Prevalence, demographics, and clinical characteristics of Latin American patients with spondyloarthritis

Large epidemiologic and clinical estimates of spondyloarthritis (SpA) in Latin America are not available. In this narrative review, our goal was to descriptively summarize the prevalence and features of SpA in Latin America, based on available small studies. A review of peer-reviewed literature identified 41 relevant publications. Of these, 11 (mostly based on Mexican data) estimated the prevalence of SpA and its subtypes, which varied from 0.28 to 0.9% (SpA), 0.02 to 0.8% (ankylosing spondylitis), 0.2 to 0.9% (axial SpA), and 0.004 to 0.08% (psoriatic arthritis). Demographic and/or clinical characteristics were reported in 31 of the 41 publications, deriving data from 3 multinational studies, as well as individual studies from Argentina, Brazil, Chile, Colombia, Costa Rica, Mexico, Peru, Uruguay, and Venezuela. Data relating to treatment, disease manifestations (articular and extra-articular), and comorbidities were summarized across the countries. Available data suggest that there is a variability in prevalence, manifestations, and comorbidities of SpA across Latin America. Basic epidemiologic and clinical data are required from several countries not currently represented. Data relating to current treatment approaches, patient outcomes, and socioeconomic impact within this large geographic region are also needed.


Introduction
Spondyloarthritis (SpA) comprises a group of inflammatory diseases of the sacroiliac and spinal joints. The Assessment in Spondyloarthritis International Society (ASAS) classification criteria [1] defines the clinical manifestations of SpA as axial, peripheral, and extraarticular. The group of interrelated diseases considered subtypes of SpA include ankylosing spondylitis (AS), psoriatic arthritis (PsA), reactive arthritis (ReA), inflammatory bowel disease (IBD)-related SpA (IBD-SpA), and undifferentiated SpA (uSpA) [1,2]. Juvenile-onset SpA (JSpA) also falls under the same umbrella categorization [1], but as JSpA has its own distinct set of challenges [3], it is out of the scope of this review. These diseases are all associated with the presence of the HLA-B27 antigen and share a number of features, including inflammatory back pain, peripheral arthritis, enthesitis, and dactylitis. Other shared features include uveitis, psoriasis, and IBD [4].
In 2016, the regional pooled prevalence of SpA around the world was estimated to range from 0.20% (95% confidence interval [CI]: 0.00-0.66) in South-East Asia to 1.61% (95% CI: 1.27-2.00) in the Northern Arctic communities [2]. In South/Latin America, the prevalence of SpA was estimated at 0.52% (95% CI: 0.10-1.25), which was similar to Europe (0.54%; 95% CI: 0.36-0.78) [2]. However, the estimates for South/Latin America were based on a limited number of available data for SpA, AS, and PsA (2,4, and 4 studies, respectively) [2]. Several of those studies were based on data from the 2 large Latin American SpA registries: Registro Ibero-Americano de Espondiloartritis (RESPONDIA, The Ibero-American Registry of Spondyloarthritis) and Registro Brasileiro de Espondiloartrites (RBE, Brazilian Registry of Spondyloarthritis). RESPONDIA represents the largest series of patients with SpA in the Americas (> 100 university centers from 8 Latin American countries [Argentina, Brazil, Chile, Costa Rica, Ecuador, Mexico, Peru, Uruguay, and Venezuela], plus Portugal and Spain), and the RBE covers the 5 Brazilian geographic macro-regions [5].
The objective of this narrative review is to summarize the available SpA prevalence data from Latin America. Specifically, we aimed to assess the demographic and clinical characteristics of Latin American patients with SpA, and to identify major gaps in the literature that could be the subject of further investigation.

Methods
A search of the literature from 2008 onwards was conducted on June 6, 2019. PubMed, the ACR and EULAR abstract archives, and the online PANLAR abstracts (from the 2018 and 2016 meetings), were searched using pre-defined search terms. PubMed searches used the thread: (spondyloarthritis OR spondyloarthropath* OR ankylosing spondylitis OR psoriatic arthritis) AND (latin america* OR argentin* OR bolivia* OR brazil* OR chile* OR colombia* OR costa rica* OR cuba* OR dominica* OR dominican republic OR ecuador* OR guatemala* OR hondura* OR mexic* OR nicaragua* OR panama* OR peru* OR urugua* OR venezuela*) AND (population OR prevalen* OR epidemiolog* OR inciden*). This search was repeated using the Latin American and Caribbean Health Sciences Literature (LILACS) platform. Online abstract archives and published abstracts were searched using the terms "population", "prevalen*" "epidemiolog*" and "inciden*" and any abstract titles containing these terms were hand-searched for relevant data (disease and geographical location).
Publications were excluded for the following reasons: 1) no data on either prevalence of SpA in Latin America, or patient demographics, or only provided for subpopulations (e.g., male vs. female); 2) not including patients with non-radiographic axSpA (nr-axSpA), AS, or PsA; 3) unidentified patient nationality (author affiliation was not taken as indicative of patient nationality); 4) SpA was not primary diagnosis, or was reported in combination with other inflammatory conditions (e.g., rheumatoid arthritis); 5) focused exclusively on JSpA; 6) combined patients from Latin America with non-Latin American countries as a single group; or 7) review of previously published data. Congress abstracts that were superseded by a journal article were excluded in favor of the article. No language restrictions were imposed. Selected publications were manually searched for data regarding prevalence and incidence of SpA, patient demographics and clinical characteristics, HLA-B27, disease manifestations, comorbidities, patient quality of life and impact of illness on work, and treatments and medication used. Data were summarized and presented descriptively; no quantitative data analysis was performed.

Demographics and clinical characteristics
Publications from Latin American countries with demographic and clinical characteristics fitting the search criteria included Argentina, Brazil, Chile, Colombia, Costa Rica, Mexico, Peru, Uruguay, and Venezuela; the majority were from Argentina and Brazil. There were also 3 multinational studies reporting data from multiple Latin American countries. Of these 31 studies, only 6 encompassed > 1000 patients: 5 were conducted in Brazil [26][27][28][29][30] and the other was a multinational study comparing Latin American patients with those from Europe [45].
In 23 of the studies, the majority (> 50%) of patients were men (Table 2). Mean or median patient age was < 50 years, except in 1 study conducted in Brazil [33], in which the mean patient age was 56 years. In 9 of the 14 studies that recorded race, more than half of the patients (57.7-96%) with SpA were white. At least 1 study from each country reported the proportion of patients with a family history of SpA (Table 2).
Most studies enrolled patients with various SpA subtypes. One study from Argentina [17] enrolled patients with AS only. Three studies from Argentina [20,21,24] and 1 from Brazil [33] enrolled only patients with PsA (Table 2). Of the studies that enrolled patients with multiple SpA subtypes, AS was the most common subtype in most of the studies. Mean age at disease onset ranged from 28 to 40 years (Table 2), which suggests that some studies included patients with JSpA [18, 19, 26, 27, 29, Table 2). The proportions of patients in single-country studies who were HLA-B27-positive ranged from 5% [21] to 71% [45] (Table 3). Only 3 such studies reported < 20% patients as HLA-B27-positive [12,21,24], which likely reflects procedural differences within practice. For example, the HLA-B27 positivity being 4% suggests that the HLA-B27 test was not performed in a representative subset of the study population (the number of patients who were tested for HLA-B27 in that study was not reported) [44]. None of the studies reporting low HLA-B27 rates reported patient race (Table 3). Eleven studies reported mean C-reactive protein levels but most did not specify measurement units (mg/dL or mg/L), which limits data interpretation.
Mean scores for Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and Bath Ankylosing Spondylitis Functional Index (BASFI) ranged from 3.3 to 6.1 and from 3.2 to 5.2, respectively (Table 3). Ankylosing Spondylitis Quality of Life (ASQoL) scores ranged from 6.5 to 8.7 (mean) and 5 to 10 (median). One study reported an ASAS Health Index score of 10 [35].

Comorbidities
Three studies reported SpA comorbidities [22,33,46], while 2 reported a history of infection without specifying the cause [27,37]. The largest of the 3 studies reporting SpA comorbidities (N = 390), which was conducted in Argentina, Colombia, and Mexico, found that the most common comorbidities were arterial hypertension (25%), hypercholesterolemia (22%), osteoporosis (9%), and gastrointestinal ulcer (8%). Arterial hypertension in Colombia and Mexico occurred at a higher rate than was seen in the general population of those countries (21% vs. 13%). Prevalence of tuberculosis (TB), defined as historic or currently active TB, was 3%, which was 10 times that observed in the general population (0.3%). Prevalence of malignancies was 2.8%, which was similar to that observed in the general population (2.6%) [46]. The other 2 studies enrolled approximately 80 patients each. In the Argentinian study [22], the most common comorbidities reported in patients with axSpA were arterial hypertension (27%), gastritis (26%), dyslipidemia (24%), gallstone disease (13%), and nephrolithiasis (12%). Of these, only nephrolithiasis occurred at a higher frequency than that seen in the general population (4%). Mean age and disease duration were significantly greater   [22]. The most common comorbidities in patients with PsA from 2 hospitals in Southern Brazil were metabolic syndrome (54%), arterial hypertension (52%), and diabetes (20%); no comparisons were made with the general population [33].

Quality of life (QoL)
ASQoL or ASAS scores were presented in 9 singlecountry studies ( Table 3). The highest mean/median ASQoL or ASAS score was 11.2 [23] and the lowest was 5 [17,18]. The relationship between SpA status and QoL may have been impacted by factors such as employment status [17], ethnicity [27,31], and treatment status [35], but their reporting was sporadic and interpreting the potential relationship is beyond the scope of this review.

Discussion
This manuscript represents a comprehensive overview of the prevalence and burden of SpA in Latin America. A 2014 systematic review of global AS prevalence found the rate in Latin America to be 10.2 per 10,000 [47]. This calculation was based on 4 studies included here [8][9][10][11]. The additional studies included in our review are generally consistent with that prevalence rate. However, the majority (60%) of the epidemiology studies included in this narrative review were from Mexico, with prevalence data largely unavailable for the majority of countries. Therefore, these data may not be representative of Latin America as a whole. Patient demographics were widely available. Sex and mean age of patients with SpA were consistently reported, whereas race, ethnicity, family history of SpA, age at disease onset, and disease duration were not.
Most patients were male and, where reported, typically had disease onset before the age of 40. These values are consistent with previously reported global estimates [4]. The proportions of patients with AS and PsA varied between countries. Patients included in studies conducted in Argentina were more likely to have PsA than AS, However, the majority of studies in these countries were small (< 100 patients) and differences in the proportion of SpA subtypes may be due to potential differences in patient selection and the diagnostic criteria used, rather than representative of prevalence as a whole.
The majority of studies reporting clinical characteristics were conducted in Argentina and Brazil. One study reported HLA-B27 status by ethnicity [27] and 1 by age [28] (both Brazil). HLA-B27-positive status was more common in white than in black patients [28], and in older than younger patients [28], consistent with previously reported global estimates [4]. HLA-B27 is not common among indigenous communities in South America [48], but no studies Consistent with the findings of the ASAS-COMOSPA study of patients from 22 countries [49], the prevalence ranges of disease manifestations reported here showed considerable variability. Peripheral articular involvement and enthesitis were similar in range between ASAS-COMOSPA and our study (3-82% vs. 12-73%, and 6-72% vs. 8-79%, respectively), whereas the study conducted in Latin America and the European Union (EU) reported higher prevalence of peripheral arthritis and enthesitis in Latin American than in EU patients [45]. We found a wider range of reported joint involvement (9-100%) and dactylitis (8-52%) than ASAS-COMOSPA (52-100% and 0-35%, respectively), whereas the range of reported extra-articular manifestations was generally lower in our study than in the ASAS-COMOSPA study.
Overall, there was a lack of data on impact of disease variables and comorbidities in Latin American patients with SpA. One study investigated the impact of variables such as disease duration or activity on QoL [17]. Other studies explored the relationship between QoL and employment status, treatment status [17,35], or patients' age or age at disease onset [28,45], but there were too few to make any meaningful conclusions. Only 3 studies presented data on comorbidities. Conditions such as osteoporosis, cardiovascular disease, cancer, and infections are key comorbidities in patients with SpA [50], yet only 1 study presented frequencies of malignancies and infections (TB) [46]. Treatment data were available from almost half of the studies reviewed, but for most countries, information was available from a single study only. The range of NSAID use across studies was largely comparable with that previously reported in the ASAS-COMOSPA study (32-98% vs. 46-98%, respectively). MTX use was higher in our review (19-67% vs. 2-51%), but SSZ (9-87% vs. 0-90%) and TNFi (5-45% vs. 44%) use was typically   [49]. In the comparison study of Latin America and Europe, Latin American patients with AS had a significantly higher use of NSAIDs, CS, MTX, SSZ, and MTX-TNFi combination therapy than their EU counterparts (P < 0.001 for all comparisons); there was no significant difference in the use of TNFi monotherapy [45].
Although the literature surveyed was relatively recent (the past 10 years), most of the estimates of treatment data date back to the beginning of that period. Since then, most Latin American countries have adopted regulatory standards for biosimilars [51] and have allowed biosimilars into their markets, and new therapies are constantly being developed. Therefore, treatment options for patients with SpA in this region have changed and the estimates summarized here may not be representative of the current situation.
This review had some limitations. Overall, we found that, outside of the 8 RESPONDIA papers, study methodology and design varied widely, and the types of data collected (prevalence, disease manifestations, comorbidities, QoL, treatments) were not consistent between studies. Several areas covered had limited data. The possibility of assessing the true extent of disease manifestations was limited due to a lack of clarity regarding data collection. Fewer than one-quarter of studies listed disease manifestations that were reported by none of the enrolled patients [34,37,39,40,42,44]. For other studies, it was not obvious if unreported disease manifestations were not found or if the data were not collected. Several treatment-related areas were not covered. For example, patient views on medication, including patient satisfaction with their treatment, were missing. Treatment adherence was not reported in any of the studies, even though poor adherence to treatments for inflammatory diseases is associated with poorer clinical outcomes [52]. Finally, study inclusion criteria for this review were specific to prevalence data or patient demographics. Consequently, studies with a sole focus on other aspects (e.g., disease manifestations) may not have been identified.
In summary, our review identified several gaps in the literature on SpA from Latin America. Prevalence studies outside Mexico are generally lacking. More than half the studies providing data on clinical characteristics came from Argentina and Brazil. There are insufficient data on comorbidities and QoL, particularly from countries other than Argentina and Brazil. Finally, data on treatment patterns are mostly out of date. Up-to-date reports on treatment adherence or patient views would add clarification to the literature. Future research efforts in this region should also focus on the need for standardized care, follow-up, and regular monitoring of disease, as well as characterizing the side effects of medications used for treatment.

Conclusions
A total of 41 publications relating to the prevalence and characteristics of SpA in Latin America were identified. Several gaps were evident in terms of the countries and data covered, particularly in terms of the treatment patterns and impact of SpA. More consistent methodologies between studies would help provide a more accurate picture of the disease burden in this region.