The incidence and clinical characteristics of peripheral arthritis in polymyalgia rheumatica and temporal arteritis: a prospective study of 231 cases
Abstract
Objective. To evaluate the incidence and characteristics of peripheral arthritis in polymyalgia rheumatica and temporal arteritis, and to ascertain the incidence of rheumatoid arthritis among such cases.
Patients and methods. In total, 231 patients were selected from a prospective population‐based study. All patients were clinically examined on several occasions and followed until cessation of therapy and permanent disease remission.
Results. Of the 231 cases, 38.5% presented peripheral arthritis either at diagnosis or during the disease course. At diagnosis, peripheral arthritis was not observed among patients with temporal arteritis. Peripheral arthritis occurring during the disease course was more often polyarticular and needed additional treatment more frequently than joint inflammation presenting at diagnosis. Only one case had distal pitting oedema. Rheumatoid arthritis developed in 4.8% of the cases and exclusively among patients with polymyalgia rheumatica.
Conclusion. Aetiopathogenic differences may exist between polymyalgia rheumatica and temporal arteritis as peripheral arthritis and the development of rheumatoid arthritis were observed among the former patient group only.
Key words
Polymyalgia rheumatica (PMR) and temporal arteritis (TA) have traditionally been regarded as one disease expressing different clinical manifestations. This view has been based on findings of a common genetic disease susceptibility [1] and a rather frequent coexistence of the two syndromes [2]. However, the significant clinical differences between PMR and TA may suggest separate aetiopathogeneses.
First, the occurrence of arteritis in PMR may not be as frequent as previously anticipated [2]. According to our own observations [3], arteritis of the temporal artery is detected histologically in only 6% of PMR when biopsies are performed in patients lacking clinical manifestations of concurrent TA. Second, the overwhelming majority of patients of the PMR/TA disease complex suffer from pure PMR without clinical signs and symptoms of TA [3]. Third, the frequent occurrence of peripheral arthritis in PMR may suggest that components in synovial tissues represent important antigenic targets in the aetiopathogenesis of PMR. Finally, our own preliminary observations [3] which indicated that peripheral arthritis may occur exclusively in PMR and not in TA may also emphasize important aetiopathogenic differences between PMR and TA.
It has also been suggested that the peripheral arthritis developing in PMR may be difficult to discriminate from that occurring in elderly onset seronegative rheumatoid arthritis (RA) [1, 4–6] and remitting seronegative symmetrical synovitis with pitting oedema (RS3PE) [7]. Thus, PMR may occasionally be more related to RA and RS3PE which affect synovial tissues than to TA whose predominant pathological finding is arteritis. However, the frequency of RA development during the disease course of PMR is not finally determined.
It was thus of interest to study the incidence and characteristics of peripheral arthritis in a large group of unselected patients with PMR and TA, followed prospectively from the onset of disease and throughout the entire disease course. The longitudinal design of the survey also allowed us to estimate the incidence of RA among such patients and to explore possible aetiopathogenic differences between PMR and TA.
Materials and methods
Geography
The county of Aust Agder, south Norway is mainly located on the coast, and in the period 1987–1994 the total population averaged 98 000 inhabitants of whom 29% were 50 yr or older.
Methods
Prior to the start of the study, all physicians in the county of Aust Agder were informed of the study, and asked to refer all patients suspected of having PMR or TA to the Department of Rheumatology as soon as possible and before initiating drug therapy. The county of Aust Agder has one general hospital including departments of rheumatology, internal medicine, geriatrics and ophthalmology. There are no private practising rheumatologists in the region.
In order to assure some minimum length of observation, only patients with a disease onset during the period 1987–1993 were included in the study. They were followed up at 3‐month intervals the first year, thereafter twice yearly until cessation of therapy and permanent disease remission. All patients were examined by a rheumatologist. In addition, all hospital records with a diagnosis of either PMR or TA were carefully reviewed to estimate the number of patients with such disorders who had been treated by the departments of internal medicine, geriatrics and ophthalmology.
The American College of Rheumatology (ACR) criteria [8] for TA were applied but only cases with histologically proven giant cell arteritis were included. For PMR, patients meeting either the criteria suggested by Bird et al. [9] or those of Hamrin [10] were selected for the study. Thus, all patients with PMR included in the present study had bilateral pain and/or stiffness of the shoulder and/or pelvic girdle which could not be explained by diseases other than PMR. The cases were divided into the following groups: (A) Patients exhibiting peripheral arthritis at diagnosis only. (B) Patients with peripheral arthritis at diagnosis and later during the disease course. (C) Patients without peripheral arthritis at diagnosis, but developing such features during the disease course. (D) Patients in whom peripheral arthritis was never observed. The mean observation time was 56.8, 75.4, 75.6 and 58.0 months in groups A, B, C and D, respectively. Peripheral arthritis was defined clinically as swelling and pain or local warmth in a joint. Magnetic resonance imaging and scintigraphic examinations were not performed.
Laboratory examinations
The following blood tests were performed in all patients: complete blood count including white cell count and thrombocytes, haemoglobin, erythrocyte sedimentation rate (Westergren) (ESR), C‐reactive protein (CRP), creatinine, creatine kinase, alaninoaminotransferase, aspartateaminotransferase, gammaglutamyltranspeptidase, alkaline phosphatase, thyroxin (T4), thyroid stimulating hormone, serum electrophoresis, uric acid, rheumatoid factors (Waaler's test), and antinuclear antibodies. Radiological examinations were performed only if development of erosive arthritic disease was suspected or the patient was evaluated for surgical intervention.
Statistics
The chi‐squared test, Fisher's exact test and Student's t‐test were used for statistical analyses, and a P value of 0.05 or less was accepted as significant.
Results
Classification of patients
Prior to the study, 26 cases were excluded. Ten patients were withdrawn from the investigation due to coexisting RA and 16 patients were lost during the follow‐up. In total, 231 patients, 151 females and 80 males were studied. There were 187 cases of pure PMR, 29 cases of TA and 15 patients with both PMR and TA.
RA
The development of RA was observed in 11 patients (4.8%), six females and five men, during the observation period (Table 1⇔). Of the 11 cases, 10 cases had pure PMR and one case suffered from both PMR and TA. None of the patients with pure TA developed RA. The mean duration of PMR at onset of RA was 62.3 months (median 59 months, range 6–130 months). Nine of the 11 patients developed RA more than 3 yr after the onset of PMR. Radiological examinations of hands and feet were performed in eight cases, of whom five had erosive changes typical of RA. Serum rheumatoid factors were determined in all cases and six of these cases possessed such autoantibodies. Thus, seropositive erosive RA developed in three cases (1.3%).
TABLE 1.
Characteristics of patients developing RA during disease course of PMR (n = 11)
Frequency of peripheral arthritis
Table 2⇔ shows the frequencies of peripheral arthritis in the various groups. In total, 38.5% of the patients presented peripheral arthritis. Peripheral arthritis occurring exclusively at diagnosis was seen in 14.7% of the cases, 7.8% presented peripheral arthritis both at diagnosis and during follow‐up, and 16% had no peripheral arthritis at diagnosis but developed such manifestations later on. At diagnosis, peripheral arthritis was not seen in TA, while one TA patient developed transient monoarticular peripheral arthritis during the observation period.
TABLE 2.
Frequencies (%) of peripheral arthritis in PMR and TA
Characteristics of the arthritis
At diagnosis, the frequencies of monoarthritis, oligoarthritis and polyarthritis were 34.0, 62.3 and 3.8%, respectively. During the disease course, the corresponding frequencies were 26.9, 48.1 and 25%, respectively. Polyarticular joint inflammation was significantly more frequent during the disease course than at diagnosis (P = 0.005), even when calculations were performed after exclusion of patients developing RA.
Among cases with peripheral arthritis contracted exclusively during the disease course, 25 of 37 (67.6%) had one episode of arthritis, while 12 patients (32.4%) had multiple episodes of arthritis. Among those with only one episode, the first episode developed during the first year of disease in 68% (range 1–86 months). Among those with multiple episodes, six of 12 (50%) developed their first arthritic attack during the first year of disease (range 6–53 months), while the last episode was seen as late as after 147 months of disease (range 9–147 months).
Distribution of arthritis
Table 3⇔ shows joint involvement at diagnosis and during the observation period. The distribution of joint involvement was similar at diagnosis and during observation. The most commonly involved joints were the knees (46.7%), metacarpophalangeal joints (MCP) (42.0%) and wrists (42.1%). Sternoclavicular joint inflammation was seen in three patients (2.8%) of whom one had erosive changes at X‐ray. Distal pitting oedema was recorded in one case only. Isolated Baker cysts without concomitant detectable knee arthritis were seen in five cases.
TABLE 3.
Joint involvement (%) in PMR and TA
Therapy of arthritis
All 231 patients except one were treated with oral corticosteroids (prednisolone). Therapy of arthritis presenting at diagnosis was recorded in 51 cases. Of these, 44 (86.3%) required no therapy in addition to oral corticosteroids, while non‐steroidal anti‐inflammatory drugs (NSAIDs) were prescribed in three cases and intra‐articular corticosteroid injections were given to four patients. During the observation period, therapy of arthritis was recorded in 55 cases. Of these, 19 cases (34.5%) were given no additional treatment, 17 cases (30.9%) were given NSAIDs, 15 cases (27.3%) received intra‐articular corticosteroid injections, seven patients (12.7%) were given disease‐modifying anti‐rheumatic drugs (DMARDs) and two patients (3.6%) had surgical knee synovectomies. The frequency of additional treatment was significantly higher during follow‐up than at diagnosis (P = 0.001).
Comparison between patients with and without arthritis
The mean age at onset of disease was similar in patients with and without peripheral arthritis (69.5 and 72.2 yr, respectively). Arthritis occurred in 41.5% of females with PMR compared with 53.6% among male PMR (P = 0.14).
The mean ESR at diagnosis was 66.9 among cases without peripheral arthritis, 67.9 in those with arthritis at diagnosis only, and 85.6 among those who developed peripheral arthritis during the disease course but exhibited no arthritis at diagnosis. Similarly, the mean CRP at diagnosis was higher (88.6) among those with arthritis later on exclusively when compared with patients with arthritis at diagnosis (65.0) and compared with cases without arthritis (59.7). However, the differences in ESR and CRP between the various groups did not reach statistical significance.
The mean minimum dose of oral corticosteroid (prednisolone) during the first year of disease was 5.5 mg in cases with arthritis at diagnosis, 6.4 mg in those with arthritis during the disease course only and 5.7 mg among patients exhibiting no joint inflammation. These differences did not reach statistical significance. The mean dose of prednisolone at the start of treatment in patients with PMR having arthritis at diagnosis only was 18.4 mg, 31.5 mg in patients with arthritis both at diagnosis and later, 27.7 mg in those with arthritis later only and 22.7 mg in cases without arthritis.
Discussion
The prospective design of the present study allowed us to study peripheral arthritis both prior to initiation of corticosteroid therapy and during the disease course. As only 16 patients (6.5%) were lost on follow‐up, the present patient material appears highly representative of the total sample. Moreover, this epidemiological survey was performed in a county having only one single hospital and no private practising rheumatologists, which further emphasizes the representativeness of the patients. As the patients were followed for the entire disease course, the risk of including cases with diseases other than PMR and TA but manifesting polymyalgic symptoms was substantially reduced. However, only clinically detectable peripheral arthritis was recorded, and routinely performed ultrasonographic examinations [11] and magnetic resonance imaging [1] could have revealed an even higher incidence of such manifestations.
It should be noted that in the present study, cases with pre‐existing RA were excluded from the calculations. The cases had long‐standing seropositive erosive RA and whether the peripheral arthritis was due to PMR or represented an exacerbation of their RA was rather difficult to evaluate. The typical characteristics of RA made it unlikely that their pre‐existing peripheral arthritis was a preceding manifestation of PMR.
The total incidence of peripheral arthritis was rather high (38.5%), confirming previous observations [4, 5, 12–14], but at variance with others [15]. An important finding was that peripheral arthritis occurred almost exclusively in patients with PMR, and no case of peripheral arthritis at onset of TA was observed. A possible explanation of why this particular absence of peripheral arthritis in TA has been infrequently noted by previous workers [16], may be that most often PMR and TA have not been evaluated as separate disease groups. The observation may, however, to some extent indicate important aetiopathogenic differences between PMR and TA. Such a view is further emphasized by our previous findings of histological arteritis being a rather rare finding in patients with pure PMR [3, 17]. Thus, synovial tissues appear as possible antigenic targets in PMR while structures in the arterial wall are of aetiopathogenic significance in TA.
Studies of HLA molecules [18, 19] have not provided findings to determine the precise relationship between PMR and TA. One concept is the presence of two different clinical expressions of the same disease, while another possibility is that PMR and TA are two separate disorders. An alternative explanation would, however, be that a common genetic predisposition exists in PMR and TA regarding disease susceptibility, but that other yet unknown factors influence the clinical expressions of the syndrome. The present observations support the view of two separate disorders, or one common disease in which host susceptibility influences clinical expressions.
On the other hand, RA developed rather infrequently among the patients which lends little support to suggestions of PMR and RA being closely related. In fact, seropositive erosive RA developed in only 1.2% of our cases. Moreover, the peripheral arthritis in PMR that presented both early and late during the disease course was most often of the oligoarticular type rather than polyarthritis which prevails in RA. In a recent study [12] of 128 patients with giant cell arteritis, six patients (4.7%) developed polyarthritis compatible with a diagnosis of RA. However, none possessed serum rheumatoid factors and only one patient developed radiological joint erosions. These findings of a rare occurrence of RA among patients with the PMR/TA disease complex are in agreement with earlier studies [20, 21], and clearly show that RA and PMR represent two separate disorders.
We were unable to confirm recent findings of a frequent coexistence of RS3PE and PMR [12]. Only one case with pitting oedema was observed, but it should be noted that the present prospective study was not constructed to look specifically for this clinical feature. Some cases may therefore have been overlooked. However, the patients were carefully evaluated at regular intervals and a frequent occurrence of a RS3PE‐like clinical picture would most likely have been noted. Thus, the results of the present investigation do not indicate significant relations between PMR and RS3PE.
Similar to earlier findings [1], the most frequently involved joints were the wrists, MCPs and knees. It has been suggested that the involvement of extra‐articular synovial tissues is of aetiopathogenic importance in PMR [1], but the frequent involvement of such distal joints as the wrists and MCPs hardly explains the predominant proximal myalgia occurring in PMR. Arthritis may, however, have been present in shoulders and hips without producing sufficient clinical findings to awaken our suspicion, a finding that has been reported by others [5, 11]. Thus, the present results do not speak against the hypothesis of synovial involvement being of primary importance in the development of PMR. However, as the overwhelming majority of our patients with peripheral arthritis recovered completely, the synovial pathogenesis in PMR is clearly different from that of RA. Recent studies [1] employing hand and foot magnetic resonance imaging have also shown that inflammation of the tenosynovial sheaths may represent the hallmark of PMR.
At diagnosis, few of the patients with peripheral arthritis required additional therapy for their arthritic manifestations. However, during the disease course, 41 of 60 events of peripheral arthritis required therapy in addition to oral corticosteroids. Moreover, polyarticular joint inflammation occurred significantly more frequently during the disease course than at diagnosis. Neither the initial dose of corticosteroids nor the maintenance dose during the first year of disease varied between these groups, and the mode of therapy could thus not explain the discrepancies in the occurrence of polyarthritis. However, the higher frequency of polyarthritis again emphasized the fact that the late peripheral arthritis was more severe than that presenting early in the disease course. Most cases were, however, managed on NSAIDs and local corticosteroid injections, but DMARDs were employed in seven cases and two knee synovectomies were performed. These findings call for careful follow‐up of patients with PMR, and suggest that such manifestations may be best treated by rheumatologists.
In conclusion, the finding of a rather high incidence of peripheral arthritis both early and late in the disease course of PMR may lend support to suggestions of synovial tissues being of aetiopathogenic significance in this disease. The almost complete absence of peripheral arthritis in TA suggests important aetiopathogenic differences between PMR and TA, but the rather infrequent development of RA among such cases does not support the view of a close relationship between PMR and RA. Although most cases exhibiting peripheral arthritis are handled by the corticosteroid therapy already in use, the frequent need for NSAIDs, intra‐articular corticosteroid injections and occasional DMARD prescriptions during the disease course do suggest that patients with PMR are best treated and observed by rheumatologists.
Footnotes
- Correspondence to: J. T. Gran, Revmatologisk Avdeling, Aust Agder Sentralsjukehus, 4800 Arendal, Norway.
- © British Society for Rheumatology
Nenhum comentário:
Postar um comentário