British Ocular Syphilis Study (BOSS): 2-Year National Surveillance Study of Intraocular Inflammation Secondary to Ocular Syphilis

Abstract

Purpose.: The British Ocular Syphilis Study (BOSS) is the first national prospective epidemiological study of intraocular syphilis (IOS) in light of the global increase in early syphilis (ES). The aims were to ascertain the UK incidence, demographics, clinical features, laboratory data, and posttreatment visual outcomes of patients with IOS.

Methods.: Prospective study of IOS, reported through the national reporting system (British Ocular Surveillance Unit) from 2009 to 2011. Case definition was any adult presenting with intraocular inflammation in ES.

Results.: A total of 41 new cases (63 eyes) of IOS were reported, giving an annual incidence of 0.3 per million UK adult population. Mean age was 48.7 years (range, 20.6–75.1); 90.2% were male. All had RPR/VDRL titers of ≥1:16. Bilateral ocular involvement occurred in 56%; in unilateral cases, the left eye was more commonly affected (P = 0.009). Mean presenting logMAR visual acuity was 0.52 (20/63 Snellen; range, −0.2 to 2.30 logMAR). Panuveitis was the commonest diagnosis, seen in 41.3%, and isolated anterior uveitis was uncommon (9.5%). Subgroup analysis between HIV-positive and -negative patients found no significant differences in terms of proportion of bilateral disease, presenting or post treatment acuity. HIV-positive patients had higher rates of panuveitis. At final follow-up, 92.1% had visual acuity ≥ 0.3 logMAR (20/40 Snellen) after antibiotic therapy.

Conclusions.: This study is the largest prospective series of ocular syphilis in the post-penicillin era. It confirms good visual outcomes for treated IOS, irrespective of HIV status or time to presentation. The study identified an unexpected preponderance for left eye involvement in uniocular cases; which is unexplained.

Introduction

The introduction of penicillin in the 1940s led to effective treatment of infectious syphilis, but its eradication has remained elusive. Infectious syphilis currently poses a significant public health problem and there has been a dramatic increase in its incidence in the United Kingdom, United States, China, and Europe. In the United Kingdom, there was a 1032% increase in the incidence of syphilis between 1999 and 2008, with a peak in 2007 of 3762 new cases and has remained high. ¹ This exponential increase has been attributed to unsafe sexual practices mainly among men who have sex with men (MSMs), who accounted for 73% of infectious syphilis. Human immunodeficiency virus (HIV) coinfection also is common and was found in 27% of cases overall and 50% of MSMs. ¹  

Intraocular syphilis (IOS) is an important manifestation of early syphilis (ES). Increasing numbers of case reports and series in the literature suggest that ocular syphilis is concurrently on an upward trend. These have been summarized in Table 1. However, there have been no prospective, population-based studies evaluating the incidence and features of ocular syphilis. 

Table 1

 

Summary of Recently Published Case Series

Study	Years	Study	No. of Patients	Concurrent HIV	Posterior Segment Involvement, Panuveitis, Posterior Uveitis, Optic Neuritis	Location
Tamesis RR2	1983–1989	Retrospective	25	5/25	12 patients	MA, USA
Shalaby IA3	1983–1995	Retrospective	13	13/13	6 patients	Baltimore, MD, USA
Villanueva AV4	1993–1996	Retrospective	20	9/20	20 patients	Detroit, MI, USA
Anshu A5	1995–2006	Retrospective	22	0/22	12/29 eyes	Singapore
Kunkel J6	1998–2006	Retrospective	24	11/24	Not specified	Berlin, Germany
Balaskas K7	1999–2009	Retrospective	26	2/26	31 eyes	Lausanne, Switzerland
Yang P8	2004–2008	Retrospective	19	4/19	28/35 eyes	Chongqing, China
Puech C 9	2005–2007	Retrospective	8	5/13	8/10 eyes	Grenoble, France
Hughes E10	2006–2009	Retrospective	13	6/13	17/19 eyes	Sydney and New South Wales, Australia
Eandi C 11	3 years	Retrospective	16	9/16	25/25 eyes	Multicenter, UK
Mathew RG	2009–2011	Prospective	41	13/41	48/63 eyes	Multicenter, UK

To answer this question, we set up the British Ocular Syphilis Study (BOSS), a national prospective population-based study to determine the incidence of ocular syphilis in adults in the United Kingdom. We report the incidence, demographics, clinical presentation patterns, and final outcome data. This study also examines the association of ocular syphilis with HIV and whether HIV coinfection influences the morbidity of IOS. BOSS was set up in conjunction with the British Ophthalmic Surveillance Unit (BOSU) and input from the Health Protection Agency (HPA). 

Methods

Patients with a new diagnosis of ocular syphilis were identified prospectively through active surveillance by BOSU. It was set up in 1997 to enable the systematic investigation of the incidence and clinical features of rare eye conditions of public health and scientific importance. Over the 2-year surveillance period, all consultant ophthalmologists in the United Kingdom were sent a “yellow reporting card,” which they returned, stating how many cases of ocular syphilis they had seen that meet the case definition. After case notification, incident and follow-up questionnaires were sent by the investigators to the reporting ophthalmologists. 

In the United Kingdom, sexually transmitted infections, including syphilis, that are seen in the Genito-Urinary Medicine (GUM) clinics are notified to the respective data collection agencies in England (HPA), Scotland (Information Services Division, NHS National Services), Wales (Communicable Disease Surveillance Centre), and Northern Ireland (Public Health Agency). Cases of ocular syphilis identified through BOSU were cross-referenced with HPA data to confirm stage of disease, sexual orientation, and HIV status. 

We undertook a 2-year period of surveillance from May 2009 to May 2011. 

The case definition was any adult patient with intraocular inflammation on presentation to the ophthalmologist with positive treponemal serology (positive Rapid Plasma Reagin [RPR]/Venereal Disease Research Laboratory [VDRL] and positive treponemal enzyme immunoassay or treponemal pallidum particle agglutination). In addition, all cases of late and treated syphilis, and other treponemal infections, such as Yaws in Afro-Caribbean individuals, were excluded by a GUM physician (BTG) on the basis of history, clinical signs, and results of the RPR/VDRL titer (<1:8), as the study was designed to investigate IOS in ES. The Standardization of Uveitis Nomenclature working group anatomic classification of uveitis was used to classify the uveitis by primary site of inflammation. ¹²  

The baseline incident questionnaire collected data on four categories: patient demographics, systemic features, visual acuity and eye examination findings, and serological investigations. The follow-up questionnaire was sent a minimum of 6 months after the initial presentation to determine the final visual acuity. This study was conducted in accordance with the tenets of the Declaration of Helsinki and received multicenter research ethics approval. 

The incidence was calculated by using the Office of National Statistics estimates for the UK population older than 18 years. Data were transferred to a database and analyzed with IBM SPSS version 21 (32 bit, IBM SPSS Statistics; IBM Corporation, Chicago, IL, USA). Visual acuities were reported in diverse notations, so were all transferred to logMAR. Low levels of acuity, such as counting finger vision or less, were also transposed to widely accepted logMAR equivalents. 

Data variable distributions were tested for normality, and where appropriate t-tests were performed, or their nonparametric equivalents. Null hypothesis testing was performed with χ² or Fisher's exact test where appropriate. The level of significance was set to P is less than 0.05. 

Results

A total of 59 cases of ocular syphilis were reported between May 2009 and May 2011. Questionnaires were not returned for six cases, three cases were a reporting error, and nine cases were excluded because of double reporting, misdiagnosis, or not meeting the inclusion criteria. 

Over the 2-year surveillance period, 41 cases met the inclusion criteria; 22 cases in the first year and 19 in the second. This gives an estimated annual incidence of ocular syphilis of 0.3 cases per 1,000,000. In total, 64 eyes were affected; one was excluded from analysis due to end-stage primary open angle glaucoma affecting the baseline acuity, leaving 63 eyes included in the analysis. The demographic data are reported in Table 2. One hundred percent of the follow-up questionnaires were returned. 

Table 2

 

Demographic Data

	n = 41 Patients; 63 Eyes
Male:female	9 males:1 female (n = 37 males; 4 females)
Mean age at presentation	48.7 y (range, 20.6–75.1 y; median, 47.5 y)
Ethnicity	90.2% (37) white; 7.1% (3) Afro-Caribbean; 2.4% (1) Arab
Sexual orientation	51.2% (21) MSM; 22.0% (9) HT; 22.0% (9) unknown; 4.9% (2) bisexual
HIV status	31.7% (13) HIV-positive; 58.5% (24) HIV-negative; 9.8% (4) unknown
Laterality	56% (23) bilateral; 44% (18) unilateral (left 13; right 5)

 HT, heterosexual.

Clinical Presentation Patterns

The mean duration of symptoms was 1 month before presentation (range, 0 days to 4 months); 56% of patients had bilateral involvement, and the mean presenting logMAR visual acuity was 0.52 (20/63 Snellen; range, −0.2 to 2.30 logMAR). Mean presenting visual acuities of affected eyes for those with unilateral disease was 0.65 logMAR, marginally worse than those with bilateral disease (0.46 logMAR) (P = 0.2). Presenting acuity was not influenced by duration of visual symptoms. In those with unilateral disease, the left eye was more commonly affected than the right eye (P = 0.008, χ² test). 

Mean IOP on presentation was 13.9 mm Hg (range, 8–27 mm Hg); in one patient (1.7%) it was higher than 22 mm Hg. 

Panuveitis was the commonest diagnosis, found in 41.3% of affected eyes. Pure anterior uveitis (AU) was seen only in 9.5% (six eyes); however, AU occurred in 69.8% of eyes overall, and 54.0% had posterior segment inflammation, either as pure posterior uveitis (12.7%) or as part of a panuveitis 41.3%. 

Vitritis was the commonest posterior segment finding, found in 65.1% of eyes. 

Fourteen eyes had optic nerve involvement, of which 50% (seven eyes) had associated AU ± vitritis. Figures 1 and 2 show the frequency of different signs and diagnoses, respectively. 

Figure 1

 

Frequency of different signs in patients presenting with ocular syphilis.

Figure 2

 

Diagnoses of patients presenting with ocular syphilis.

Of the 41 patients with IOS, 29.3% (12) had systemic signs of ES, such as generalized rash, mucosal lesions, or generalized lymphadenopathy; 31.7% (13) had no signs, and the presence or absence of signs was not reported in 39.0% (16). 

Human Immunodeficiency Virus Status and Sexual Orientation

Through the HPA and our questionnaire, we were able to ascertain HIV status of individuals; 31.7% (13) were HIV-positive, 58.5% (24) were HIV-negative, and 9.8% (4) had unknown HIV status. The mean age of the HIV-positive patients was significantly younger (38.6 years) than the HIV-negative patients (54.6 years) (P = 0.001). All HIV-positive patients were men. 

In total, 51.2% (21) of patients were MSM, 22.0% (9) were heterosexual, 22.0% (9) had unknown sexual orientation, and 4.9% (2) were bisexual. Of the 13 patients who were HIV-positive, nine were MSM, one was heterosexual, and three had unknown sexual orientation. One hundred percent of ophthalmologists referred their patients to a GUM specialist. 

Rapid Plasma Reagin/VDRL Titers in Relation to HIV Status

Venereal Disease Research Laboratory titer was available in 31 (76%) of 41 patients with a median VDRL of 1:128 (range, 1:16–1:2048). In HIV-positive patients, the median titer was 1:256 (range, 1:128–1:512), and higher than the median titer of HIV-negative patients (median, 1:128, range, 1:16–1:2048). This approached significance (P = 0.053, Mann-Whitney U test). There were 92.6% who had a titer greater than 1:32. 

Ophthalmic Clinical Features in Those With HIV

Subgroup analysis between HIV-positive and HIV-negative patients found no significant differences between the groups in terms of the proportion of bilateral disease, presenting acuities, or final acuity outcome of affected eyes. 

In terms of clinical features, affected eyes from HIV-positive patients were much less likely to have uveitis limited to anterior or intermediate segments (1/21 eyes) compared with known HIV-negative patients (13/37 eyes) (P = 0.02, Fisher's Exact test). Although the HIV-positive eyes had higher rates of panuveitis, rates of macular edema were less common in this group (1/21), compared with HIV-negative affected eyes (10/37) (P = 0.04, Fisher's Exact test). 

Final Outcomes

Final outcome data were available for 100% of patients. At final follow-up, 92.1% had visual acuity of 0.3 logMAR (20/40 Snellen) or better, after antibiotic therapy (see Fig. 3). There was no correlation between presenting acuity and posttreatment acuity; HIV status did not influence the final visual acuity outcome. Only four eyes (6.3%) had final acuities worse than their presenting acuities. Details of the four cases are outlined in Table 3. All patients had posterior segment involvement. 

Figure 3

 

Presenting acuity plotted against the final acuity.

Table 3

 

Details of the Four Cases With Worse Visual Outcomes Following Treatment

Case	Bi-/Unilateral	Eye Affected	HIV Status, 1 = positive	Presenting Acuity, logMAR	Final Acuity, logMAR	Time to Presentation, mo	Diagnosis
18	Bilateral	L	1	0.0	0.18	0.5	Panuveitis
31	Bilateral	L	0	0.0	0.18	0.5	Posterior uveitis
34	Unilateral	L	0	1.0	1.86	1.0	Posterior uveitis
42	Bilateral	L	0	0.0	0.3	0.5	Panuveitis

Discussion

Infectious syphilis has increased exponentially in the United Kingdom, Europe, United States, and China over the past decade and poses a significant public health problem. Diagnosis in the early stages is key, as it is a treatable condition. BOSS was set up as a prospective study to determine whether new cases of IOS were increasing in tandem with systemic syphilis in the United Kingdom. 

Our study found an estimated annual incidence of ocular syphilis of 0.3 cases per 1,000,000. In 2009, there were 2854 cases of syphilis and in 2010, 2654 cases in England alone. ¹³ If we include only IOS reported in England (17 cases in 2009 and 16 cases in 2010), we can calculate that it affects approximately 0.6% of all those affected with infectious syphilis every year. To our knowledge, there are no other national data for ocular syphilis. 

It is unclear why the incidence has not reflected the exponential increase seen in infectious syphilis. We postulate that there may be increased awareness among individuals, particularly MSMs, due to various nationwide health campaigns and as a result improved health-seeking behavior. Early diagnosis and treatment of infectious syphilis may not allow the condition to progress through the stages and cause ocular syphilis. Interestingly, a study looking at patients with secondary syphilis, found subclinical anterior uveitis in almost 50% of patients; this may again account for the low incidence. ¹⁴  

The mean age of patients was 48.7 years, with wide range of 20.9 to 75.1 years. This is in keeping with reports in the literature of 41 to 58 years. ^4,10,15 The mean age of patients with ocular syphilis in HIV populations has been found to be slightly lower, at 37 to 40 years; our findings agree with this. ^3,16–18  

This is the first prospective national epidemiologically robust study to confirm a strong male predominance, with nine men for every one woman diagnosed. This trend also is found in ES, highlighting this condition in the MSM population. Most of our patients were Caucasian, which is in contrast to American studies where most patients were black. ^4,19  

Ocular syphilis can present in all stages of the disease; studies often do not discriminate the staging of ocular syphilis. Our study looked at ocular presentation in ES and excluded late latent or tertiary syphilis. Unfortunately, systemic features of syphilis, such as rash, lymphadenopathy, mucosal lesions, or chancre were poorly reported in our study. Of the cases reported, 29.3% had signs and 31.7% had no signs; 39.0% were not reported. There are mixed reports in the literature about systemic features of syphilis appearing before or in conjunction with ocular syphilis. A paucity of systemic signs were reported in a Southeast Asian cohort, although these patients were all diagnosed in the latent phase of disease. ⁵ This study did not differentiate between early or late latent stages of the disease; no signs would be expected in late latent syphilis. Essentially, those with “early” latent syphilis should be classified as secondary syphilis and those with “late” latent as tertiary syphilis. Other studies have found that 40% to 69% of ocular syphilis patients have systemic features. ^3,10,15,16 A systematic review of 101 case reports and series in the literature of HIV-positive individuals with ocular syphilis found 55% to have rash related to infectious syphilis. ¹⁷ The elicitation of systemic features is important, as it may aid diagnosis. In our study, 29.3% had associated systemic features of ES, and syphilis serological tests are required to confirm diagnosis. Several series in the literature have reported initial vitreous biopsy and treatment for viral retinitis. ^10,20 Indeed, in our series, reporting ophthalmologists were considering potential diagnoses of acute retinal necrosis in two cases and cytomegalovirus retinitis in one case. Other diagnoses that were considered by reporting ophthalmologists included sarcoid uveitis, diabetic retinopathy, giant cell arteritis, and ocular lymphoma. A high index of suspicion may prevent unnecessary investigations and treatment. Syphilis should be excluded in all patients with ocular inflammation. 

Neurosyphilis has been shown to be associated with an RPR greater than or equal to 1:32 and HIV-induced immune impairment. ²¹ In our study, the lowest titer was 1:16, with a median of 1:128, and in HIV-positive the lowest was 1:128 with a median of 1:256. There were 93% with a titer greater than 1:32, which is consistent with the findings of that study. 

The mean presenting visual acuity was 0.52 logMAR (20/63 Snellen) (−0.2 to 2.3 logMAR); 50% had a presenting acuity of 20/60 or worse. However, at final follow-up, 92% had acuities better than 20/40, which is the UK visual acuity driving standard. This highlights the importance of diagnosing ocular syphilis, particularly in the early stages, as most cases make excellent visual recovery. 

The mean time to presentation was 1 month, with a range of 0 (incidental finding) to 4 months. Those with worse visual acuities tended to present early and, interestingly, those who presented late did not have poor acuities. 

Bilateral ocular involvement was seen in 56%. Several reports in the literature have found higher rates of bilateral disease in HIV-positive patients, 63% to 83%, and suggested HIV may modulate the disease process in some way. ^3,17 However, this was not the case in our prospective study, with HIV status bearing no influence on disease laterality. However, we were unable to correlate disease laterality with severity of HIV disease, such as CD4 cell count, and whether patients were already on antiretroviral therapy. 

Interestingly, there was a preponderance of left eye involvement in those with unilateral disease (P = 0.009, Fisher's Exact test). This has not been previously reported in ocular syphilis. We could not find any other studies pertaining to infectious uveitis in the literature where the left eye was predominantly affected. The reasons for the difference are currently unknown. It is postulated that syphilis is primarily a vasculitis, and, therefore, the eye receiving more direct blood flow may be more likely to be affected. On the left-hand side, the common carotid originates directly from the aortic arch, whereas on the right side, the brachiocephalic artery emanates from the aortic arch and then bifurcates into the right subclavian artery and common carotid, which supplies the right eye. In the same way, plantar rashes in secondary syphilis are less likely to be on pressure areas of the sole, compared with nonpressure areas, due to a reduced blood supply. 

Of note, patients with normal-tension glaucoma have been found to have worse disease in their left eye, compared with the right, ²² although the reasons for this remain unknown. 

Panuveitis was the commonest diagnosis, found in 41.3% of our series. Vitritis was the most frequent posterior segment finding, followed by retinitis. Another study also found vitritis to be the commonest complaint. ¹⁵ In a series of 13 HIV-positive patients, panuveitis was the commonest diagnosis, found in 38% of patients. ³ It should be noted that isolated anterior uveitis was found in only 9.5% of our cases, suggesting inflammation related to ocular syphilis is more likely to affect the posterior segment or the anterior and posterior segments together. No cases of isolated anterior uveitis were found in a series of 26 patients with ocular syphilis. ⁷  

Our prospective epidemiological study showed that 92% of patients had a final acuity of 20/40 or better. This is corroborated by other studies. ^7,18 The presenting acuity did not appear to prejudice posttreatment outcome acuity. Interestingly, patients with coexisting HIV disease did not do worse in terms of outcome acuities compared with HIV-negative patients. This contrasts with another study, which found worse presenting acuity and HIV positivity to have a negative influence on final acuity. ⁸  

In our study, 51.2% were MSM and 31.7% were HIV-positive overall. Of the 13 patients who were HIV-positive, nine were MSM, one was heterosexual, and three had unknown sexual orientation. It should be noted that HIV patients had presenting acuities and final acuities no worse than those without HIV. This group did not tend to have more bilateral disease, which is in contrast to previous case series. ^3,23  

Certain forms of infectious uveitis commonly present with an associated raised IOP in the acute phases, namely herpetic and toxoplasma-related uveitis. A retrospective case series suggested syphilitic uveitis also was associated with an acute hypertensive phase. ²⁴ However, our study found only one case presenting with an ocular pressure of 27 mm Hg. This was the only case with an IOP greater than 22 mm Hg. 

Although the incidence of ocular syphilis is low, our study highlights that it is an important diagnosis to make, as most patients achieved excellent acuities after treatment. BOSS also underlines the importance of ascertaining HIV status in this group of patients, as one-third of our patients were HIV-positive. This reflects the sexually transmitted nature of both conditions, and a diagnosis of IOS may bring to light an HIV diagnosis for the first time. ⁶ Thus, all cases of syphilis should have an HIV test. 

However, as with all similar reporting systems, complete case ascertainment through the BOSU is not possible, and the sexually transmitted infections surveillance systems in the United Kingdom do not collect information on ocular disease related to syphilis, so we were not able to perform capture-recapture analysis. Therefore, the incidence of ocular syphilis may be slightly higher. We were able to verify some demographic data, as well as HIV status, by using probability matching between our cases and those on the HPA syphilis register. Another study limitation was that we were not able to gather information on CD4 counts and whether the patients with a known diagnosis of HIV were on antiretroviral therapy and which patients were diagnosed with HIV because of their IOS presentation. This is important, as patients with low CD4 counts are more likely to develop neurosyphilis. ²¹  

In summary, we believe BOSS to be the largest prospective study looking at the incidence of ocular syphilis in a Western European population. Ocular syphilis is a rare but important diagnosis to make, as it responds extremely well to treatment and can cause serious morbidity if untreated. A high index of suspicion is required to diagnose IOS secondary to ES, as it can be the only feature of ES. Elicitation of systemic signs and syphilis serology are required to aid diagnosis. It commonly causes panuveitis, and HIV coinfection is not uncommon and should be excluded in all cases. Commoner left eye presentations in unilateral disease is a new finding in this study and the reason remains unknown. 

Acknowledgments

The authors thank the British Ophthalmological Surveillance Unit for generously supporting this study, in particular Barny Foot and Miles Stanford for their invaluable guidance in planning the study. We also thank Susan O'Hea and Sarah Davies for their help with study administration, and Ian Simms from the HPA for his help with cross-referencing data. 

We thank the following ophthalmologists for their contribution of cases to the study: PL Atkinson, W Ayliffe, R Bates, D Charteris, Y D'Souza, J Deutsch, H Devonport, A Dick, P Fox, E Graham, P Hassett, L Herbert, R Humphry, G Jayamanne, C Jones, J Keenan, J Khan, M Kulshreshtha, G Larkin, S Madill, B Matthews, M McKibbin, P Murray, P Nithianandan, V Pai, R Pandit, S Patterson-Brown, A Plumb, A Raghu Ram, A Smith, S Spencer, M Stanford, and G Williams. 

Supported by Fight for Sight, UK, Guide Dogs for the Blind, UK, and an ARVO travel grant. 

Disclosure: R.G. Mathew, None; B.T. Goh, None; M.C. Westcott, None