Statistics from Altmetric.com
Posterior vitreous detachment (PVD) is a common condition after the fifth decade of life.1 Synchisis of the vitreous progresses in proportion to age, creating holes in the posterior hyaloid membrane and allowing PVD to occur.2 The incidence of rhegmatogenous retinal detachment is generally accepted to be season dependent, with a higher incidence in the summer months.3,4 However, there is no evidence that PVD is season dependent. We performed this study to determine whether there is any correlation between ambient temperatures, humidity, or solar radiance on the incidence of PVD.
Patients were selected for this study from the eye casualty database at Oxford Eye Hospital. All patient records, which have been diagnostically coded as acute PVD over a 2 year period, were reviewed. Cases where there was a precipitating cause for PVD such as blunt trauma, retinal vascular disease, diabetic retinopathy, previous surgery, or laser treatment were excluded. Only cases with spontaneous PVD were included in the study.
We used three environmental variables—mean daily temperatures in degrees centigrade (°C), percentage relative humidity (both averaged over measurements taken every 10 seconds), and total solar radiation in the day (0.3–3.0 nm) measured in mega joules per square metre (MJ/m2). These data were obtained from the environmental change network site at Wytham, Oxford, UK.5 In order to reduce short term variation, we aggregated the occurrences of PVD cases and thus considered the number of cases observed in each week. We also averaged the daily data for environmental variables over each week.
We considered 567 cases with a mean age of 63 years (range 36–89 years). Of these, 319 (53%) were referred by general practitioner, 170 (30%) were self referred, 68 (11.4%) were optician referrals, and the remaining 10 (1.6%) were referred from another hospital. The average number of new patients attending the eye casualty department at Oxford was 876 per month (range 763–965). The average total number of patients attending the eye casualty department during summer months (April to September) was 912 compared to 839 in the colder months (October to March). Even though there were more visitors in the warmer months compared to the colder months, this difference was not significant (Wilcoxon-Mann-Whitney test p value = 0.1). We also fitted a sine/cosine linear model to test for a seasonal pattern, but did not find it supportive (p = 0.32). We also examined the effect of a cyclical pattern for the weekly or monthly occurrences—for instance, the numbers of cases for June, July, and August were 60, 42, and 44 respectively; for December, January, and February they were 27, 48, and 49 respectively. Thus, we proceeded to model the number of PVD cases directly as a function of environmental variables.
There was a significant increase in the average number of occurrences of PVD from 4.23 to 5.49 cases per week between 1996 and 1997 (t test, p=0.0019), so we included year in all models (Fig 1).
We first fitted a model including relative humidity, air temperature, solar radiation, and year. There was evidence of a strong association (p=0.035) between weekly averaged air temperature and the number of occurrences of PVD (Fig 1). However, we did not find a significant association between weekly occurrences of PVD and relative humidity or solar radiation.
We found a significant increase in the variance of the number of occurrences of PVD with increasing weekly temperatures. Accordingly, we used the natural logarithm of occurrences as our response variable; this transformation stabilised the variances. We then fitted a generalised linear model using air temperatures and year as explanatory variables. Both had a significant and direct effect on the log number of occurrences; the regression coefficients were 0.118 (95% confidence interval = 0.022, 0.214) for year and 0.0189 (confidence interval = 0.0015, 0.036) for air temperature. This means that the weekly number of PVD cases increased faster for higher temperatures—for example, for an increase from 5–6 degrees in average air temperature, the rate of increase in weekly number of PVD cases was 0.09 per week and for a temperature change from 20 to 21 it was 0.12. We did not find any evidence of an interaction between air temperature and year; thus, the effect of year was only additive, and the rate of change of log occurrences with respect to air temperatures was the same for both years.
We also found a strong association (p=0.028) between the air temperatures of the previous week and the number of weekly occurrences and the results were very similar to the same week.
Vitreous liquefaction, which advances with increasing age, is an important event in the pathogenesis of PVD.6
In our study there was no evidence of a cyclical pattern for the weekly or monthly occurrences, therefore we modelled the number of PVD cases directly as a function of the air temperature. Our data suggest a highly significant correlation between weekly average temperatures and the incidence of PVD.
The lack of any previous literature on this subject makes it difficult to draw any conclusions about the mechanism for an increase in PVD occurrences with increasing temperatures. However, we postulate that increasing physical activity and dehydration associated with increasing temperatures may have a role to play.
Further work is necessary in order to investigate the effect of increasing temperatures and/or dehydration on the biochemical structure of the vitreous.