Chlamydiae as etiologic agents in chronic undifferentiated spondylarthritis

Abstract

Objective

The majority of patients with Chlamydia-induced reactive arthritis do not present with the classic triad of arthritis, conjunctivitis/iritis, and urethritis. Moreover, acute chlamydial infections are often asymptomatic. The aim of the present study was to assess the prevalence of synovial Chlamydia trachomatis and Chlamydia pneumoniae infections in patients with chronic undifferentiated spondylarthritis (uSpA).

Methods

Study patients met the European Spondylarthropathy Study Group criteria for SpA, without evidence of ankylosing spondylitis, psoriasis, inflammatory bowel disease, or preceding dysentery. Symptoms were present for ≥6 months. Each patient underwent a synovial biopsy; tissue and concomitantly obtained peripheral blood mononuclear cells (PBMCs) were analyzed by polymerase chain reaction (PCR) for C trachomatis and C pneumoniae DNA. Other data collected on the day of the biopsy included standard demographic information and medical history, including any known history of C trachomatis or C pneumoniae. Physical examination (including joint count, evaluation for dactylitis and/or enthesitis, and skin examination) and HLA–B27 typing were performed. Synovial tissue (ST) samples from 167 patients with osteoarthritis (OA) were used as controls.

Results

Twenty-six patients met the entry criteria and underwent synovial biopsy (25 knee, 1 wrist). Sixteen of them (62%) were positive for C trachomatis and/or C pneumoniae DNA (10 for C trachomatis, 4 for C pneumoniae, and 2 for both). PCR analysis of ST revealed the presence of Chlamydia significantly more frequently in patients with uSpA than in OA controls (P < 0.0001). No specific clinical characteristics differentiated Chlamydia-positive from Chlamydia-negative patients. PBMCs from 4 of the 26 uSpA patients (15%) were positive for Chlamydia, and Chlamydia was found in ST from 2 of these 4 patients. No significant correlation between PCR positivity and HLA–B27 positivity was found.

Conclusion

The frequency of Chlamydia-positive ST samples, as determined by PCR, was found to be significantly higher in patients with uSpA than in patients with OA. Our results suggest that in many patients with uSpA, chlamydial infection, which is often occult, may be the cause.

In 1978, a group of rheumatoid factor–negative inflammatory arthritides was recognized as a unified entity and collectively termed seronegative polyarthritis (1). Now known as the spondylarthritides (SpA), these arthritides share clinical and radiographic features and include ankylosing spondylitis (AS), psoriatic arthritis, inflammatory bowel disease–related arthritis, reactive arthritis (ReA), and undifferentiated SpA (uSpA). The term “uSpA” is used to designate patients with clinical and radiographic features consistent with SpA who do not fulfill the classification criteria for any of the established disease categories.

In 1991, the European Spondylarthropathy Study Group (ESSG) criteria were developed in order to establish diagnostic criteria for the spondylarthritides (2). These criteria include both clinical and radiographic features and have been demonstrated to be 87% sensitive and specific for the diagnosis of SpA. They are somewhat less well-suited for the diagnosis of disease in its early stages but are extremely useful for recognizing established disease. In 1995, the Amor criteria for SpA were developed (3). These criteria include many of the same clinical and radiographic features as the ESSG criteria, but also include the HLA–B27 status.

The prevalence of SpA in white populations is estimated to be 0.5–2.0% (4). The true prevalence of uSpA is difficult to assess quantitatively in these or in other groups because of a lack of specific diagnostic criteria and because the disease often goes unrecognized. However, AS and uSpA are considered to be the most common types of SpA (5, 6). Indeed, a large study assessing the prevalence of SpA in Germany demonstrated an overall prevalence of 1.9%, with AS being the most common type, followed by uSpA (0.86% and 0.67%, respectively) (7). Another report indicated that uSpA was more than twice as common as AS, with 40% of the subjects studied having the former condition (8). Importantly with regard to the present study, uSpA may well be the most underdiagnosed type of SpA. For example, a study in Spain evaluated 514 patients with anterior uveitis and showed that 53% of these patients were not diagnosed as having an SpA until after their first episode of uveitis; for those patients with an eventual diagnosis of uSpA, 91% remained undiagnosed until after the first episode of uveitis (9).

Some investigators have argued that the term uSpA should be viewed solely as a working label (10). The onus remains on the clinicians to make the correct diagnosis. This is often done by long-term followup, which can take years, and during that time, the patient often develops disease sequelae before a definitive diagnosis can be made. In contrast, identification of a causative organism, if present, could allow for an immediate diagnosis.

It has been argued that uSpA is a forme fruste of ReA (11). Chlamydia trachomatis is the most common etiologic agent in ReA in the US (12, 13), and the initial infection is often asymptomatic (14, 15). A number of published reports also indicate that Chlamydia pneumoniae, a related respiratory pathogen, is another causative agent in ReA, albeit at a lower frequency than C trachomatis (16, 17). The aim of the present study was to determine the percentage of patients with uSpA whose disease may be a result of infection with C trachomatis and/or C pneumoniae.

PATIENTS AND METHODS

Subjects and controls.

We assessed the prevalence of synovial C trachomatis and C pneumoniae in patients who met the ESSG criteria for SpA. Individuals were recruited from May 2006 through December 2007 as part of an ongoing clinical trial to assess a novel therapy for chronic Chlamydia-induced ReA (ClinicalTrials.gov identifier: NCT00351273). Clinical study sites were the University of South Florida College of Medicine, the Louisiana State University Health Sciences Center, and the University of Toronto (Toronto, Ontario, Canada). Patients were enrolled in the trial if they had chronic uSpA or classic Chlamydia-induced ReA. All patients had to meet the ESSG criteria for SpA (2) without evidence of AS, psoriasis, inflammatory bowel disease, or preceding dysentery.

Men and women at least 18 years old whose symptoms had been present for ≥6 months were eligible to participate. Patients also had to meet the following inclusion criteria: 1) inflammatory spinal pain (defined as spinal pain in the back, dorsal, or cervical region, with at least 4 of the following: [a] onset before the age of 45 years, [b] insidious onset, [c] improved by exercise, [d] associated with morning stiffness, or [e] at least 3 months' duration) or 2) synovitis, and 1 or more of the following: 3) positive family history (defined as the presence in first-degree or second-degree relatives of any of the following: [a] AS, [b] psoriasis, [c] acute uveitis, [d] ReA, or [e] inflammatory bowel disease), 4) urethritis or cervicitis within 1 month before arthritis, 5) alternating buttock pain, 6) enthesopathy (defined as past or present spontaneous pain or tenderness on examination of the site of insertion of the Achilles tendon or plantar fascia), or 7) sacroiliitis (defined as at least radiographic grade 2 unilateral, where 0 = normal, 1 = possible, 2 = minimal, 3 = moderate, and 4 = ankylosis). Exclusion criteria included use of anticoagulants, current pregnancy, and previous prolonged exposure to antibiotics (>2 weeks) as a specific therapy for possible ReA.

Frozen synovial tissue (ST) samples stored at the institution of one of the authors (APH) were used as controls for this study. These samples were from patients with osteoarthritis (OA) and had been obtained by needle biopsy using the Parker-Pearson method (18) or at the time of joint replacement surgery. No peripheral blood mononuclear cells (PBMCs) were available from the OA patients. The primary outcome measure of the study was Chlamydia positivity in ST, as determined by polymerase chain reaction (PCR). PCR findings in PBMCs also are reported.

Sample procurement and handling.

One blood sample (5 ml) was obtained from each study subject and shipped at ambient temperature via overnight courier from the clinical sites to the laboratory. From patients with active synovitis who provided consent, ST was obtained at the same time as the blood sample, by synovial biopsy using a Parker-Pearson needle (18). The site of the biopsy was the knee in all patients, with the exception of 1 sample obtained from the wrist by open surgical procedure. Three or 4 tissue samples were obtained under sterile conditions from the suprapatellar pouch in each patient (except for the patient whose biopsy site was the wrist). These samples were immediately snap-frozen in liquid nitrogen and stored at −80°C until they were shipped on dry ice, again using an overnight courier, to the laboratory.

Data collection.

On the day of synovial biopsy, data collected from the patients included standard demographic information and medical history, including any known previous exposure to C trachomatis or C pneumoniae. If patients reported such an exposure, the timing of that infection was documented in relation to the onset of their uSpA. In addition, the blood sample was collected, and physical examination and HLA–B27 typing were performed. The physical examination included swollen and tender joint counts (66 and 68 joints, respectively), evaluation for dactylitis and active enthesitis (Achilles tendinitis or plantar fasciitis), and a skin examination (specifically for keratoderma blennorrhagicum and circinate balanitis).

PCR analysis.

The use of PCR assays to assess chlamydial DNA in ST or in other samples has been reported for both C trachomatis and C pneumoniae, and PCR was performed as described previously (19–25). Screening PCR assays were performed in duplicate by each of 2 workers. Standard analyses included the targeting of at least 2 different DNA sequences from each organism. The C trachomatis–directed assays targeted omp1 and the 16S ribosomal RNA (rRNA) genes. Assays to assess the presence of C pneumoniae DNA targeted the homologous genes in that organism (C trachomatis–directed primer systems do not amplify sequences in C pneumoniae, and vice versa). Samples were considered positive if a positive result was obtained in both duplicates of all assays. In the case of discrepancies, a third set of assays targeting any of several other genes was used for resolution. Extreme care was taken to avoid contamination of PCR-related materials.

Statistical analysis.

Patients with uSpA whose ST was positive for Chlamydia were compared with patients with OA by Fisher's exact test. P values less than or equal to 0.05 were considered significant.

RESULTS

A total of 80 study subjects were screened as part of the clinical trial of a novel therapy for Chlamydia-induced ReA. Of these 80, 26 patients met the entry criteria for the current study and underwent synovial biopsy. These ST samples were collected at the University of South Florida College of Medicine and Louisiana State University Health Sciences Center. Demographic and clinical information on the 26 patients (16 men and 10 women) is summarized in Table 1. The average age was 48.4 years (range 22–70 years), and the average disease duration was 9.1 years (range 0.5–36 years). The biopsy site was the knee in 25 patients and the wrist in 1. All biopsied joints had active synovitis on the day of ST procurement. The control group for this study comprised 167 patients with OA (Table 1).

Table 1. Characteristics of the uSpA patients and OA controls*

	uSpA patients (n = 26)	OA controls (n = 167)
* uSpA = undifferentiated spondylarthritis; OA = osteoarthritis. † Data were available for 70 controls. ‡ Data were available for 25 patients.
Age, mean (range) years	48.4 (22–70)	43.3 (21–74)
Sex, no. female/male	10/16	72/95
Race, no.
White	19	125
Black	6	25
Hispanic	1	7
Asian	0	5
Not reported	0	5
Disease duration, mean (range) years†	9.1 (0.5–36)	1.1 (0.1–2)
HLA–B27 positivity, no. (%)‡	4 (16)	–

As indicated above, all 26 study subjects met the ESSG criteria, and all had a disease duration of ≥6 months. Specific clinical characteristics of the patients on the day of synovial biopsy are summarized in Table 2. The mean score for the physician's global assessment and the patient's global assessment for all 26 subjects was 57.7 (0–100 scale, with 0 being the best and 100 being the worst) and 41.6 (0 being the worst and 100 being the best), respectively. In 16 of the 26 patients (62%), radiography of the sacroiliac joints was performed at the time of study entry. Fourteen of these 16 (88%) had asymmetric sacroiliitis of grade 2 or 3 (no subject had grade 4 radiographic sacroiliitis or symmetric sacroiliitis). Seven patients were taking a disease-modifying antirheumatic drug for their SpA at the time of biopsy (sulfasalazine in 3, methotrexate in 2, and hydroxychloroquine in 2).

Table 2. Clinical characteristics of the patients with undifferentiated spondylarthritis (uSpA)*

* Except where indicated otherwise, values are the number (%). † Defined according to the European Spondylarthropathy Study Group criteria. ‡ Defined as Achilles tendinitis and/or plantar fasciitis. § Date were available for 16 male patients.
Axial arthritis†	19 (73)
Peripheral arthritis	26 (100)
Swollen joint count, mean (range)	3.3 (1–8)
Tender joint count, mean (range)	7.2 (1–26)
Active enthesitis‡	15 (58)
Active dactylitis	7 (27)
History of uveitis	2 (8)
Active uveitis	1 (4)
History of keratoderma blennorrhagicum	6 (23)
Active keratoderma blennorrhagicum	4 (15)
History of circinate balanitis§	2 (13)
Active circinate balanitis	1 (6)
Active urethritis	1 (4)
Health Assessment Questionnaire, mean (range) score	0.89 (0–1.63)
Low back morning stiffness, mean (range) hours	1.1 (0–2.5)
History of Chlamydia trachomatis at any time	11 (42)
History of C trachomatis within 6 weeks of uSpA onset	2 (8)
History of Chlamydia pneumoniae	0 (0)

PCR analysis revealed that 16 of the 26 study subjects (62%) were positive for either C trachomatis or C pneumoniae DNA, or both, in multiple assays using ST nucleic acid preparations (Table 3). Of the biopsies from the 16 patients who tested positive for Chlamydia, 10 (63%) were positive for C trachomatis only, 4 (25%) were positive for C pneumoniae only, and 2 (13%) were positive for both. Patients with uSpA were significantly more likely than OA controls to be positive for Chlamydia (20 of 167 controls [12%]; P < 0.0001). Interestingly, only 2 of the 26 uSpA patients (8%) exhibited clinical symptoms of a possible C trachomatis infection within 6 weeks before the onset of ReA symptoms. Because we had neither documented proof of the original C trachomatis infection nor evidence of the actual timing of the infection, these 2 patients were not originally classified as having Chlamydia-induced ReA. Seven of the 10 patients (70%) whose ST was positive for C trachomatis only had no known clinical history of infection with the organism at any time previously, and no patient had a known clinical history of C pneumoniae infection. None of the specific clinical characteristics of these study subjects (Table 2), including a history of C trachomatis at any time, and none of the medications they were taking significantly correlated with Chlamydia positivity (P = 0.25). Both of the patients who had a history of C trachomatis infection within the 6 weeks prior to the onset of uSpA tested positive for C trachomatis, but the numbers were too low to determine if this was a significant correlation.

Table 3. Synovial tissue positivity for Chlamydia trachomatis or Chlamydia pneumoniae in the uSpA patients and OA controls*

	uSpA patients (n = 26)	OA controls (n = 167)	P
* Values are the number (%) positive. uSpA = undifferentiated spondylarthritis; OA = osteoarthritis.
Chlamydia	16 (62)	20 (12)	<0.0001
C trachomatis	10 (38)	19 (11)	0.001
C pneumoniae	4 (15)	0 (0)	0.0003
Both C trachomatis and C pneumoniae	2 (8)	1 (0.6)	0.048

As stated above, each ST sample was analyzed in duplicate by 2 observers using PCR to test for the presence of 2 genes, specifically, omp1 and 16S rRNA to assay for C trachomatis and 2 homologous genes to assay for C pneumoniae. This resulted in 8 assays per ST sample for each organism. Only samples that were positive for Chlamydia in all 8 assays were considered to be true positive. In no uSpA or OA patient was there a positive result in only some of the PCR assays (i.e., positive results in <8 of the 8 assays). Although 3–4 ST samples were obtained from each patient, we did not assay more than a single tissue sample per patient in order to conserve the samples for future studies.

The disease duration was known for only 70 of 167 OA patients (Table 1). In many of these patients, the ST sample was collected at the time of arthroscopy that was performed for routine clinical care. OA disease inception was defined as the time of initial diagnosis by a physician. These factors contributed to the short disease duration and relatively young age of the OA patients.

PBMCs from only 4 of the 26 patients (15%) were positive for chlamydiae (3 C trachomatis and 1 C pneumoniae). ST samples from 2 of these 4 patients were also positive for chlamydiae (1 C trachomatis and 1 C pneumoniae), and PBMCs from both of these patients were positive for C trachomatis. Four of 25 patients (16%; test not performed on 1 subject) were HLA–B27 positive. ST samples from all 4 of these patients were positive for chlamydiae (3 C trachomatis and 1 C pneumoniae), but there was not a significant correlation (P = 0.26).

DISCUSSION

Undifferentiated SpA is one of the most common forms of SpA, and some data suggest that it is the most common presentation of all spondylarthritides (8). Chlamydial infections are the most common cause of ReA (12, 13), although the initial infection frequently can be asymptomatic (14, 15). Importantly with regard to the present study, uSpA has been argued to be a forme fruste of ReA (11). The data presented here demonstrate that the rate of Chlamydia positivity in the ST of patients with uSpA (62%) was significantly higher than the rate in OA controls (12%), suggesting that chlamydial infection, which is often occult, may be the cause of uSpA in many patients. Background C trachomatis positivity rates of ∼5–20% in synovial samples have been reported (26, 27), and urogenital cultures from patients with suspected Chlamydia-induced ReA have higher rates of positivity for C trachomatis compared with patients with other types of SpA (28).

New genital infections with C trachomatis must be reported to the Centers for Disease Control and Prevention (CDC). The CDC has estimated that as many as 3 million new cases per year occur in the US, with as many as 4–6 million cases active at any one time (29, 30). Some reports have indicated that ∼5% of patients develop objective features consistent with ReA after C trachomatis or nongonococcal infection (31). If ReA develops in 5% of the 3 million new cases of chlamydial infection in a given year, as many as 150,000 cases of acute Chlamydia-induced ReA would appear in the US each year, which is a low estimate, since it does not include cases that result from C pneumoniae. Of these 150,000 cases, ∼30–50% would progress to chronicity, with one report suggesting that 63% of patients would experience chronic symptoms (32). For comparison, the estimated annual incidence of rheumatoid arthritis (RA) in the US is 44.6 cases per 100,000 people (33). If the population is ∼281 million (according to the 2000 US Census), then ∼125,000 new cases of RA occur every year. A 2002 study in Sweden showed the annual incidence of ReA to be higher than that of RA (34). Thus, Chlamydia-induced ReA represents a considerable burden on the health care systems of the US and other nations, and its impact on those systems may well be significantly underrecognized.

The pathologic sequelae of chlamydial infections can be severe, which is an issue of importance, especially in women, in whom initial genital infections with C trachomatis are often asymptomatic. Interestingly, in one study, 78% of subjects who developed ReA features after C trachomatis or nongonococcal infections had an asymptomatic initial infection (31). This mirrors the data reported here, that in 14 of 16 patients (88%) with an ST assay revealing positivity for C trachomatis, the initial infection was asymptomatic. For C pneumoniae, as many as 70% of acute infections are asymptomatic (35, 36); even when a patient is symptomatic from an acute C pneumoniae exposure, definitive identification of the organism is rare. Thus, relying on identification of a symptomatic infection may result in routine underdiagnosis, or misdiagnosis, of Chlamydia-induced ReA.

Because ReA is a type of SpA and because the majority of patients with ReA do not present with the classic triad of symptoms (13), it is reasonable to contend that C trachomatis has an etiologic role in uSpA. Although there is no pathognomonic diagnostic test for Chlamydia-induced ReA, using PCR to reveal the presence of chlamydial DNA in ST samples from patients who fulfill the clinical criteria for ReA represents the most accurate means of diagnosing the condition. The contention that ST analysis yields the most accurate results is supported by the fact that PBMCs from the majority of subjects studied here (85%) tested negative for Chlamydia, and only 2 of 16 patients (12.5%) whose ST tested positive for C trachomatis also had C trachomatis in their PBMCs.

Although there were significantly fewer OA patients with chlamydiae positivity, ST from a small percentage of these patients (12%) did harbor this organism. As stated above, similar findings have been reported in other OA patients (27) and even in a very small percentage of asymptomatic volunteers (26). This highlights the importance of host genetic variability and host tolerance. Various hosts might respond differently to the same pathogen. Furthermore, in the case of C trachomatis, there are several different serovars, and they may portend diverse prognoses that include variable pathogenic sequelae. This study did not address these important questions.

Only 4 of the 25 study subjects (16%) who underwent HLA–B27 typing and only 4 of the 16 (25%) whose ST was positive for Chlamydia were HLA–B27 positive. These numbers are lower than the previously reported overall prevalence of HLA–B27 in patients with ReA. However, the majority of epidemiologic studies in ReA suggest that the background prevalence of HLA–B27 is 30–50% (13), rather than the higher prevalences frequently reported (37). Furthermore, HLA–B27–negative patients who develop Chlamydia-induced ReA may display a less fulminant disease course or may be less likely to manifest the complete triad of symptoms. The fact that the majority of these subjects in the present study were HLA–B27 negative might explain the presence of fewer of the “classic” symptoms of ReA. Regardless, the data presented here suggest that chlamydial infections, which are often occult, are etiologic in many cases of uSpA.

AUTHOR CONTRIBUTIONS

Dr. Carter had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Carter, Gérard, Hudson.

Acquisition of data. Carter, Gérard, Espinoza, Ricca, Valeriano, Snelgrove, Oszust, Vasey, Hudson.

Analysis and interpretation of data. Carter, Hudson.

Manuscript preparation. Carter, Valeriano, Vasey, Hudson.

Statistical analysis. Carter.