Coinfections and differential diagnosis in immunocompetent patients with uveitis of infectious origin

Toxoplasma gondii is one of the most common human zoonosis, affecting about a third of the world’s population [1]. Around 10% of people that acquire this infection postnatally [2, 3], and up to 80% of children congenitally infected [4, 5], develop ocular toxoplasmosis (OT). This clinical form of toxoplasmosis is the most common etiology of posterior uveitis worldwide [1, 6].

Although in clinical practice a majority of cases of OT are diagnosed by a combination of consistent clinical features and supportive serological results [7], in cases of atypical presentations it is of utmost importance to differentiate OT from other causes of posterior uveitis that share similar clinical characteristics [8–14].

A definitive diagnosis is only obtained after direct evidence of the presence of the parasite in aqueous humor (AH) by polymerase chain reaction (PCR) that amplifies specific Toxoplasma DNA or by determining the eye’s own antibody production through Goldmann-Witmer coefficient (GWC) [15, 16]. These methods cannot only confirm the OT diagnosis but can also rule out other similar infectious diseases [17].

It has been described that the analysis of AH by PCR changed the diagnosis and treatment in more than a third of patients, and it should be considered for uveitis of an atypical clinical form, recurrent severe uveitis of unclear etiology, and therapy refractory cases [18]. As the relative importance of different etiologies changes from one geographical site to another, we aim to evaluate the differential diagnosis of this parasitic infection in immunocompetent patients seen in an Ocular Immunology and Uveitis Service, in Bogotá, Colombia. No previous description of this diagnostic approach has been presented in Latin America.

Five hundred forty-seven patients presented with the diagnosis of uveitis from July 2014 to April 2016, 174 with active uveitis, from which 67 patients were suspected to have uveitis of infectious origin; and 66 patients, who accepted to provide written informed consent, were included in the study period.

Sixty-six consecutive patients – 32 females (47.8%), and 34 males (51.5%), aged between 13 and 79, with a median age of 37 years – who presented with uveitis that was clinically consistent with OT were first studied by Toxoplasma serological tests. According to the results of serum tests, specific Toxoplasma PCR and GWC analyses were performed in AH samples obtained by diagnostic anterior chamber paracentesis (Fig. 1). There were 22 cases of confirmed OT, eight of confirmed OU and five COINF. In total, 35 patients (53%) had etiological diagnosis confirmed by PCR or GWC (Fig. 1). OT was conclusively diagnosed for 27 patients through examination of aqueous humor samples (five of them with COINF): 9 tested positive on both tests (PCR and GWC) in aqueous humor; in 15 patients diagnosis was made by the high load of specific antibodies in aqueous humor alone and in three by the PCR alone. In confirmed OT, 12 out of 27 patients with PCR analysis were tested positive by PCR (44.4%) and 24 out of 27 with GWC analysis (88.8%) were found to be positive.

Patients with UOT (n = 11) because of insufficient AH sample (n = 2) or because they did not undergo AH sampling (n = 3), might have been diagnosed with viral uveítis and/or ocular toxoplasmosis if AH could have been tested.

Considering the clinical signs, panuveitis was more common in patients with COINF, than in patients with UOT (80% vs 6.3%). While posterior uveitis was more common in patients with UOT (87% vs 20%), the presence of anterior uveitis was more frequent in patients with OU including viral origin than in patients with UOT or patients with COINF (21.7%, vs 6.3% vs 0% respectively). The specific clinical signs are shown in Table 1.

Of interest, IgM serological tests for the herpesvirus family (HSV1, HSV2, HSV3 and EBV) were positive without relation to the presence of the virus as detected by qPCR in aqueous humor (Table 1). In total, 9 patients were positive for these IgM assays (7 for EBV and 2 HSV1). One patient was positive simultaneously for IgM anti- HSV1 and EBV. Patients with IgM for several viruses might have had recent infection with one of them, inducing cross-IgM detection. Nevertheless, in our study none of them presented positive anti-Toxoplasma IgM. Also, none of these patients had positive serum tests for syphilis, thus we did not include aqueous humor PCR for Treponema pallidum in the patients. The five COINF patients were further investigated by IgM for CMV, and none of them had positive results.

Intraocular pressure was very high in eight patients (12%). One of these patients with COINF was initially diagnosed as typical recurrent OT, with a bad response to conventional treatment; this patient presented OT + VZV COINF. Another patient with triple COINF (TO+VZV + CMV), had bilateral compromise and did not have any history or clinical or laboratory evidence of immunodeficiency.

Diagnostic doubts are a factor in the management of uveitis [15, 21–24]. There are several differential diagnoses that must be considered when dealing with patients with suspected infectious uveitis [15, 21–24]. Several studies have demonstrated that by combining PCR and GWC the diagnostic certainty for toxoplasmic retinochoroiditis is increased, reaching 80–93% sensitivity and specificity of ~ 93% [23]. In our study, we made an accurate diagnosis of uveitis causes using GWC and PCR in 53% of our patients. The frequency of PCR positivity was similar to that reported in previous studies [16, 25, 26]. To the best of our knowledge, there are no previous studies about the presence of COINF in immunocompetent Colombian patients with OT. In the present work, combined laboratory techniques, in serum and AH, were of significant diagnostic value for patients diagnosed as having other intraocular infections (VZV, EBV, tuberculosis).

GWC is based on the comparison of levels of specific antibodies in ocular fluid and serum samples [26]. It has been reported in a study in French patients that the time interval from symptoms onset to anterior chamber tap is highly correlated with the GWC positivity. In our cases, there were no differences in days of symptoms in patients with positive or negative PCR; a similar result has previously been reported [26]. This could be explained because Toxoplasma infection is caused by different strains in Colombia and Europe. The load of parasites is higher in Colombian patients compared with European patients and this can induce a rapid increase in antibodies [27]. Colombian strains are usually known as more virulent, triggering a higher intraocular inflammatory response [28]. This might influence the delay or intensity of DNA detection or antibody detection in AH. The different strains explain why test results and clinical decisions vary according to the geographical origin of the ocular infection [29]. Our group in Colombia has consistently demonstrated that South American strains producing ocular toxoplasmosis are different and induce a different humoral response with different cytokine mediators [30]. This can help to understand why diagnostic performances of serological and molecular tests differ [31]. It also has been proven that the performance of RE target for PCR amplification is different when used in South American patients [29, 32, 33].

Patients with UOT (n = 16) were treated with anti-Toxoplasma therapy. All of them had a good clinical response. This is consistent with the traditional practice around the world (treating patients with suspected clinical picture of OT with positive serological anti-Toxoplasma tests). If the outcome was favorable under anti-Toxoplasma therapy, then patients could be definitely considered as OT even with aqueous humor negative. PCR is used only in patients with atypical clinical presentations or in patients with bad response to anti-Toxoplasma treatment [34, 35]. However, in our study we carried out these diagnostic techniques in patients with typical and atypical presentation. We have to take into account that a negative PCR result for Toxoplasma gondii does not rule out the diagnosis and that, in these cases, the clinical response to therapy is crucial to confirm the diagnoses [35]. Although there were no typical granulomas in the posterior pole suggesting ocular infection by Toxocara canis in our patients, toxocariasis was ruled out serologically when patients presented posterior uveitis or panuveitis and none of the patients screened had positive antibodies. Although Lyme disease is also an infectious cause of uveitis, it was not screened for, because there are not cases of Lyme disease reported as a cause of uveitis in Colombia and patients did not have any traveling history to the northern hemisphere. Thus, differential diagnosis by PCR in AH was not done for toxocariasis or Lyme disease [36] . As it cannot be excluded that OT was not diagnosed due to lack of sensitivity of the PCR or GWC, we merge the 3 patients with positive serological assay for Toxoplasma with the 20 patients with “other uveitis”. Considering that Mycobacterium tuberculosis and Leptospira [37] can also cause uveitis, we prefer to keep the term as “Other Uveitis (OU)” instead to consider cases of other uveitis only as of viral origin. This population was clinically homogenous regarding the characteristics of uveitis, leading as to think in infectious origin. It is necessary to improve the sensitivity of diagnostic techniques in AH in order to get definitive and conclusive diagnosis in these patients.

Multiple infectious diseases can be associated in immune-deficient patients with human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS), presented as simultaneous or separate infections leading to severe intraocular compromise [38, 39]. In our study, all patients were immunocompetent, including those with co-infection. It has been reported that EBV-DNA was found in HIV negative immunocompromised patients with uveitis. However, authors hypothesized that EBV is not a direct cause of uveitis, but it may play a role as a secondary factor in the pathogenesis of uveitis, producing a homologue of IL-10. This interleukin is known as an immunosuppressant that influences the course of intra-ocular inflammation caused by other pathogens as Toxoplasma, HSV, VZV and CMV [40].

CMV has been associated with anterior, chronic, hypertensive, unilateral and recurrent uveitis in immunocompetent patients [41]. In our work, we did not find CMV infection as a single cause of uveitis in this group of patients. We found CMV infection in coexistence with OT + VZV in one immunocompetent patient. To the best of our knowledge, there are no previous reports about triple intraocular co-infection in the literature in immunocompetent patients. Intraocular COINF of OT associated with viral infection was biologically proven in our patients. COINF presented in immune-competent patients deserves careful consideration. For instance, in a recent study, three immunocompetent patients were AH PCR double positive, for VZV + EBV, Toxoplasma + EBV, and Toxoplasma + HSV [18]. In these double positive cases, therapy was decided according to the clinical criteria, with close follow-up. Double positive PCR has previously been reported in immunosuppressed patients: it is probably due to secondary reactivation of latent parasitic or viral infections triggered by acute infectious uveitis. In the positive cases for EBV, we have to consider that the EBV genome could be present in a latent phase in B lymphocytes, leading to cross reactivity, which can produce false PCR positive results [42]. In immunosuppressed patients, PCR testing for multiple pathogens in ocular fluids can be useful for selecting treatment, since clinical characteristics could be atypical in these cases [43].

PCR in AH samples, GWC and a differential diagnosis assisted with therapeutic trials have been studied by other authors in anterior uveitis [44]. CMV was identified by PCR in aqueous and vitreous samples as the most frequently recognized infectious organism in posterior and panuveitis of HIV-1-negative Thai patients [45]. In our study, none of our COINF patients presented any associated systemic diseases or alterations in ancillary tests. Thus, we excluded primary or secondary immunodeficiencies.

It has been reported that intraocular immune response in OT differs depending on the infecting Toxoplasma gondii strain [46]. Virulent parasites, such as South American strains, may cause more severe OT due to an inhibition of the protective effect of interferon gamma (IFN-γ) [46]. IFN-γ is crucial in the immune response to viral diseases, providing antiviral protection from CMV infection [47]. There is also an important antiviral role for virus-specific CD4(+) T cells in protecting from pathogenic CMV infection [47]. In Colombian patients with OT, the immune response after antigenic stimulation has been found to have a preferential Th2-skewed response, regardless of the nature of antigen stimulus [48] and more severe clinical characteristics in patients infected by type I/III strains [30]; strain-dependent Th2 skewed response should be investigated to identify if it plays a role in favoring this type of co-infections. The Th2-skewed response in patients infected by virulent Toxoplasma gondii strains might have a detrimental role in the defense against viruses, aiding parasites and viral co-infection. T-helper response induction may be confounded by co-infection of a single host by multiple intracellular pathogens. Due to normally adaptive feedback loops that tend to polarize T-helper responses, it can become very difficult for the immune system to mount effective, conflicting responses [48].

Ocular fluids PCR is useful in the diagnosis of uveitis with unusual clinical appearance, severe recurrent intraocular inflammation of unclear origin, and therapy-resistant patients. Additional studies are necessary to analyze how the co-infection affect the therapeutic response and prognosis in this group of patients, and to investigate the need to include additional tests. In future studies, additional molecular analysis for multiple pathogens in aqueous humor would be needed in order to determine if some etiologies were dismissed.