Background People with chronic pain commonly believe that their pain is affected by the weather. Despite a century's worth of research, there is no scientific consensus on the existence of a relationship between weather and chronic pain.
Objectives A systematic literature review to (1) gain an overview of existing research on the weather-pain relationship, and (2) summarise the methodologies, methodological rigour and risk of bias in published studies of patients with musculoskeletal conditions.
Methods Systematic search strategies were developed for 5 databases (Medline, Embase, PsycInfo, Scopus, Web of Science) to retrieve studies that investigated the relationship between weather conditions and chronic pain. Original articles describing observational studies that related chronic pain (primary outcome) to weather conditions (exposure), were included. Study characteristics, methodology (e.g. measurements of pain and weather, aspects of statistical analysis) and data on risk of bias (validation of participants' exposure and diagnosis, coverage of weather variation, correction for confounders) was extracted. Methodological rigour was summarised by mapping the methodological variability among studies and ranking these from least to most rigorous.
Results The searches returned 16,081 articles. After removing 3090 duplicates and excluding 12726 articles during title screening, 265 abstracts were assessed for eligibility. Of 64 observational studies that met the inclusion criteria, 30 (47%) investigated pain associated with musculoskeletal conditions, 24 (38%) investigated headache and 10 (16%) investigated pain associated with other conditions such as sickle cell disease and dental pain. After full text assessment, 10 of 30 papers on musculoskeletal conditions were excluded because they investigated risk of acute pain episodes rather than chronic pain symptoms (n=6), or investigated multiple conditions (n=4). The 20 included studies investigated rheumatoid arthritis (6, 30%), osteoarthritis (6, 30%), fibromyalgia (6, 30%) and low back pain (2, 10%). A total of 15 studies (75%) reported some effect of weather on chronic pain, and this was clinically significant (consistent and sufficient size of effect) in 6. Participant numbers varied from 19 to 2491. 1 study was cross-sectional and 19 longitudinal with follow ups ranging from two weeks to 3.5 years, with participants scoring their pain once to daily for 12 months. Pain was measured with a Visual Analog Scale or Numeric Rating Scale in 17 (85%). Weather conditions were retrieved from local weather stations, with most studies assuming that participants stayed in the area/city (65%) or postcode area (20%) where they lived.
In 12 (60%) studies, participants were blinded to the study hypothesis. Methods for correcting for confounders were highly variable, with 7 studies not addressing any, and the remainder addressing one or more of 32 confounding variables.
Conclusions Methodological variability of studies investigating the relationship between weather and chronic musculoskeletal pain is high. This methodological review will inform best research practice for those investigating the relationship between the weather and chronic pain.
Acknowledgements Mary Ingram, librarian
Disclosure of Interest None declared