Article Text

Download PDFPDF

Seasonal pattern in the onset of polymyalgia rheumatica
  1. F Perfetto1,
  2. A Moggi-Pignone1,
  3. A Becucci1,
  4. F Cantini2,
  5. M Di Natale2,
  6. R Livi1,
  7. A Tempestini1,
  8. M Matucci-Cerinic1
  1. 1Department of Medicine, Rheumatology Division, University of Florence, Italy
  2. 22nd Division of Medicine-Rheumatology Unit, Ospedale di Prato, Italy
  1. Correspondence to:
    Dr F Perfetto
    Department of Medicine, Rheumatology Division, University of Florence, Viale Pieraccini 18, 50139 Florence, Italy;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The aetiology of polymyalgia rheumatica (PMR) is unknown, but its sudden onset and the wide variation in incidence reported from various parts of the world suggest the existence of an environmental agent or genetic factors, or both.1,2 Interestingly, epidemiological studies showed a regular cyclical pattern of incidence of PMR,1–,3 whereas others failed to confirm any seasonal trend.4–,6

To evaluate the seasonal distribution in the onset of symptoms of PMR, this retrospective study considered 201 patients (mean age 72.6 years; 126 women, 75 men) with a diagnosis of PMR,7 seen by rheumatology units of the Prato (125 patients) and Florence Hospitals (76 patients), Italy, from January 1992 to December 2002. All patients were identified by searching hospital (74 patients) or ambulatory records (127 patients). In 21 patients, PMR was associated with giant cell arteritis (GCA).

Differences between seasonal incidence were assessed by the χ2 test and by the single cosinor analysis. The parameters calculated with the cosinor procedure were the MESOR (the rhythm adjusted mean over the time period analysed), amplitude (half the distance between the absolute maximum and minimum of the fitted curve), and acrophase (peak time of rhythmic change). Significance was defined as p<0.05.

In 71 patients symptoms started in winter, 49 in spring, 33 in summer, and 48 in autumn. The χ2 test showed that the frequency of onset of PMR was significantly (p<0.001) increased in winter for all the patients and also for subsets by sex (men, p<0.01; women, p<0.04) and age (age <75, p<0.001; age >75, p>0.001). Rhythm analysis (table 1) identified a circannual variation in the occurrence of PMR, with a significant (p<0.001) peak in January. A similar circannual pattern was also identified for subsets by age (age <75, p<0.001; age >75, p<0.01) and sex (p = 0.03) (table 1). No rhythm was detected in patients with PMR/GCA.

Table 1

 Circannual rhythm parameters of single cosinor analysis fitted to monthly data of occurrence of PMR/GCA

This study shows a significant winter periodicity in the onset of PMR. These results are in agreement with many previous reports1–,3 but contrast with others4–,6 that failed to demonstrate seasonality. Differences in the composition of the group (such as high proportion of patients with PMR/GCA), size of the group, geographical location (northern versus southern Europe), and statistical procedure may explain these discrepancies. Unlike this study, previous reports checked the statistical significance of the time effect on the onset of symptoms only by the χ2 test. That test, however, only indicates whether the counts are different in a different season of the year without necessarily implying the presence of a circannual cycle, and it gives no information as to its timing. On the contrary, the cosinor method provides a better chronobiological analysis and yields rhythmic characteristics such as the MESOR, amplitude, and acrophase, each with its variance estimate.

Two potentially synergistic mechanisms may have a role in the seasonality of PMR. Firstly, the infectious-like onset of PMR suggests a precipitating environmental factor as one of the causes, and a close temporal relationship between epidemics of Mycoplasma pneumoniae, Chlamydia pneumoniae, and parvovirus B19 and peaks of cases of PMR/GCA has been reported.1,2 However, reports of material with an infectious origin in temporal arteries have not been corroborated,8,9 and Peris found no relationship between the onset of PMR and parvovirus B19 infection.6 Secondly, some authors have reported seasonal variation of the immune system and inflammatory responsiveness.10 These seasonal changes, when associated with several unknown infectious agents, may increase the susceptibility to develop some diseases, including PMR. Further longitudinal studies conducted on larger samples and in other countries at different latitudes are needed.