting syphilitic inflammation. Again aqueous sampling was performed and again both herpesviral and syphilis PCR were negative. He therefore underwent vitrectomy, specimens being sent for bacterial (including mycobacterial) and fungal culture, and further PCR, which on this occasion was positive for syphilis. Being allergic to penicillin he was treated with intravenous ceftriaxone 2 g/day for 2 weeks. He did not undergo lumbar puncture as he was already being treated as for neurosyphilis. His uveitis rapidly responded to treatment. At his last visit, his retinitis and occlusive vasculitis had resolved, leaving him with 0.1 logMAR acuity but with a severely constricted visual field because of retinal infarction.

Discussion

Ocular syphilis can manifest in protean fashion, including episcleritis, scleritis, keratitis, and optic neuritis. However, the great majority of patients develop uveitis in one or both eyes. This may be limited to the anterior segment (in which case it may be granulomatous or nongranulomatous) but the infection more commonly causes posterior uveitis. There may be retinal vasculitis, which can involve arterioles and which may be occlusive. The most usual expression, however, is of widespread retinitis, often associated with punctate lesions either as adjacent “satellite” lesions or overlying the zonal inflammation. Most patients present during the early stages of secondary syphilis and often have persistent global headache with or without a maculopapular skin rash of variable extent. Typical ocular appearances in combination with suggestive systemic signs and syphilis serology indicating active disease is adequate to presume intraocular syphilis and to treat appropriately. However, neither of our patients initially exhibited suggestive features and in both cases the context and the appearance suggested a different diagnosis. It is in such circumstances that PCR testing of intraocular fluid assumes significance.

The usefulness of real-time PCR in the diagnosis of nonocular syphilis is well established in a genitourinary medicine clinic,2 with a claimed sensitivity of 73%, a specificity of 95%, and positive predictive value of 89%. A recent metanalysis3 of 69 studies examining sampling results from multiple sites found the highest sensitivity (83%) in blood from infected neonates, with pooled specificity again 95%. But for the detection of treponemal DNA in ocular fluid, although reported in a small number of patients,4–6 it is not well-validated and for both aqueous and vitreous humor, neither sensitivity nor specificity are clear. A TaqMan probe-enhanced real-time PCR is used by our laboratory, with 40–45 amplification cycles, positivity being reported at ≤40 cycles.

Our first patient presented with signs suggestive of intraocular candidiasis and also had demonstrable risk factors for that disease (repeated intravenous access, leukopenia). However, the positive syphilis ELISA in combination with a high-titer RPR made it clear that he had active syphilis and that in retrospect it had probably been responsible for the patient's (previously presumed fungal) rash and for his snail-track mucosal ulceration. Despite the absence of suggestive intraocular signs, the positive syphilis PCR, the negative tests for fungi, and the resolution on treatment prove the diagnosis of intraocular syphilis.

To confirm a diagnosis of neurosyphilis, the UK National Guidelines on the Management of Syphilis 2008 (commissioned by the British Association for Sexual Health & HIV) require a TPPA titer of 1:320 from cerebrospinal fluid7 because lower titers may reflect passive dissemination of antibodies from serum. Our first patient (titer 1:160) did not quite achieve that threshold, but since most specialists regard ocular involvement as a form of neurosyphilis, he was in any case given standard neurosyphilis treatment. Our second patient provoked interesting questions about atypical disease presentations, false-positive and false-negative tests, and the possibility of dual (or even triple) pathology in immunodeficient patients. Despite a clinical suspicion of syphilitic retinitis in this patient, we were unable to obtain a positive PCR for T. pallidum from two successive aqueous taps, but were diverted by a positive PCR test for HSV-1, which may or may not have been true positive. For safety, antiviral treatment was also used. In due course PCR testing on a vitreous sample was positive and the response to treatment was supportive of that diagnosis. It is not yet known whether the sensitivity of treponemal PCR testing is superior from vitreous compared to aqueous, but it may be, and it is clear, firstly, that not all diagnostic efforts are immediately rewarded, and, secondly, that delayed diagnosis and treatment may worsen the visual outcome, as seen here.

Conclusion

For the diagnosis of active syphilis, serology remains fundamental. However, in a patient with uveitis, either with equivocal syphilis serology, or with atypical clinical signs, or if HIV positive where co-infection is possible, PCR on intraocular fluid may prove the diagnosis and steer doctors toward appropriate therapy.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.