The landscape of reactive arthritis in Colombia (2017–2022): A national registry-based study

Abstract

Introduction/Objectives

Registries allow for ascertaining the epidemiology of chronic diseases such as reactive arthritis (ReA). The Colombian Ministry of Health has implemented a National Health Registry (SISPRO, Sistema Integrado de Información de la Protección Social, by its in Spanish acronym) that collects data from each medical contact in the system, which provides close to universal coverage (around 98%). We aimed to estimate the 6-year prevalence of ReA in Colombia and to describe its demographics, using data from January 1st, 2017 to December 31st, 2022.

Material and methods

We performed an observational, cross-sectional study using the International Classification of Diseases terms associated with ReA based on RIPS (Registros Individuales de Prestación de Servicios de Salud) database on SISPRO and calculated the period prevalence per 100,000 inhabitants.

Results

We found 19,081 patients with a primary diagnosis of ReA, with an estimated 6-year prevalence of 40 cases per 100,000 inhabitants (0.04%). We observed a higher prevalence in women (male-to- female ratio 0.5:1). We found the highest prevalence of cases in the 75–79 years group. Most cases were reported in Atlántico, Quindío, Bolivar, Sucre, and Magdalena. The absolute number of cases was higher in pre-COVID-19 pandemic years (2017: 3630, 2018: 3843, 2019: 4711) when contrasted with COVID-19 pandemic (2020: 2715, 2021: 3021) and post-COVID-19 pandemic (2022: 2451) years.

Conclusion

This is the first study to describe the demographic characteristics of ReA in Colombia. The Caribbean region presents the highest prevalence and elderly patients were most commonly affected, possibly due to misclassification with similar conditions like inflammatory ostearthritis or pseudogout.

Resumen

Introducción/Objetivos

Los registros permiten conocer la epidemiología de enfermedades crónicas como la artritis reactiva (ARe). El Ministerio de Salud de Colombia ha implementado un Sistema Integrado de Información de la Protección Social (SISPRO) que recopila datos de cada contacto médico en el sistema, proporcionando una cobertura cercana a la universal (aproximadamente 98%). Nuestro objetivo fue estimar la prevalencia de 6 años de la ARe en Colombia y describir sus características demográficas, utilizando datos del 1.^o de enero de 2017 al 31 de diciembre de 2022.

Materiales y métodos

Realizamos un estudio observacional transversal utilizando los términos de la Clasificación Internacional de Enfermedades asociados con la ARe basados en los RIPS (Registros Individuales de Prestación de Servicios de Salud) de SISPRO, y calculamos el periodo de prevalencia por cada 100.000 habitantes.

Resultados

Identificamos 19.081 pacientes con diagnóstico primario de ARe, con una prevalencia estimada de 6 años de 40 casos por cada 100.000 habitantes (0,04%). Observamos una mayor prevalencia en mujeres (proporción hombre-mujer de 0.5:1). El grupo etario con mayor prevalencia de casos fue el de 75 a 79 años. La mayoría de los casos se reportaron en Atlántico, Quindío, Bolívar, Sucre y Magdalena. El número absoluto de casos fue mayor en los años previos a la pandemia de COVID-19 (2017: 3.630; 2018: 3.843; 2019: 4.711) en comparación con los años pandémicos (2020: 2.715; 2021: 3.021) y pospandémicos (2022: 2.451).

Conclusión

Este es el primer estudio que describe las características demográficas de la ARe en Colombia. La región del Caribe presenta la mayor prevalencia y los pacientes de edad avanzada fueron los más comúnmente afectados, posiblemente debido a una clasificación errónea con patologías similares como la osteoartrosis o la pseudogota.

Introduction

Reactive arthritis (ReA) is a type of inflammatory spondylarthritis that primarily affects the peripheral joints, usually preceded by bacterial infections of the gastrointestinal or genitourinary tract [1,2]. Sexually acquired ReA is more frequent than the enteric form [3]. ReA presents as an asymmetric monoarthritis or oligoarthritis affecting large joints and occurs about 2–4 weeks after infection [1,4]. Furthermore, it may be accompanied by extra-articular manifestations such as uveitis, cervicitis, keratoderma blennorrhagicum, circinate balanitis, aphthous ulcers, erythema nodosum, among others [1,2]. The most common clinical pattern is the acute form, with most patients presenting with a resolution of symptoms within 6 months; 10–30% of cases may develop chronic ReA [5].

ReA usually affects young adults between 18 and 40, with a peak incidence between 20 and 29 [5]. Worldwide, ReA has an incidence between 0.6 and 27 cases per 100,000 individuals [6]. A rise in post-COVID-19 arthritis cases, classified as ReA, has been documented. These cases exhibit a wide range of symptoms, from small joint arthritis to spondyloarthritis-like phenotypes, affecting individuals of all ages and genders [7,8].

Data describing the prevalence or demographic characteristics of patients suffering from this disease in Latin America are scarce. A Central American cohort reported a high prevalence of ReA (47%) [9], though these results may not be generalizable due to regional differences in infection rates and genetic factors.

We aimed to estimate the 6-year prevalence of ReA in Colombia and to describe its demographics. We reported the geographical distribution of cases and explored whether their prevalence changed during or after the COVID-19 pandemic.

Methods

According to data from the Colombian Ministry of Health, as of November 2024, 98.62% of the population was covered by one of Latin America's most extensive healthcare systems [10]. The Ministry created a database called Integrated Social Protection Information System (Sistema Integrado de Información de la Protección Social, SISPRO, by its in Spanish acronym), which stores and manages the crucial data needed for the system's regulation and control processes. Medical staff gather demographic and clinical information during every outpatient or inpatient visit, and these data are compiled into the Individual Health Services Delivery Registry (Registro Individual de Prestación de Servicios de Salud, RIPS, by its in Spanish acronym). These databases are accessible to the public for scientific analysis, with prior approval (username and password) from the data warehouse (http://www.sispro.gov.co/); we obtained the data for this study from the online dynamic tables. We retrieved data from January 1st, 2017, to December 31st, 2022. We performed an observational, cross-sectional study analyzing the RIPS database using the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) codes for conditions compatible with ReA (M023, M028, M029, M036; Supplementary Table 1). The diagnostic criteria for ReA in the SISPRO database are not standardized and depend on each physician. Moreover, only the “main diagnosis” is included, which may result in underreporting for patients with multiple diagnoses. Following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines, we analyzed the age distribution in 6-year intervals based on data from the latest census (2018) [11]. We determined period prevalence per 100,000 inhabitants by using the number of patients diagnosed with a code compatible with ReA (counted once) as the numerator. The denominator consisted of the population the National Administrative Department of Statistics (DANE) reported for each age group or geographic area. We adjusted the prevalence by age, gender, and region using the direct method. We also explored the period prevalence of ReA during pre-COVID-19 pandemic (2017–2019), COVID-19 pandemic (2020–2021) and post-COVID-19 pandemic (2022) years. Data were recorded and analyzed with Microsoft Excel (Microsoft Corp., Redmond, WA, USA).

Despite its limitations, such as the lack of clearly defined diagnostic criteria, SISPRO represents a valuable tool for generating knowledge in public health. It enables the estimation of disease prevalence and the analysis of geographic distribution. Data quality is higher when diagnoses are made by specialists or require paraclinical confirmation. Its use can support evidence-based decision-making [12].

Results

We found 19,081 individual patients (“cases”) with a primary diagnosis of ReA during the six years. To calculate the period prevalence, we used the estimated population of the 2018 census as a denominator: 48,258,494 inhabitants. Thus, we estimated a 6-year period prevalence of 40 cases per 100,000 inhabitants (0.04%). The male and female prevalence was 25 and 53 cases per 100,000 habitants, respectively, with a male-to-female ratio 0.5:1.

We observed a gradual increase in age group prevalence, with the highest prevalence occurring in the 75–79 year-old age group (see Fig. 1a). Table 1 shows the prevalence by age group and gender. At the same time, Fig. 2 illustrates the geographic distribution by department. The general prevalence ranged between 1.4 and 108 cases per 100,000 inhabitants; male prevalence ranged between 0 and 75 cases, whereas female prevalence ranged between 2.7 and 139 cases. Most cases were reported in Atlántico, Quindío, Bolivar, Sucre, and Magdalena.

Fig. 1.

Aged-specific prevalence of patients with reactive arthritis (M023, M028, M029, M036) during the years 2017–2022. Prevalence calculated with the average population of the period as denominator per 100,000 population.

Table 1.

Patients with a main diagnosis of reactive arthritis (M023, M028, M029, M036) by gender and age group between 2017 and 2022.

Age group (years)	Male		Female		Total population
	Patients	Prevalence^a	Patients	Prevalence^a	Patients	Prevalence^a
0–4	178	9	147	8	325	9
5–9	206	10	214	11	426	11
10–14	264	13	306	16	572	14
15–19	248	12	382	19	631	15
20–24	255	12	427	20	682	16
25–29	318	16	571	29	889	22
30–34	363	20	720	39	1083	30
35–39	392	24	880	50	1272	37
40–44	406	28	1104	70	1510	50
45–49	485	36	1381	91	1866	65
50–54	501	39	1713	116	2214	80
55–59	566	50	1506	115	2072	85
60–64	532	59	1254	118	1786	91
65–69	496	72	1019	125	1515	101
70–74	391	79	697	118	1088	100
75–79	271	83	519	130	790	109
80 or older	286	70	528	100	814	87
Total	5987	25	13,085	53	19,081	40

a

Calculated with the average information was missing for 5 population of the period as denominator per 100,000 population. Gender information was missing for 9 patients.

Fig. 2.

Geographic distribution of the adjusted prevalence of patients with reactive arthritis (M023, M028, M029, M036; a global, b male, and c female).

We explored whether the prevalence of patients diagnosed with ReA changed between pre-pandemic, pandemic, and post-pandemic years. The absolute number of cases was higher in pre-pandemic years (2017: 3630, 2018: 3843, 2019: 4711) when contrasted with pandemic (2020: 2715, 2021: 3021) and post-pandemic (2022: 2451) years (Fig. 3).

Fig. 3.

Number of patients with reactive arthritis in the prepandemic (blue), pandemic (red), and post-pandemic (green) years.

Discussion

Using the National Health Registry data, we estimated the prevalence of ReA in Colombia. We found a 6-year prevalence estimated at 40 cases per 100,000 people (0.04%). We observed a higher prevalence in women, with a male-to-female ratio of 0.5:1. The highest case prevalence was in the 75–79 age group. The departments in the Caribbean region had the highest prevalence. We did not observe a higher f prevalence of ReA during the pandemic or post-pandemic years. To the best of our knowledge, our report is the first to describe the demographic characteristics of ReA in Colombia using an official country-level health database.

Previous studies have applied ICD codes to analyze epidemiological data related to ReA [13–15]. We estimated a 6-year adjusted prevalence rate of 40 cases per 100,000 individuals (0.04%), aligning with some global and regional estimates previously reported [16]. In indigenous peoples of the Northern Arctic, the prevalence of ReA ranged from 0.25% to 1.0%. In Europe, prevalence ranged from 0.21% in Lithuania to 0.03% in Greece. A study from India did not identify cases of ReA among the 8145 people interviewed. A hospital-based study in Zimbabwe reported a prevalence of 0.001% [16]. In the USA, the prevalence of ReA ranges from 3.5 to 5 patients per 100,000 population [17]. It is important to emphasize that differences in the epidemiological data on ReA may be attributed to the absence of universally standardized diagnostic criteria and the heterogeneity of its clinical manifestations [1].

In contrast to the literature, we found a higher prevalence of ReA in women with a male-to-female ratio of 0.5:1 (Table 1) [2,5]. We hypothesize some possible factors. First, in Colombia, the prevalence of Chlamydia trachomatis and Neisseria gonorrhoeae is higher in women than in men (30.4% vs 14.5% and 25.7% vs 21.3%, respectively), as shown by a previous study using molecular diagnostic techniques [18]. Second, many patients initially see primary care physicians, who may not have the specialized training to diagnose joint pain accurately [19,20]. As a result, they might misdiagnose other inflammatory arthropathies such as ReA; our group has proposed an algorithm to overcome this limitation [19,20]. For instance, the prevalence of conditions such as osteoarthritis and acute arthritis induced by calcium pyrophosphate crystals appears to be higher among older women [21,22], which might have been coded as ReA and contribute to the observed differences. This could also explain why our analysis revealed that the highest prevalence of ReA was in the 75–79 age group, in contrast to the literature, which indicates a higher prevalence among young adults [2,5].

Regarding geographical distribution, we found a particular clustering of ReA diagnosis in the Caribbean region (Atlántico, Bolivar, Sucre, and Magdalena). The development of ReA depends on multiple factors, such as the etiological agent and genetic factors (e.g., HLA-B27) [1,2,5]. Besides its association with sexually transmitted agents [3], gastrointestinal pathogens such as Salmonella enterica, Yersinia spp., Campylobacter jejuni, Clostridioides difficile, and Escherichia coli are well-known triggers of ReA [1,2,5,23]. Thus, we hypothesize that social determinants of health, such as access to potable drinking water, may influence its prevalence. A study about the microbiological water quality and contamination sources along the coast of the Atlántico department (located in Colombia's Caribbean region) found high concentrations of bacteria such as E. coli, Enterococcus faecalis, and Clostridioides perfringens in water sources, including beaches and streams [24]. This poses residents of this region with a greater risk of gastrointestinal illnesses and, consequently, ReA.

Further, genetic background may favor a higher prevalence, such as the strong link between HLA-B27 and spondyloarthritis risk [1,2,5]; this allele is frequently found in people with European ancestry [25]. HLA-B27 positivity is not essential for developing ReA, but it increases the susceptibility to the condition [26]. HLA-B27 influences the immune response by presenting antigenic fragments to T cells, altering immune self-tolerance, elevating tumor necrosis factor-alpha levels, and extending the survival of microorganisms within the host [26]. A study that explored the ancestry landscape of Colombian admixed populations found a high prevalence of European ancestry in the Caribbean region of Colombia [27]. However, research in the Colombian population suggests that the antigen's prevalence in our population is lower than in others, where positivity rates can reach up to 90% [28–43]. Noteworthy, Londoño and colleagues reported in Colombian patients with axial spondyloarthritis that HLA-B15 was associated with peripheral disease, whereas HLA-B27 was associated with axial manifestations [44]. Previous studies in Colombia have described a prevalence of HLA-B15 between 2% and 6% in healthy people from different regions [31,33,38,45,46]. However, a study from Bolivar (a department located in the Caribbean region) found a very low prevalence of HLA-B15 antigens in this population [30]. As this study was performed more than three decades ago, the influence of migrations and globalization may have changed the HLA landscape in this region; future studies should explore this theory.

Due to the COVID-19 pandemic, an increase in reports on post-COVID-19 ReA has led to a resurgence of global interest in ReA [7,8,26,47,48]. Post-COVID-19 ReA has a transient phenotype, can present at extremes of age, seems to affect both sexes equally, and exhibits diverse manifestations [47]. Additionally, post-COVID-19 vaccination ReA cases have also been reported [49]. However, in our study, we found a higher number of cases in pre-pandemic years compared to pandemic and post-pandemic years, which might be attributed to the widespread reduction in healthcare utilization for non-COVID-19 conditions during the COVID-19 pandemic [50,51]. A recent pooled analysis suggests a higher prevalence of new-onset immune-mediated diseases in patients with previous COVID-19 infection, including spondyloarthritis [52]. Whether this may be observed also for ReA remains to be elucidated.

Limitations

We recognize some limitations in our study. First, since data were obtained from physicians’ coding in medical records, it's unclear whether the diagnoses fully align with the classification criteria. This could lead to the prevalence being either underestimated or overestimated [12]. However, this database includes the primary statistics from the Colombian Ministry of Health and is the basis for public health policy decisions. Second, prevalence might be underestimated in underserved areas, as formal diagnoses are more likely to be recorded in larger cities where rheumatologists are concentrated. Third, comparing our findings with previous studies is difficult due to different methodologies. Nonetheless, we consider that, as an official source with national coverage, SISPRO remains a valuable tool for generating hypotheses, describing epidemiological patterns, and guiding public health decisions, provided that its interpretation is carried out with appropriate caution. Fourth, the database used contains aggregated data. The lack of granularity makes it difficult to analyze specific patterns or trends in detail, particularly at the individual level. Future studies should explore strategies to explore individual details such as previous infections and previous antibiotic use, among other characteristics.

Conclusion

Based on the National Health Registry data, we estimated a 6-year prevalence of ReA of 40 cases per 100,000 inhabitants (0.04%), with a higher prevalence observed in women and older adults, contrasting with previous reports. Geographical clustering was noted in northern regions, likely influenced by environmental and genetic factors. We did not observe a higher prevalence of ReA cases during the pandemic or post-pandemic years. Our findings underscore the need for improved diagnostic accuracy, particularly in primary care, and further research is needed to understand the interaction between genetic, environmental, and infectious contributors to ReA prevalence and distribution in Colombia.

Author's contributions

All authors contributed to the conception and design of the study. Material preparation, data collection, and analysis were performed by Liz Beltran-Medina, Julián Barahona-Correa and Karen Gualdrón-Chaparro. The manuscript was drafted by Liz Beltran-Medina, Julián Barahona-Correa and Santiago Bernal-Macías. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Ethical considerations

The study was classified as a no-risk investigation, in accordance with the Helsinki Declaration, and received approval from the Institutional Research and Ethics Committee.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT to enhance the manuscript's writing. After using this tool/service, the authors reviewed and edited the content as needed, assuming full responsibility for the final publication's content.

Funding

Open Access funding provided by Colombia Consortium.

Conflict of interests

None of the authors have any conflicts of interest to disclose.