Diagnosis of polymyalgia rheumatica usually means a favourable outcome for your patient

Introduction

Polymyalgia rheumatica (PMR) is an auto-inflammatory rheumatic disease of people over 50 years, presenting with pain and stiffness in the neck, shoulder and hip girdles1. The term PMR was first used to underline that it seemed substantially milder from rheumatoid arthritis (RA) as no joint damage had been observed2. However, this name may be misleading as PMR is a disease but not a non-specific myalgic or paraneoplastic syndrome. Secondary PMR does not exist. The short definition of PMR may be also misunderstood as it does not suggest how unique the clinical picture of ‘musculoskeletal pain and stiffness’ might be. What makes PMR so special is a sudden onset of severe musculoskeletal symptoms that strongly reduce daily performance and quality of life, together with systemic inflammatory reaction in elderly but usually generally healthy patient, frequently accompanied by depressive reaction, lack of pathognomonic laboratory or radiologic findings, and splendid reaction to low-dose corticosteroids (CSs). Strong auto-inflammatory response is normally unexpected in an elderly patient, but immunosenescence3 seems not to apply in PMR. Good prognosis is also surprising as there is no good explanation to the fact that PMR patients apparently live longer than matched controls45. However, PMR is frequently accompanied by giant cell arteritis (GCA). Therefore, there is an increased probability of all the ischaemic complications related to GCA in PMR6. Further, generally favourable outcome of PMR may be easily wasted in case of excessive CSs treatment, resulting in side effect that eventually worsens patient's quality of life7.

The aim of this review article is to discuss about correct diagnosis of PMR.

Epidemiology

Annual incidence of PMR is up to 50/100,000 population over 50 yr5. It is uncommon in India89. The highest prevalence is found in Caucasians, in Northern European countries, especially in Scandinavia. Vikings' ancestry is associated with increased risk of PMR (marked by migration from Scandinavia10, in Western European lands invaded by Normands11 or Eastern European settled by Varangians12). The incidence grows with age, starting from 50 yr, with peak above 70 yr5. It is easier to find a report of wrong diagnosis of PMR in patients under 50 yr13, than an undoubtful case report of PMR or GCA in young14.

Polymyalgia Rheumatica Versus Giant Cell Arteritis

PMR and GCA are closely related. These have similar genetic background and epidemiology, and frequently overlap15. However, PMR represents a non-specific immune-mediated inflammatory reaction triggered by innate immunity activation. GCA is manifested by vascular inflammation caused by faulty adaptive immune reaction that is mainly T cell dependent16. The clinical outcome is so different that it is continuously debated why some patients express only one pathway. Evidence of vascular inflammation in needed to diagnose GCA. It is usually recorded in about 10 per cent of PMR patients17, but detailed vascular examination may substantially increase that percentage18. The historical term polymyalgia arteritica19 might still be accurate because it underlines suspected vascular pathogenesis of PMR. Silent vasculitis has been demonstrated in some of PMR patients20. Regardless of PMR and GCA pathophysiologic relations, from a practical point of view, it must not be forgotten that patients diagnosed with PMR are also at an increased risk of GCA-associated complications. The concomitant GCA implies the need for more aggressive treatment strategy (Table I). PMR symptoms can dominate over the clinical picture of GCA. Small CSs doses used for PMR do not suppress concomitant GCA (although this is more generally accepted than proved opinion)21 that may progress to cause blindness.

Table I Differences in treatment strategy underlining the need for identifying concomitant giant cell arteritis (GCA) in patients with polymyalgia rheumatic (PMR)

When Should Polymyalgia Rheumatica Be Suspected?

Proximal, musculoskeletal pain and stiffness are the leading manifestations and these raise little doubts about the diagnosis in a typical PMR case. However, PMR pivotal manifestations may be less specific: fewer or raised C-reactive protein (CRP) of unknown origin, general weakness, weight loss, depressive reaction, decompensation of chronic diseases (e.g. chronic heart failure) due to increased metabolism121.

How to Diagnose Polymyalgia Rheumatica?

There is no universal answer to this question. Although clinical presentation is typical, in some cases, the disease may be surprising. The best strategy is a combination of different approaches (based on clinical picture, classification criteria, exclusions and ex juvantibus diagnosing), together with extensive diagnostics and follow up observation of atypical cases24. Ninety per cent of rheumatologists in the United Kingdom (UK) declare the use of the classification criteria for PMR diagnosis although they admit not to adhere to the guidelines to exclude other diseases25. This approach may be accepted only in countries with high incidence of PMR resulting in relatively low number of PMR mimics26.

Diagnosis Based on Clinical Picture

Clinical assessment is most important for PMR diagnosis. There are no specific antibodies (PMR and GCA are T cell-dependent diseases1627) or additional testing to confirm PMR. Clinical manifestations come from a combination of musculoskeletal pain and stiffness and acute inflammatory response, which is quite different in elderly compared to young patients.

Diagnosis Based on the Classification Criteria

Classification criteria are not developed for diagnosing but are generally used for this purpose. Systems of classification of the diseases are demanded by clinical trials, public health and insurance systems. These criteria define most typical manifestation of the disease and provide an organized summary of its main manifestations. That is why the ‘old’ criteria such as formulated by Bird et al36 or Jones and Hazleman37 are still valid as these were formulated based on clinical observation (Table II). Yet, PMR is not a disease diagnosed by simple ticking signs and symptoms on a checklist. If it were so, a coincidence of depression and shoulder joints osteoarthritis (OA) in an elderly patient would be misclassified as PMR according to Bird's criteria. The interpretation of the importance of complaints and findings in an individual patient by an expert remains fundamental for PMR diagnosis. Shoulder girdle pain must have inflammatory character and depression cannot be endogenous to be accounted in Bird's criteria. Jones and Hazleman's criteria have a potential superiority in countries with low PMR prevalence as they have a higher specificity and include a list of most important exclusions.

Table II Summary of similar classification criteria for polymyalgia rheumatic a (PMR)

In 2012, European League against Rheumatism (EULAR) and American College of Rheumatology (ACR) formulated new classification criteria for PMR (Fig. 1)3839. These may only be applied in patients meeting preliminary criteria: age over 50 yr, bilateral shoulder aching, elevated CRP or ESR. These are a good example of classification philosophy to reduce heterogeneity of positively classified cases by selecting the subset of the most typical manifestations. Therefore, these criteria are not to be met in atypical cases of PMR. On the other hand, these have a potential to reduce the rate of false positive diagnosis38.

Close

Fig. 1

Patient meeting preliminary criteria must score at least 4 points to be classified as polymyalgia rheumatica (PMR) according to European League against Rheumatism/American College of Rheumatology criteria39, which gives three possible scenarios encompassing the combination of 2 points for morning stiffness; 2 points for absence of rheumatoid arthritis specific antibodies; 1+1 point for hip girdle involvement and absence of other joint involvement. Source: Ref 39.

Export to PPT

Ultrasound

Musculoskeletal ultrasound gains importance in rheumatology. PMR criteria integrated ultrasonographic evaluation into classification process for the first time in rheumatology39. Ultrasound criterion requires examining both shoulders (for glenohumeral synovitis, bursitis or biceps tenosynovitis) and hips (for joint synovitis or trochanteric bursitis). The intention for this is assessment of the symmetry of changes between inflammatory changes in both shoulders and between upper and lower limbs involvement. Each of these will score an additional point to the scoring algorithm. Final sum of five out of eight (6 from algorithm without ultrasound plus 2 from ultrasound examination) points enables to classify a patient as PMR with 66 per cent sensitivity and 81 per cent specificity39. Ultrasound evaluation is relatively simple to perform as the findings of merely the presence of joint effusion, tenosynovitis or bursitis are sufficient. However, these abnormalities are hardly specific for PMR. A short ultrasound examination of proximal joints usually fails to demonstrate differences between inflammatory joint diseases. The differentiation between PMR and RA is a common dilemma. A more detailed ultrasound assessment can demonstrate joints' erosions and extensive synovial proliferation of both small and large joints that are more typical for RA. Ultrasound can demonstrate degenerative or post-traumatic joint lesions as well as detect GCA overlap40 (Fig. 2).

Close

Fig. 2

Ultrasound examination of glenohumeral joint from axillary approach revealing no significant effusion inside a joint capsule (bottom of the picture). With the same approach, axillary artery is simultaneously examined with colour duplex ultrasound (left, upper side of the picture): it reveals homogenous, hypoechoic vessel wall thickening of 1 mm (indicated between “+” signs), well delineated towards the luminal side being consistent with axillary arteritis.

Export to PPT

Additional tests to help with the diagnosis

There are no pathognomonic antibodies or other PMR-specific markers discovered. Although ESR and CRP serum levels are not specific, it is hard to find PMR with normal inflammatory parameters41. Other acute phase markers (fibrinogenemia, thrombocythemia and elevated IL-6, the latter correlates best with the disease activity42) are also present. Anaemia of chronic disease type is common and is reversed shortly after CSs treatment initiation. Sometimes, slightly increased transaminases and alkaline phosphatase levels are present. Sparse studies indicated a significantly higher occurrence of anti-phospholipid antibodies, but these have not been proved to be associated with ischaemic or thromboembolic complications42434445.

The benefits and drawbacks of classification criteria sets must be duly considered before applying them for diagnosis. The 2012 proposed EULAR/ACR PMR criteria38 (opposed to the previous sets) aim to minimize the role of clinical intuition and build on more objective features and additional tests. However, ESR, rheumatoid factor and ultrasound changes still have only limited specificity. The drawbacks of all of the PMR criteria sets are their unsatisfactory sensitivity and specificity. They were also formulated in populations with a high PMR prevalence. If classification criteria are not met (which usually takes place in atypical PMR), the disease should not be diagnosed hastly but only after excluding other causes of similar symptoms2324.

Diagnosis by Excluding Other Causes of Similar Symptoms

The need for considering PMR exclusions was underlined in the previous criteria37 and is also found in the current guidelines23. The typical clinical picture of PMR requires only basic differential diagnostics. The more atypical the clinical picture, the wider differential diagnostics is required. The differential diagnostics in countries with low PMR incidence requires considering the relatively higher number of PMR mimics. It was illustrated in a study from Turkey that 30 per cent of patients with final PMR diagnosis were hospitalized, 30 per cent were treated with antibiotics, and in 29, 22 and 19 per cent abdominal, chest and brain computed tomography (CT), respectively, were performed. This approach seemed reasonable; however, there were still 13±13 months delay from onset of symptoms to diagnosis46.

Why do PMR-like manifestations mask the symptoms of other diseases? PMR pathogenesis is mediated by innate immunity. It triggers non-specific inflammatory reaction which is not unique for PMR47. Acute inflammatory response can mask more characteristic symptoms of a disease that are not reported by patients. For example, elderly onset RA may go with systemic inflammatory manifestations and large joint involvement that cause patient immobilization. Therefore, small joints inflammation is not reported by a patient whose main complaint is inability to get out of bed. Further, serious, GCA-associated ischaemic manifestations (such as double vision or jaw claudication that are typical prodromal symptoms of vision loss) can be unreported by patients seeking medical advice because of much more disturbing manifestations of overlapping PMR. Paraneoplastic syndromes that can be manifested long before an appearance of symptoms associated with tumour growth may also be misclassified as PMR48.

Diagnosis ex juvantibus

Rapid and spectacular improvement shortly after CSs introduction enables concluding on a cause based on an observed response to the treatment. Therefore, PMR patients are much pleased shortly after treatment initiation and grateful to their doctors55. Lack of improvement within five days suggest a need to verify the PMR diagnosis (to some different disease or GCA)22. Diagnosis ex juvantibus was included in Jones and Hazleman's criteria37. New, 2012 PMR classification criteria do not include a good response to CSs in the diagnostic process39. It raised some discussion as this criterion is widely used in a daily practice56. It was argued that treatment response was non-specific and difficult to define feature. Indeed, elderly onset RA or other inflammatory conditions also respond well to CSs. However, this response would be weak in OA and transient if applied in infection or neoplasm. In these cases review of the initial PMR diagnosis is needed. It is not a mistake to reassess the initial diagnosis and change it accordingly. About 10 per cent of patients initially diagnosed with PMR are later reclassified as having elderly onset RA57. For that purpose, careful monitoring of patients is needed. PMR patients require regular medical check-ups.

Diagnosing ex juvantibus may also be made eagerly because it does not require an effort of time-consuming and expensive procedures. Spectacular efficacy of CSs in PMR may cause a temptation to overuse them. There has been increased interest in GCA and PMR evoked by modern treatment possibilities5859 and fast track GCA clinics60. Based on our own experience, PMR is easy to overdiagnose. Establishing rational PMR diagnosis illustrates a challenge to resist fashion and wishful thinking in medicine. Adherence to the 2012 PMR classification criteria could be beneficial in preventing overdiagnosis because these do not include response to therapy. However, the ‘response criterion’ will not disappear from the daily practice of the physicians who can appreciate how unique PMR symptoms react to CSs. At least as long as there are no better disease markers.

Conclusion

Lack of specific biomarkers of PMR is problematic and research is needed. Up to now, the diagnosis remains clinical. There are many clinical subtleties to be considered. Usually, PMR responds well to CSs treatment and outcomes are favourable. However, differential diagnosis encompasses diseases with bad prognosis; therefore, PMR overdiagnosis can be detrimental.