Undifferentiated Spondyloarthritis: A Longterm Followup

MATERIALS AND METHODS

Our study was designed in January 1994 and intended to investigate the longterm outcome of consecutive adult patients with the diagnosis of uSpA who attended the spondyloarthritis outpatient clinic of the Hospital de Clínicas of the University of Campinas, a tertiary referral university hospital in Brazil. Initially, adult patients not fulfilling the criteria for a definite diagnosis of any of the SpA at the initial visit and who fulfilled the ESSG⁷ criteria for SpA were included. Subsequently, it was also necessary to fulfill at least 6 points in the Amor⁶ set of criteria to be considered a patient with uSpA.

A total of 156 consecutive adult patients were referred to our institution between January 1994 and December 2002 with the diagnosis of uSpA. Forty-five patients had to be excluded from our study because they fulfilled 4 or 5 points in the Amor set of criteria or had a diagnosis of a definite SpA. So 111 patients with uSpA were included in our study.

Ethnicity was divided into whites and nonwhites; the latter included African Brazilians (black patients of unmixed ancestry and mulattoes, i.e., originating from the mixture of white and black populations) and East Asians (Japanese origin). Most patients came from the metropolitan area of Campinas, with a population of roughly 3 million, located in an industrial region of the state of São Paulo in southeast Brazil, where the white population includes people of predominantly Portuguese, Italian, and/or Spanish ancestry.

Clinical evaluation included a search for initial symptoms and the presence of axial and peripheral joint involvement, and heel enthesopathies. Initial symptoms included ILBP, asymmetric oligoarthritis (predominantly affecting large joints of the lower limbs), and heel enthesopathies (Achilles tendinitis and/or plantar fasciitis). ILBP was considered according to established criteria¹⁴.

Imaging methods performed at study entry included pelvic and calcaneal radiography. After 2, 5, 7, and 10 years of followup, another pelvic radiography examination was performed. Each sacroiliac joint was scored according to the modified New York criteria¹⁵; to be enrolled in the study, patients had to present sacroiliitis graded as normal or below grade II bilateral or below grade III unilateral. The radiographs were independently analyzed by 1 rheumatologist and 1 blinded radiologist (who changed 3 times during the 15-year study period). When there were different interpretations of a radiograph, the 2 analysts conferred and agreed on a diagnosis.

After 2, 5, 7, and 10-year followup, patients who fulfilled the modified New York criteria¹⁵ were diagnosed as having AS and patients who presented with characteristic psoriatic skin lesions were diagnosed as having PsA. Disease remission was defined as a complete remission of clinical symptoms for at least 1 year (without symptomatic medication).

HLA-B alleles were typed at low resolution using polymerase chain reaction-amplified DNA hybridized with sequence-specific oligonucleotide primers (Dynal Biotech Ltd., Merseyside, UK).

Statistical analysis

The chi-squared and Fisher’s exact tests were used to verify the association of the variables of interest; p values ≤ 0.05 were considered to be significant, and 0.10 ≤ p < 0.05 were considered to be a statistical trend. To verify the influence of risk factors for progression to AS (compared to the patients who progressed to remission) in the 5, 7, and 10-year outcome of the patients with uSpA, univariate and multivariate logistic regression analyses were used.

RESULTS

There were 90 men (81.1%) and 21 women (18.9%) patients, with a male:female ratio of 4.3:1. There were 87 whites (78.4%), 22 African Brazilians (19.8%), and 2 East Asians (1.8%) of Japanese origin; the African Brazilian group included only mulattoes, with no patient presenting pure black ancestry. The mean age at onset was 27.2 years (SD ± 9.5 yrs), and the mean age at diagnosis was 29.9 years (SD ± 9.6 yrs); the mean time to diagnosis was 2.7 years. HLA-B27 was positive in 68 (61.3%) patients, with 13 (11.7%) of them having a family history of SpA, predominantly AS.

The most frequent initial symptoms were peripheral arthritis (45.9%), ILBP (28.8%), enthesitis (15.3%), and buttock pain (8.1%). Patients’ characteristics at the initial evaluation and the joint manifestations shown during the longterm followup are described in Table 1.

Peripheral involvement was common during the study period, affecting 93 (83.8%) patients. The most frequent peripheral joints involved were ankles (79.3%), knees (58.6%), wrists (11.7%), and elbows (6.3%); involvement of small joints in upper and lower limbs was infrequent. During the followup, 67 (60.4%) patients had enthesopathy; the most frequent entheses involved the Achilles tendon (43.2%) and plantar fascia (35.1%). Sausage fingers were observed in 8 patients (7.2%).

During the study period, 75 patients (67.6%) had ILBP, 59 (53.2%) had buttock pain, and 12 (10.8%) had cervical inflammatory pain in the followup. Involvement of the girdle joints was noted in 37 patients (33.3%), mainly affecting hip (16.2%), shoulder (10.8%), and sternoclavicular joints (9%).

Twenty-four patients (21.6%) presented grade I sacroiliitis at the initial radiograph. Fourteen patients presented unilateral and 10 bilateral grade I sacroiliitis; no patient presented grades II, III, or IV sacroiliitis (unilateral or bilateral) at the initial radiograph.

All the patients took nonsteroidal antiinflammatory drugs (NSAID), and 32 (28.8%) took glucocorticoids, mainly low doses of prednisone. Methotrexate was prescribed for 30.6% and sulfasalazine for 71.2% of the patients during the study period. In most patients, the followup ended before biologic drugs became regularly available for SpA in Brazil.

All the patients were followed for 2 years, 105 (94.6%) for 5 years, 84 (75.7%) for 7 years, and 42 (37.8%) for 10 years. Two patients died during the followup for reasons not related to SpA. Four patients were considered lost to followup at 5 years, 25 at 7 years, and 67 at 10 years. Twenty-seven patients progressed to AS (24.3%) and 3 to PsA (2.7%), while 25 patients (22.5%) went into remission during the followup. The outcomes at 2, 5, 7, and 10 years are shown in Figure 1. The proportion of patients with progression to AS and to remission was similar at 2, 5, 7, and 10 years. The percentage of patients with the diagnosis of uSpA decreased significantly with progress of the followup, from 82.9% at 2 years to 16.7% at 10 years.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Patient outcomes after 2, 5, 7, and 10 years. N = 111. AS: ankylosing spondylitis; PsA: psoriatic arthritis; REM: remission; USpA: undifferentiated spondyloarthritis.

Univariate logistic regression analysis revealed that ethnicity, HLA-B27, buttock pain, inflammatory low back pain, ankle involvement, sacroiliitis grade I at the beginning of the study, and the use of sulfasalazine were statistically associated with progression to AS, compared to the patients who went into remission during the study period (Table 2).

Multivariate logistic regression analysis revealed that HLA-B27 (p = 0.035, OR 6.720, 95% CI 11.45–39.43) and buttock pain (p = 0.009, OR 6.211, 95% CI 1.591–24.25) were statistically associated with progression to AS, compared to the patients who went into remission (Table 3).

DISCUSSION

Our study confirms that uSpA can be considered a provisional diagnosis in a significant number of patients, especially in those with longterm followup, and that AS is the main outcome variable in the SpA group. Also important is the association of the axial and peripheral components in the characterization of the uSpA diagnosis in a heterogeneous population such as that in Brazil. A recent publication of the Brazilian Registry of Spondyloarthritis, analyzing 1036 patients who submitted to a common protocol of investigation in 28 university centers, showed that pure axial disease was found in 36.7% of the patients and that pure peripheral involvement was observed in 10.7%, while the mixed pattern (axial, peripheral, and entheseal) was observed in 47.9% of the patients; uSpA was diagnosed in 6.3% of these patients¹⁶.

Among the classification criteria for SpA, the ESSG criteria⁷ are the most widely used, with a sensitivity of 87% and a specificity of 87%; they are practical and easy to apply. The Amor criteria⁶ are less frequently used in daily practice by rheumatologists. Although both sets of criteria are efficient to classify SpA, we demonstrated that fulfilling the 2 sets can be important to better characterize the population under study with uSpA. Collantes, et al, analyzing 102 Spanish patients with uSpA in a 5-year followup, found that the Amor and ESSG criteria both allowed early identification of uSpA, but that the Amor criteria performed better¹⁷.

The selection of patients with ILBP, asymmetric oligoarthritis (affecting predominantly large joints in the lower limbs), and enthesopathies (Achilles tendinitis and/or plantar fasciitis) was intended to include only patients with symptomatic SpA, to establish patterns of disease outcome. Peripheral involvement was the most common initial symptom (45.9%), and was found in 83.8% of the patients during the followup. Enthesitis was found in 15.3% as an initial symptom and in 60.4% during the followup. These data emphasize the importance and the high percentage of peripheral symptoms in Brazilian patients with uSpA, as shown in a recent study of 1036 Brazilian patients with SpA¹⁶. Although ILBP was found as an initial symptom in 28.8% of patients and buttock pain in 8.1%, during the longterm followup, ILBP was found in 67.6% and buttock pain in 53.2%. In the short-term and longterm followup, no patient progressed to other diseases not included in the SpA group.

The epidemiologic variables of the patients with uSpA in our study were similar to those observed in patients with AS at our institution¹⁸, with a predominance of white (78.4%) and male (81.1%) patients, with young age at onset (mean 23.7 yrs), and family history of SpA in 11.7%. The HLA-B27 positivity was lower in patients with uSpA (61.3% vs 78.2%), compared to patients with AS. Another Latin American study, from Mexico¹⁹, following up on 62 patients with uSpA, with a predominance of male (52%) and HLA-B27-positive (63.8%) patients, found younger age at onset (mean 20.4 yrs) and positive family history in 16%. A study by Collantes, et al, in Spain¹⁷ observed male predominance (53%), with older age at onset (mean 31.8 yrs) and lower HLA-B27 positivity (20.6%), and with a family history in 8.8%. The study by Mau, et al, in Germany²⁰, analyzing 88 patients in a longterm followup, also presented male predominance (male:female ratio 1.9:1).

Peripheral involvement was an important point in the characterization of the patients with uSpA in our study. Arthritis predominating in large joints in the lower limbs was the most frequent initial symptom (45.9%), with a significant number of patients presenting knee (58.6%) and ankle (78.3%) involvement and enthesopathy (60.4%) during the followup. Analyzing studies involving patients with uSpA, peripheral arthritis was found in 43.1%¹⁷, 82%¹⁹, and 60%²⁰ in various studies.

White ethnicity was statistically associated with progression to AS in the univariate analysis in this study. The Brazilian white population has predominantly European ancestry (from Portugal, Spain, and/or Italy); it is noteworthy that, among the African Brazilians, no pure black patient was diagnosed as having uSpA; mixing among white and black individuals is an important characteristic of the African Brazilian population.

Radiographic grade I sacroiliitis is another variable associated with progression to AS at the univariate analysis. The same results were observed by Huerta-Sil, et al in a Mexican series¹⁹ that showed radiographic sacroiliitis grade < II bilateral or grade < III unilateral with an OR = 11.18, and grade I bilateral with an OR = 12.58. These results emphasize the importance of grade I radiographic sacroiliitis in patients with uSpA. This finding could be considered an early recognition of AS.

Magnetic resonance imaging (MRI) is the most sensitive method to detect early inflammatory changes in sacroiliac joints²¹. The new Assessment of Spondyloarthritis International Society criteria for axial SpA include MRI as an important instrument for the diagnosis of early disease^9,10.

All the patients took NSAID during the followup in our study. Because of the large number of patients with uSpA who had peripheral joint involvement, methotrexate (30.6%) and sulfasalazine (71.2%) were frequently prescribed. Sulfasalazine use presented a statistical association with progression to AS in the univariate analysis.

In a longterm followup, there was a significant and progressive number of patients presenting a diagnosis of AS (24.3%). The Mau series showed that 59% of 54 patients with uSpA in a 10-year followup progressed to AS, while 19% continued with the diagnosis of uSpA²⁰. In an Indian study analyzing the 11-year followup of 22 patients with uSpA, 15 (68%) presented AS, 1 PsA, 2 went into natural remission, and 4 continued with uSpA²². Although the percentage of patients with uSpA who progressed to AS was significantly lower (24.3%) in our study, compared to the German (59%) and the Indian (68%) series, the percentage of patients continuing with uSpA in the longterm followup was similar in the 3 studies (16.7%, 19%, and 16%). In our study, HLA-B27 and buttock pain were associated with AS in the multivariate analysis, compared to the patients who went into remission. Buttock pain was an infrequent initial symptom (8.2%) in the patients with uSpA, but was found in 50.3% of them during the followup, and presented an OR = 6.211 at the multivariate analysis. These results were similar to those observed in the 2-year followup of the initial patients with uSpA observed in our service¹². Progression of uSpA to PsA is rather unusual; in our study, 3 patients progressed to PsA.

Our study had some deficiencies that should be mentioned. First, the percentage of patients lost to followup at 7 years (24.3%) and 10 years (62.2%) does not allow definite conclusions. Second, the fact that the radiographic scoring was not blinded, and the interobserver and intraobserver reliability was not analyzed, places doubt on the accuracy of the diagnoses. Indeed, these are problems of a real-world longterm followup that should be observed in future studies.

Our study showed that AS is the most frequent definite disease diagnosed in the longterm followup of a series of Brazilian patients with uSpA, and that HLA-B27 and buttock pain were the best predictors of progression to AS.