Seasonal variations in hospital admissions for mania: Examining for associations with weather variables over time

Highlights

• This register-based nationwide study correlated weather and admissions with mania.

• Admissions with mania followed a seasonal pattern with peaks in summer.

• Admission rates correlated with sunshine, ultraviolet dose, temperature and snow cover.

• The seasonal pattern in admissions was unaffected by changes in climate variables.

Abstract

Background

Bipolar disorder is characterized by a seasonal pattern with emerging evidence that weather conditions may trigger symptoms. Thus, our aims were to investigate if year-to-year variations in admissions with mania correlated with year-to-year variations in key meteorological variables, if there was a secular trend in light of climate change and if gender or admission status influenced the seasonal pattern.

Methods

We undertook a Danish register-based nationwide historical cohort study. We included all adults hospitalized to psychiatric care from 1995 to 2012 with mania using the Danish Psychiatric Central Research Register. The Danish Meteorological Institute provided the meteorological variables. The association between weather and admissions was tested using linear regression.

Results

Our database comprised 24,313 admissions with mania. There was a seasonal pattern with admission rates peaking in summer. Higher admission rates were associated with more sunshine, more ultraviolet radiation, higher temperature and less snow but were unassociated with rainfall. We did not find a secular trend in the seasonal pattern. Finally, neither gender nor admission status impacted on the overall seasonal pattern of admissions with mania.

Limitations

Only patients in psychiatric care were included. We could not subdivide by type of bipolar disorder.

Conclusion

This cohort study based on more than 24,000 admissions identified a distinct seasonal pattern in hospital admissions for those with mania. We found no secular trend. This could indicate that the climate change is not impacting on seasonal patterns, that there is no link between the proposed variables or that change is currently not sufficiently distinctive.

Introduction

Bipolar disorder exhibits a seasonal pattern in that, in the majority of studies in most non-equatorial regions, admissions for mania peak in spring and summer whereas admissions for depression peak in autumn and winter (Geoffroy et al., 2014). Females have been identified as more subject to seasonal variation in some studies than males (Kerr-Correa, 1998; Suhail, 1998). Partonen and Lönnqvist (1996) found that the likelihood of depression was highest in autumn for first-ever admissions but not readmissions: whether admission status (first-ever admissions vs. readmissions) also influence the pattern of seasonality in admissions with mania is unclear. The seasonal pattern has long been attributed to meteorological variables, especially the impact of light. One southern hemisphere study demonstrated that the peak in spring correlated with the rapid increase in hours of sunshine or luminance at that time, allowing the authors to hypothesize that such a change in luminance stimulated the pineal gland and so contributed to the seasonal increase in admissions of those with mania (Parker and Walter, 1982). Patients with bipolar disorder experience a disrupted circadian rhythm with seasonal changes effecting sleep efficiency and mood (Hakkarainen et al., 2003, McKenna et al., 2014). Short light wave lengths suppress melatonin, a regulator of circadian rhythm, and supersensitivity to light is a suggested trait marker for bipolar disorder (Lewy et al., 1985, Nathan et al., 1999, Wright and Lack, 2001). Thus, increased exposure to light during spring and summer might increase the risk of a manic episode through suppression of melatonin, although the impact of other meteorological variables needs to be conceded.

In a review by Geoffroy et al. (2014), the authors called for large-scale studies exploring the influence of varying meteorological variables (e.g. sunlight, ultraviolet radiation, rain, snow cover and temperature) on admission rates of those with bipolar disorder. The association of rain and temperature with admissions for those with mania remains controversial (Carney et al., 1988, Lee et al., 2002, Lee et al., 2007, Mawson and Smith, 1981, McWilliams et al., 2014, Myers and Davies, 1978, Sayer et al., 1991; Suhail, 1998; Volpe et al., 2010, Volpe and Porto, 2006). Some studies find that increased rain is associated with a decrease in admissions (Lee et al., 2007, Mawson and Smith, 1981, Volpe et al., 2010, Volpe and Porto, 2006) and higher temperature is associated with an increase in admissions (Myers and Davies, 1978, Sayer et al., 1991). The causal link remains unclear. Higher temperature could affect the individual's ability to initiate and maintain sleep which would lead to a disrupted circadian rhythm (Pandey et al., 2005). Rain could be associated with less sunshine, a behavioral change towards staying indoors or have a biological impact. Snow cover has not been studied. Snow would only be prevalent in winter, however, as snow increases, the reflection of light could have an impact on the seasonal variation in admissions for mania.

If seasonality of admissions for mania is influenced by increases in luminance or temperature, we would anticipate a significant increase in admissions for mania during warmer months and/or during sunnier summers. Likewise, we expect that changes in climate might impact on admission rates. As ‘climate change’ (or warming of the earth) is increasingly recognized as a distinct phenomenon with adverse health effects (McMichael et al., 2006), it raises the question as to whether such a change has had an impact on seasonal admission patterns for bipolar disorder.

The aim of this nationwide 18-year historical cohort study was to investigate three hypotheses. Firstly, that year-to-year variations in admissions with mania correlate with year-to-year variation in sunlight, ultraviolet radiation, temperature, rain and snow cover. Secondly, if there has been any secular change over time in admissions if it was associated with climate change variables. Thirdly, to determine if gender or admission status (first-ever/readmission) influence the seasonal pattern.