Incidence and Prevalence of Polymyalgia Rheumatica and Giant Cell Arteritis in a Healthcare Management Organization in Buenos Aires, Argentina

DISCUSSION

To our knowledge, our study is the first to provide PMR and GCA epidemiological data in Argentina and one of the first ones in Latin America. Data published around the world are variable. PMR incidence is high in Nordic countries,^9-11 as well as in Olmsted County in the United States,¹² where most of the population descends from Northern Europe. One of the highest IRs is the one published in the United Kingdom, which was 84.2/100,000 PY in the period between 1990 and 2001.¹³ In contrast, in Southern European countries, IRs are low, at approximately 22/100,000 PY.¹⁴ In a previous review of the epidemiology of PMR in Italy, an incidence between 12 and 23 cases per 100,000 inhabitants was reported.¹⁵

The IR in our population was higher at 78.8/100,000 PY (95% CI 73.4-84.2), but it was within the range of rates reported in a recent systematic review and close to those reported in the UK and US.¹⁶ Regarding Latin America, in Colombia, a PMR prevalence was reported of 200 per 100,000 inhabitants aged > 50 years¹⁷; these results are very similar to our data (255/100,000).

There are several differences between the reported studies that might explain varying results, such as the criteria used for diagnosis (some studies used Bird or Healey criteria, others the ACR criteria, and others physician diagnosis), the use of normal ESR as an exclusion criterion (as in one of the Italian studies),¹⁵ different cut-off values for ESR as inclusion criteria, and different populations included (such as patients diagnosed by general practitioners, those referred to rheumatology centers, population-based studies). With all these differences among studies, it is difficult to draw a conclusion on which data are more accurate. We took a population-based approach, although our population might have some special characteristics that are not necessarily generalizable, as discussed below.

In a recently published metanalysis that included 107 studies, the pooled incidence (95% CI) of GCA was 10.00 (9.22-10.78) cases per 100,000 people aged > 50 years.¹⁸ The incidence was highest in Scandinavia at 21.57 (18.90-24.23), followed by North and South America at 10.89 (8.78-13.00), Europe at 7.26 (6.05-8.47), and Oceania at 7.85 (−1.48 to 17.19). These figures are similar to what we found in our study. Pooled prevalence in that metanalysis was 51.74 (42.04-61.43) cases per 100,000 people aged > 50 years.¹⁸ Our prevalence was within that range, lower than the ones reported in Denmark, Italy, the US, and Germany, and higher than those reported in Tunisia, Japan, Turkey, and Spain.¹⁸ As in PMR, several factors might explain the differences. First, there is the lack of a standardized definition for GCA, in particular when administrative databases are used, as there are no specific billing codes for GCA. Second, there is inconsistency in the inclusion criteria used, wherein the majority of hospital-based studies included only biopsy-proven cases, whereas most population or community-based studies (such as ours) also included clinical diagnoses. Last, most studies on prevalence included cases with GCA or PMR with a past diagnosis who were alive and living in the study area at the time of the study.¹⁶ Because as we consider that a patient who has stopped all treatment and has no symptoms does not have prevalent disease, we did not include those patients; this exclusion might explain our lower prevalence.

This work has several limitations. One is that the HIMCP is a medical care plan that does not incorporate “sick” patients. Therefore, our prevalent cases could be reduced because people with either disease would not have been allowed into the plan. Although this limitation would affect prevalence and would not affect incidence, it is possible that PMR or GCA is more often diagnosed in healthy people rather than in patients with multiple comorbidities. It is also possible that healthier patients receive corticosteroids for extended periods of time (usually, we are more concerned about steroid use in patients with multiple comorbidities), and that bias could affect prevalence.

Another limitation is the generalizability. Although the population of the HIMCP is similar in age and economic distribution to the overall CABA population (Supplementary Tables S1-3, available with the online version of this article), we have to be cautious in considering this valid for the whole city and even more so to the rest of the country. However, ours are the first results published in Argentina, to our knowledge, and may be useful for future comparisons.

Another difficulty is GCA definition. As we currently know, GCA is a more complex disease affecting not only cranial arteries but also large vessels, mainly thoracic. In this study and almost every study published around the world, GCA epidemiology takes into account mainly cranial disease, since large vessel involvement is not easily recognized nor properly registered in databases and EMRs. Therefore, the true incidence and prevalence of the disease may have been underestimated.

We did not investigate whether patients with both GCA and PMR symptoms fulfilled PMR classification criteria also; we classified them as GCA only. In our cohort, there were 53 patients with incident GCA and PMR symptoms (Table 1). In the case that these patients would have fulfilled PMR classification criteria and we counted them as having incident PMR, the PMR IR would have been 83.9/100,000 PY, a little bit higher than the one we are reporting but within the 95% CI. In cases where both entities coexist, GCA diagnosis is the one that will lead treatment decisions; that is the reason why we have classified it as such, but we do recognize that it can be another study limitation.

Differences and similarities in incidence and prevalence with cohorts from other parts of the world are difficult to explain. True genetic and environmental differences may be, in part, confounded by referrals bias and case definitions, particularly for GCA.

To our knowledge, this is the first GCA epidemiological study published from Latin America and one of the first ones showing epidemiological PMR data of this region. We know that the Latin American population is different from those in other sites in the world, and having appropriate data may allow us to have a better understanding of diseases and to plan for proper medical care of these patients. Clearly, population-based epidemiology studies are needed in Latin America.