Syphilitic uveitis: a review of clinical manifestations and treatment outcomes of syphilitic uveitis in human immunodeficiency virus-positive and negative patients

Abstract

The incidence of syphilis and syphilitic uveitis in our community is increasing. The prevalence of associated neurosyphilis is unknown, and it remains unclear whether syphilitic uveitis should be treated as secondary syphilis with intramuscular penicillin or neurosyphilis with intravenous penicillin. The (English language) literature was reviewed for all unique cases of syphilitic uveitis reported from 1984 to June 2008. For each case the following data were recorded: the clinical features of the syphilis, the uveitis and any associated neurosyphilis, the human immunodeficiency virus (HIV) status, lumbar puncture findings, treatment and follow up. We identified 143 patients in 41 original reports of syphilitic uveitis (93 HIV-positive and 50 HIV-negative). Posterior uveitis was reported in 79 patients (55.2%); panuveitis was reported in 36 patients (25.2%); anterior/intermediate uveitis was reported in only 28 patients (19.6%). Lumbar puncture findings were abnormal in 82 patients (57%), and the majority of these patients (76%, 62 out of 82) were HIV-positive. One hundred and ten (77%) patients were treated with intravenous therapy, usually penicillin. Most recovered from the syphilis, however, a proportion did not recover full vision. There were 13 (9%) treatment failures, which tended to occur in patients who were HIV-positive (n = 11), had abnormal lumbar puncture findings (n = 8) and/or were treated (n = 11) intravenously. There is a high incidence of abnormal lumbar puncture findings in patients with syphilitic uveitis and a strong association with HIV infection. Most received appropriate therapy with a low relapse rate, which was not related to the type of therapy.

Introduction

Uveitis is the most common ocular manifestation of acquired syphilis and the incidence of syphilitic uveitis in our community is currently increasing.^1–3 However, the prevalence of neurosyphilis in patients with syphilitic uveitis is unknown. This is an important distinction as the presence of neurosyphilis determines the treatment recommendation. It is currently unclear whether all syphilitic uveitis should be treated as secondary syphilis with intramuscular (i.m.) penicillin or as neurosyphilis with 2 weeks of intravenous (i.v.) penicillin.³

It has been proposed that retinitis and optic neuritis should all be treated as neurosyphilis because if the eye is an extension of the brain, then involvement of anything derived from neuroepithelium could sensibly be regarded as neurosyphilis. There is however ongoing debate about anterior uveitis. Uveitis is considered by some only as a risk factor for neurosyphilis. Many authors suggest that all ocular involvement should be considered identical to neurosyphilis^4–8 and also encourage cerebrospinal fluid (CSF) examination^3,4 in all such patients.

Co-infection with human immunodeficiency virus (HIV) further complicates the picture. Patients with HIV are more likely to contract syphilis.¹ Coexisting HIV infection may also accelerate the progression of syphilis^7,9 and increase the probability of relapse.¹⁰ Shalaby et al.¹¹ found that the prevalence of neurosyphilis was high in HIV-positive patients with syphilitic uveitis, with seven out of 11 patients testing CSF Venereal Disease Research Laboratory test (VDRL) positive for neurosyphilis. Although there are numerous cases reported in the literature, there are few prospective studies. Frohman and Lama performed a review in 2000¹² and found that failure of serological recovery was more common in HIV-positive patients. This was based on one study from 1997¹³ of patients with early syphilis, using as treatment i.m. penicillin versus i.m. penicillin with 10 days of oral amoxycillin and probenecid. Interestingly, 25% of these patients had Treponema pallidum detected in the CSF even though they had early stage syphilis. In addition, this review identified other studies that reported non-penicillin therapies. Passo and Rosenbaum reported one HIV-positive patient who tested negative to both HIV and syphilis on first presentation, but 6 months later tested positive to both HIV and syphilis with a high VDRL titre of 1 : 256.¹⁴

The current treatment of syphilitic uveitis relies on our understanding of the natural history, clinical features and response to treatment of syphilitic uveitis in HIV-positive and HIV-negative patients; however, there is not yet a definitive study to provide guidelines. A review of the literature was performed, specifically to determine the incidence of associated neurosyphilis in patients presenting with syphilitic uveitis, and the requirement for aggressive i.v. penicillin in patients with syphilitic uveitis.

Method

A review of the literature was performed using Medline via OVID from 1984 to June 2008. The Medline search was performed using search terms ‘Eye’ or ‘Ophthal*’ and ‘Syphilis’ or ‘syphil*’, limiting the search to English language literature and humans. The identified papers were reference mined to ensure that no relevant papers were ignored. All original case reports or papers describing unique patients presenting with syphilitic uveitis were examined. Only papers specifically identifying the HIV serology (whether positive or negative) of the patients were included.

For each case, the presenting features of the uveitis (e.g. anterior uveitis or retinitis), the systemic features of the syphilis, the clinical features of neurosyphilis (if any), investigatory evidence of neurosyphilis (e.g. lumbar puncture results), treatment type and treatment outcome were recorded.

Lumbar puncture findings were considered abnormal if there was a positive VDRL on the CSF (clinically confirmed neurosyphilis according to the Centres for Disease Control (CDC) criteria^15,16) or any abnormal elevation in cells or protein in the CSF (clinically probable neurosyphilis according to the CDC criteria^15,16). That is any CSF abnormality without an obvious alternative explanation was regarded as indicative of neurosyphilis in this review.

Treatment was grouped according to the antibiotic therapy. The first group were those who received i.v. penicillin for a 6–21-day course (sometimes followed by weekly i.m. course especially if the i.v. course was of shorter duration). The second group received daily i.m. penicillin for 10–21 days, which was considered equivalent to i.v. therapy by the authors who used this therapy. The third group received weekly i.m. penicillin followed by i.v. penicillin. The fourth group received weekly i.m. penicillin. The final group included any other antibiotic therapy.

Treatment outcome and patient recovery was determined from documented subsequent serology and visual symptoms.

The data were tabulated to summarize the findings and the frequency of each category was determined. Statistical analysis was performed to determine associations between HIV status and presentation of uveitis, HIV status and CSF findings and presentation of uveitis and CSF findings. The confidence intervals for proportions were calculated using Fisher's exact test, and the confidence intervals for relative risk were calculated using Taylor Series.

Analysis of predictive factors (HIV status, site of uveitis and treatment regimen) for treatment outcome was limited by short follow-up periods (as described following).

Results

One hundred and forty-three original descriptions of patients with syphilitic uveitis were identified. These cases were identified in 41 different papers.^{2,6,7,9–11,14,17–52}

Data for certain parameters could not be collected as they were not described in the original articles. In particular, systemic features of syphilis were recorded in only 39, and clinical features of neurosyphilis were recorded in only 15 of 143 patients. Although the treatment outcome (clinical features and/or serology) was described in all patients, the follow-up period for these patients was variable. Forty patients were followed up for an unspecified period or less than 1 month; 65 patients were followed up for between 1 and 6 months; 20 patients were followed up for between 6 months and 1 year; only 18 patients were followed up for longer than 1 year.

In addition, two papers^10,53 did not describe individual patients so the data provided by these papers lacked sufficient detail to determine relationships between risk factors and outcomes.

Clinical presentation

The type of uveitis, systemic features of neurosyphilis and HIV status were determined for each patient (when reported) as defined in the Method. Table 1 describes the main presenting features of the syphilitic uveitis. Sixty-five per cent of patients in this review were HIV-positive (65%; 95% CI 57–73%). The most common presentation in HIV-positive patients was posterior uveitis (51.6%; 95% CI 41–62%). Of the patients who were diagnosed with isolated anterior uveitis, all but one, 96% were HIV-positive (95% CI 82–99.9%). From these data, patients who presented with syphilitic anterior uveitis were 14.5 times (relative risk) more likely to be HIV-positive than HIV-negative (P = 0.00002; 95% CI 2.03–103.7).

Table 1.  Presentation of uveitis

	Anterior uveitis or intermediate uveitis	Posterior uveitis	Panuveitis	Total patients
HIV, human immunodeficiency virus.
HIV-positive	27 (96.4%)	48 (60.8%)	18 (50%)	93 (65%)
HIV-negative	1 (3.6%)	31 (39.2%)	18 (50%)	50 (35%)
Total	28 (20%)	79 (55%)	36 (25%)	143

Cerebrospinal fluid abnormalities

The incidence of CSF abnormalities (as defined in the Method) was also high in the group of patients reviewed (see Table 2). Eighty-one per cent of HIV-positive patients underwent a lumbar puncture, compared with 66% of HIV-negative patients. Eighty-three per cent of the HIV-positive patients and 61% of HIV-negative patients who underwent lumbar punctures had abnormal CSF findings. In the group of patients who had lumbar punctures, those who were HIV-positive were 1.3 times more likely to have abnormal CSF (95% CI 1.017–1.83, 2-tailed P-value of 0.013). Table 3 arranges those patients that did have lumbar punctures performed according to their clinical pictures (27 patients were excluded because of lack of detail on individual patients). Table 3 shows a trend towards patients with HIV and posterior uveitis being most likely to show CSF abnormalities. Using logistic regression to analyse the relationship between lumbar puncture findings, site of uveitis and HIV status, a statistically significant association between HIV status and lumbar puncture abnormality was found (OR 3.68; 95% CI 1.21–11.22). No significant association between site of uveitis (anterior, posterior and panuveitis) and lumbar puncture status could be demonstrated after adjusting for HIV status.

Table 2.  CSF analysis for neurosyphilis

	Lumbar puncture abnormal	Lumbar puncture normal	Lumbar puncture not performed	Total patients
CSF, cerebrospinal fluid; HIV, human immunodeficiency virus.
HIV-positive	62	13	18	93
HIV-negative	20	13	17	50
Total	82 (57%)	26 (18%)	35	143

Table 3.  Relationship of uveitis to CSF abnormalities^†

	HIV-positive; LP abnormal	HIV-positive; LP normal	HIV-negative; LP abnormal	HIV-negative; LP normal	Total patients
^† Only patients who had LPs performed are included in Table 3, and a further 27 patients were excluded because of lack of detail on individual patients. CSF, cerebrospinal fluid; HIV, human immunodeficiency virus; LP, lumbar puncture.
Anterior uveitis/intermediate uveitis	7	1	0	0	8
Posterior uveitis	22	0	16	9	47
Panuveitis	12	6	4	4	26
Total	41	7	20	13	81

Treatment

The treatment protocols reported in the literature for syphilis are outlined in Table 4. A variety of treatment protocols were identified. Most patients were treated with a neurosyphilis regimen, either with i.v. penicillin for 10 days or more (76%, n = 108), or with daily i.m. penicillin for the same time period (n = 6), which was considered equivalent by the authors using this particular protocol. Only a small number of patients received lesser therapy of either weekly i.m. penicillin or oral medication. There was only one patient who did not receive any treatment at all and was lost to follow up (see Table 4).

Table 4.  Treatment of syphilis

	HIV-positive	HIV-negative	Total patients
HIV, human immunodeficiency virus. i.m., intramuscular; i.v., intravenous.
i.v. penicillin (10–21  days or equivalent)	76	32	108
i.v. ceftriaxone	2	2	4
i.v. chloramphenicol	1	0	1
Daily i.m. penicillin	3	3	6
Weekly i.m. penicillin	9	7	16
i.m. penicillin followed by i.v.	1	3	4
Oral tetracycline	1	1	2
Oral ampicillin	0	1	1
No treatment	0	1	1
Total	93	50	143

Follow up

Analysis of predictive factors (HIV status, site of uveitis and treatment regimen) for treatment outcome was limited by short and variable follow-up periods. Despite the variable follow-up period, the treatment outcome (clinical features and serology) was described in all patients. Thirteen patients were reported to have failed treatment and required a repeat course of treatment. Eleven of these 13 patients were HIV-positive, 8 of the 13 patients had an abnormal lumbar puncture, and 11 had been treated initially with i.v. penicillin. Of the two patients who had been treated with i.m. penicillin, one had received a weekly i.m. dose for 3 weeks the other had received a neurosyphilis regimen of daily i.m. penicillin for 14 days.

Discussion

Clinical presentation

The large proportion of HIV-positive patients identified in this study agrees with the findings of previous authors;^1,11 however, they may be overrepresented as they may have presented a management challenge and therefore warranted reports in the literature. As only English literature was examined, we do not know whether coexistent HIV with syphilis may be more common in Western countries than elsewhere.

Posterior uveitis was the most common presentation of syphilitic uveitis reported in the literature. An isolated anterior uveitis did occur, but was less common. Interestingly, isolated anterior uveitis was primarily reported in HIV-positive patients; only one HIV-negative patient had anterior uveitis in the literature. This may indicate that a patient with isolated anterior uveitis secondary to syphilis is likely to be HIV-positive. This may be a real phenomenon, but it may also be due to a lack of case reports describing patients with uncomplicated anterior uveitis. The site of uveitis, whether anterior or posterior, could not be related with statistical significance to risk of CSF involvement.

Previous authors including Barile and Flynn⁵⁴ have reported 71% (17/24) of syphilitic uveitis patients present with isolated anterior uveitis, and a further 13% (3/24) with panuveitis. Other authors^55,56 (and older texts) state that anterior uveitis is the most common presentation of syphilitic uveitis. In the literature reviewed we identified that all but one patient with isolated anterior uveitis was HIV-positive, predicting that patients with syphilitic anterior uveitis are far more likely (14.5 times relative risk) to be HIV-positive than HIV-negative. However, this result may be affected by reporting bias. Many of the reports of syphilitic uveitis are from the infectious diseases literature, and there may be some simplification of uveitis definitions or unorthodox reporting of uveitis clinical features.

Treatment of HIV has progressed and changed considerably in the period of this literature review (1984–2008). It would not be possible to make any useful comparison across HIV treatment groups and as such we did not include HIV treatment as a factor in this review.

Cerebrospinal fluid abnormalities

Although our data support an increased incidence of neurosyphilis in HIV-positive patients, it is important to acknowledge that HIV-positive patients can also have CSF abnormalities for other reasons. It is difficult to judge the significance of CSF abnormalities in HIV-positive patients; however, these patients require treatment as for neurosyphilis and close monitoring.

The CDC defines patients with syphilis as confirmed neurosyphilis if they have a positive CSF VDRL; or probable neurosyphilis if they have a negative VDRL but an elevated protein or white cell count in the CSF, with clinical signs or symptoms. Ocular findings are acceptable clinical signs within this definition.^15,16 We considered all CSF abnormalities to indicate a presumptive diagnosis of neurosyphilis for two reasons. First, it is possible that neurosyphilis may present with any type of CSF abnormality.^8,14 Second, as our purpose was to ascertain how often patients with uveitis had CSF changes, which would necessitate aggressive treatment and close monitoring, it was necessary to consider all CSF abnormalities as probable neurosyphilis as per the CDC definition.¹⁵ To be absolutely certain that any CSF changes are not due to syphilis, repeated testing is required but this was not reported routinely in the literature. There are a variety of CSF abnormalities that are consistent with neurosyphilis; however, these are not yet well defined.¹⁶ Using this definition of neurosyphilis it is possible that we have overestimated the incidence of neurosyphilis in our patient group as there may be abnormal CSF findings in HIV-positive patients for many reasons other than neurosyphilis. Nonetheless in the HIV-negative group the rate of CSF abnormality was still 66% so conservatively this is the minimum rate of associated neurosyphilis in patients with syphilitic uveitis.

Anatomically, one would assume that patients with posterior uveitis would be more likely to have associated neurosyphilis than patients with anterior or intermediate uveitis. However, we did not find a statistically significant association between the site of uveitis and CSF abnormalities, especially after correcting for HIV status.

There are a number of possible explanations for the association found between neurosyphilis and HIV. It may be that by the time the uveitis has developed (generally in the course of secondary syphilis) there has been widespread haematogenous dissemination and neurosyphilis simply forms a part of this spectrum. It may be that patients who develop syphilitic uveitis are at increased risk of developing neurosyphilis because of the common origin of the uveal tract and the CNS, and similarities between the blood-ocular barrier and the blood–brain barrier. It may be that the incidence of neurosyphilis is not actually as high as identified by this review as many HIV-positive patients have CSF abnormalities for reasons that are unrelated to syphilis.

Treatment

The variety of treatment protocols used in the literature makes comparison of treatment outcomes difficult. Both the type and duration of therapy reported varied greatly, but most protocols involved penicillin as the first line of therapy. Optimal therapy for neurosyphilis is considered to be four hourly i.v. penicillin from 10 to 21 days, but in some cases daily i.m. penicillin is used for 10 days or longer and considered equivalent to the i.v. therapy by some authors.^{22,26,31,33,36,46} The CDC does consider a 14-day course of daily i.m. penicillin as an alternative to i.v. penicillin,¹⁵ but only when accompanied with oral probenecid.

There was no significant difference in efficacy for the different treatment modalities in this review. It is important to note that even with aggressive full-dose i.v. therapy there is still the possibility of treatment failure and these patients all need monitoring post treatment. Given the apparent lack of definitive benefit of i.v. therapy over i.m. therapy, a case could be made for treating syphilitic uveitis in HIV-negative patients with a lesser regimen of i.m. penicillin, as long as there was adequate follow up.

Follow up

Of the 143 patients included in this review, only 13 were reported to fail treatment. However, because of the highly variable follow-up period, a meaningful statistical analysis could not be performed to assess the significance of these treatment failures. A trend is however apparent, identifying that patients who fail treatment are likely to be HIV-positive, have CSF abnormalities and to have received penicillin as initial therapy. There is inadequate evidence to indicate the optimal duration of ongoing follow up.

Conclusion

This review emphasizes the association between syphilitic uveitis, HIV infection and neurosyphilis. Posterior uveitis was the most common presentation reported. There is a high incidence of abnormal CSF findings in patients with syphilitic uveitis but the site of uveitis, whether anterior or posterior, was not related to CSF findings. Treatment failure appears more likely in patients who are HIV-positive and have CSF abnormalities, and may occur even after full-dose i.v. penicillin therapy. The primary risk factor for failure of treatment appears to be coexisting HIV infection rather than the treatment regimen employed, and these patients require close monitoring even after standard i.v. therapy.

Acknowledgement

The authors would like to thank Dr R Wijemunige MBBS, Department of Epidemiology and Preventive Medicine, Monash University, for assistance with the statistical analysis.