Osteoarthritis (OA) is the most common rheumatic disease in the general population and its prevalence increases with age.1,2 Genetic, mechanical, and several other factors have been implicated in its origin.3 OA symptoms such as joint pain, reduced mobility, crepitus, and swelling result from articular cartilage loss, subchondral bone proliferation, bone misalignment, and synovitis. The management of patients with OA includes a number of pharmacologic and non-pharmacological recommendations tailored in accordance to the joints involved.4–7 General physicians (GPs) see most patients with OA; referrals to specialized departments depend on surgical needs and patient preference.8

The knee is the largest joint most commonly involved in OA. The prevalence of knee OA ranges from 1.4% in urban Philippines15 to 19.3% in rural Iran16 whereas that of “painful knee OA” ranges from 6.5% to 28%.9–13 The age-standardized prevalence of knee OA is 1770 per 100000 males and 2693 per 100000 females.14 Along with hip OA, knee OA impairs physical functioning and decreases the quality of life of patients with OA.17–19 Mechanical factors, particularly those involved in the wear and tear processes in the joints, including body overweight and heavy-load work are major causes of the disease.20 There is an important association between knee pain and disability, which depends on the age of the patient, disease characteristics, and outcome measures.10–12,21

The prevalence and characteristics of OA in Latin America is partially known. Studies made according to World Health Organization/International League Associations for Rheumatology Community Oriented Program for the Control of Rheumatic Diseases (COPCORD) have shown a that the prevalence of OA in 2 Mexican studies is 2.3%22 and 10.5%,23 in Guatemala 2.8%,24 in Brazil 4.1%,25 in Peru 14.4%,26 and in Cuba 20.4%.27 Although these variations may reflect ethnical factors, it is also possible that methodological differences, demographic characteristics, and medical care accessibility might also influence those findings. Consequently, we developed an initiative for the study of knee OA in order to document its clinical features and consequences in patients living in Argentina, Brazil, and México. Specifically, we aimed to describe the level of symptoms and functional limitations in patients with clinical knee OA, the types of health care they access and the strategies followed in each country to deal with signs and symptoms of the disease.

This was an observational, cross-sectional study of patients with symptomatic knee OA carried out in Argentina, Brazil, and México. Institutional review boards at each country approved the protocol and informed consent of this study. The participants were informed about the nature of the study before signing an informed consent.

The population included in this study was a convenient sample of ambulatory patients with knee pain attributed to OA28 by trained General Physicians (GPs) in primary medical care units from where patients were referred to Rheumatology specialized departments in Argentina, Brazil, and Mexico in a 12-month period. Case definition referred to men and women >18 years old who complained of unilateral or bilateral weight bearing, non-traumatic, and non-inflammatory knee pain for >6 months and a current episode of continuous or sporadic knee pain of >1 month. Patients with a confirmed diagnosis of any active inflammatory rheumatic disease, including rheumatoid arthritis, spondyloarthritis, gout, calcium pyrophosphate dehydrate crystal deposition disease, connective tissue diseases, and infectious arthritis were not included in the study. Likewise, patients with congenital lower limb malformations, serious functional impairment, disability or chronic disease causing impaired functioning as consequence of foot, ankle, or hips involvement were not included in the study. The diagnosis of knee OA was established by GPs according to the American College of Rheumatology classification criteria.28 Patients’ assessment, including demographic data, clinical symptoms, physical function, health care utilization, and therapeutics was performed at the Rheumatology department.

We collected information on age, gender, occupation, height, weight, and body mass index (BMI). Ethnicity was operationally defined in accordance to parents and all 4 grandparents’ ethnicities as follows:29 (1) Whites: subjects with White European ancestors; (2) Mestizos: subjects with Amerindian and White European ancestors; (3) African Latin Americans: subjects who have at least one African ancestor irrespective of whether other ancestors were White or Amerindian; and (4) Pure Amerindians: subjects with all autochthonous ancestors.

Clinical variables included the duration of the disease and of the current episode of pain, its characteristics, and intensity in the last 6 months. We classified the localization of knee pain either as generalized (tenderness in all the knee) or localized (tenderness either in the internal, external, or patello-femoral aspects of the knee). We also looked for joint crepitus during movement and knee swelling. The assessment also included a characterization of diagnostic and therapeutic medical and non-medical interventions, type of health insurance coverage, and physical functioning using the Spanish version of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC).30 We also incorporated an item exploring the ability to perform activities of daily living and its relationship with keen pain as follows: (1) I do not have any problem to perform my normal daily activities, (2) I have some difficulties to perform some of my daily activities due to knee pain, (3) I have difficulties to perform most of my daily activities due to knee pain and (4) I am unable to perform most of my daily activities due to knee pain.

We have assessed the characteristics of knee pain related to OA and its impact on 1210 patients referred from primary health clinics to Rheumatology departments in Argentina, Brazil, and México. Rather than finding a common pattern, our study showed important differences in the frequency of most demographic and clinical variables. Some of such differences have already been described Latin American studies of patients with systemic lupus erythematosus,29 spondyloarthritis,31 and early rheumatoid arthritis32 and have been explained of the basis of ethnical, environmental, and demographic factors. In addition, the role of educational, socioeconomic factors, and health care system should also be considered.33

Nevertheless, general assumptions in this study include the onset of knee OA pain before the age of 65 years and more women than men affected by the disease. Despite a small proportion of our patients had a formal job, very few had lost it because of the disease. Only 4% of the patients were unable to do any physical activity and 25% had low WOMAC scores and no physical limitations at all. These findings suggest that the severity of knee OA in our study was actually mild to moderate and that its pattern is certainly similar to that of patients attending primary care units in which OA referrals to specialized departments depends on the need for joint replacement, physical therapy and rehabilitation, comorbidities management, and demographic factors.8,34–40 Despite all patients in our study were under the care of GPs, most of them were already seen by rheumatologists and orthopedic surgeons.

The overall and last episode of pain lasted for more than 5 years and around 6 months, respectively and their intensity was 5.2 on a VAS. These findings seem to agree with studies suggesting that the longest the duration of knee OA complaints, the higher the intensity of pain.40,41 In the regression model, having knee pain >5.2 was associated with being treated in Argentina or Brazil, the number of medical visits in the last year, and having a drug prescription for knee pain.

Nearly 3 quarters of the patients had taken medications throughout the course of the disease, but very few took them on a daily basis; 52% had taken NSAIDs, 32.5% non-narcotic analgesics, and 27.8% anti-OA medications. The consumption of analgesic and NSAIDs relied on the severity of symptoms than on a regular basis schedule. Analgesic and NSAIDs prescription in patients with knee OA has been linked to disease perception,8 age, gender, medical and social contexts,42 health care utilization and health professional contact,37,39 performance of heavy load tasks,43 and need of surgical treatment.44 In our study, the level of pain, age, and country of origin were associated with NSAIDs’ prescription in the regression analysis. Finally, 30% of the studied population received OA modifying drugs, but differences between countries were also found. Overall, the therapeutic approach for knee OA in this study seems consistent with current recommendations for the management of the disease.4–7

Several studies have shown that age, gender, BMI, number of co-morbidities, type of job, access to medical care, and disease duration in patients with knee OA are associated with reduced HRQoL, poor functioning, job loss, and increased burden of disease.17–19,45–55 Except for those cases requiring surgery, the radiographic features in knee OA do not correlate with clinical findings and do not determine HRQoL or physical functioning.45,49,51,52

Comparisons between countries yielded some interesting differences in demographic characteristics, health care utilization, and some clinical and therapeutic features. This should not be unexpected since previous studies have linked the characteristics of knee OA to ethnicity, living conditions, and access to care.14,20,35,37–40,55–60 As consequence the fragmentation of the health system in Latin America, only a variable percentage of the population is either fully or partially covered by the state. For those who are partially covered and those who are not covered by any of such systems, the cost of disease is paid with their own money, either through direct payments or the acquisition of a private medical insurance. Health care access is limited and the burden of disease for the patient and householders is huge.61 Enabling variables, such as pain and reduced function and not the need of care have a direct effect on the utilization of health service, specifically the availability of health coverage, the existence of a regular source of care, and the out-of-the pocket money. This finding might indicate that patients with chronic diseases such as knee OA do not only learn to live with their disease, recognizing their symptoms, but also know how to deal with the health-care system.

We were interested in the role of some of such as well as other independent variables in pain level, disability, and use of NSAIDs. The contribution of country origin was significant in all 3 logistic regression models. Patients from Argentina were over 3 times more likely to report pain ≥5.2, but were 2.97 times more likely to be on NSAIDs than Mexicans. Compared with Mexicans, Brazilians were 0.29 and 0.22 times less likely to be on NSAIDs or present disability than Mexicans. Argentinians were 0.23 times less likely to present disability than Mexican patients. The significance of the variable country of origin as predictor of pain level, disability, and NSAIDs usage might support the role of each country health-care systems and less likely ethnicity in the burden of knee OA.

For every additional medical visit, patients were 0.83 times less likely to report pain ≥5.2, but 1.7 times more likely to present disability. Each incremental unit in VAS pain was associated with 1.29 and 1.19 more chances of being on NSAIDs and having disability. Patients who had a drug prescription were 2.8 times more likely to present pain equal or above to 5.2 on a VAS and 2.18 times more likely to present disability. For each year increment on age there is a 0.98 times less chance to receive NSAIDs in this population.

There were some potential limitations to our study. Regarding the representativeness of the populations included in our study, our approach was the identification of knee OA at primary health care clinics. In this regard, we followed the usual steps took by the patients with knee OA in the 3 countries. Following a standard protocol, GPs referred consecutive knee OA patients to the rheumatologist for detailed characterization of demographic and clinical variables. Differences between countries in regard to demographic and access to care seemed to reflect the characteristics of the population and health care in each of the 3 countries.

In conclusion, the burden of knee OA in Latin American countries seems to be related with demographic factors, including the characteristics of the health system. Knee OA in primary health clinics seems to be a mild form of disease with mild impact on functioning. At this stage, knee OA could be the target of disease modifying agents as treatment. We think that the assessment of disease, particularly knee OA should include the individual analysis of each Latin American country. By following that approach, health care planning, from detection of disease to its treatment would be more appropriate to the needs of each country.

Supported by an unrestrictive fund from Expanscience Laboratories.

Hafid Halhol is employed by Laboratories Expanscience. The other authors declare no conflict of interest.