In our study, we found a seasonal change in IOP of approximately 0.4 mm Hg between winter and summer months with higher IOP during winter. Daily mean temperature and age directly correlated significantly with IOP in our multiple linear regression model, while daily sunshine duration did not correlate directly with IOP. Multiple linear regression analysis with total sunshine duration of the ten preceding days before IOP measurement showed a significant correlation of sunshine duration and IOP.
Our major observation is the global tendency towards lower IOP measurements during summer months in comparison to winter months in a city in the northern hemisphere with continental climate. This association is well established and several hypotheses of physiological mechanisms for this annual fluctuation in IOP have been postulated. These include ambient temperature, atmospheric pressure [18], sunshine duration [19], melatonin secretion [20, 21] and tear break-up-time [22].
One interventional prospective study with a small cohort examined the correlation of temperature and IOP, but saw only a modest increase in IOP in non-acclimated subjects after exposition to 40°C and 50% humidity at 3 hours and 15 minutes compared to acclimated subjects and subjects at 20°C and 50% humidity, which is inverse to the correlation that we observed [23]. A more recent study including 13 young healthy individuals showed an inverse correlation between IOP and environmental temperature and sunshine duration [19]. Our data support the hypothesis that higher temperature is associated with lower IOP. A potential explanation for the diverging findings could be that higher temperatures have different short- and long-term effects. A potential mechanism by which temperature changes lead to changes in IOP is cold-induced vasoconstriction [24]. This effect might also affect episcleral vessels [25]. The aqueous humor is eventually drained into the episcleral veins. Changes in episcleral venous pressure induced by cold might lead to higher IOP. Vice versa, increased arterial blood flow caused by vasodilation induced by higher temperature could potentially result in increased aqueous humor secretion, although experiments, supporting this assumption, are missing.
We also investigated the effect of sunshine duration on IOP. However, separation of the effect of sunshine duration and temperature is impeded because these variables are intercorrelated [26]. Interpretation of our data should also consider that sunlight exposure and sunshine duration need to be distinguished. Sunshine exposure in winter months might be low even at long sunshine duration due to protective clothing and less time spent outdoors compared to summer. It also remains unclear whether light exposure of the eye or on the skin is relevant for the changes in IOP. We have no data regarding the real individual sunlight exposure of the patients in our retrospective study population. Therefore, validity of our findings is limited. Nevertheless, we think that on a greater perspective it is reasonable to assume that sunlight exposure is higher in summer months than in winter months. The melatonin pathway is a potential explanation by which light exposure leads to changes in IOP [20, 27]. The mechanism by which this occurs appears to involve coupling of the Melatonin 3 receptor with the sympathetic component that controls aqueous humor synthesis and efflux [28]. Melatonin and its analogs have been shown to modulate the α2 adrenoreceptors [29]. Melatonin levels could not be evaluated in this study. If the melatonin pathway is the intraocular transducer of sunlight exposure to changes in IOP can not be answered from this study. Additionally, the effect strength of sunlight exposure remains elusive. Still our data support further investigation of sunlight and its effects on IOP.
In our study, no significant correlation of daily mean atmospheric pressure was observed, even though it had been described previously in a cohort of 109 patients [30]. Interpretation of data regarding IOP and atmospheric pressure has to take into account that IOP is actually the trans-corneal pressure difference. The true pressure in the eye with a measured IOP of 15 mm Hg is 783 mm Hg [768 mm Hg (atmospheric pressure equivalent to 1024 hPa) plus 15 mm Hg] as 1 hPa is equivalent to 0.75 mm Hg. Accordingly a difference of atmospheric pressure of 5 hPa between January and June is equivalent to approximately 3.75 mm Hg. At a stable true IOP the higher atmospheric pressure during winter would reduce the measured trans-corneal pressure. This is contrary to the observation in this study with higher IOP during winter months. Atmospheric pressure can change by 7–8 hPa within 24 hours. Accordingly, mean daily atmospheric pressure is potentially not precise enough to detect changes in measured IOP caused by atmospheric pressure changes. Our findings do not rule out effects of atmospheric pressure on IOP. However, atmospheric pressure is unlikely to be the main factor contributing to seasonal IOP changes.
Additionally to the factors contributing to seasonal IOP changes investigated in this study, other factors should be discussed. Variation of blood pressure is a potential factor contributing to seasonal IOP variation. Systolic and diastolic blood pressure present seasonal variations with higher pressure in winter than in summer [31–35]. Positive correlation between blood pressure and IOP has been reported previously [36–38]. Mitchell et al. reported an increase by 0.28 mm Hg (95% confidence interval, 0.23–0.34) for each 10-mm Hg increase in systolic blood pressure [36]. Modesti et al. reported a blood-pressure decrease by 0.15 mm Hg to 0.19 mm Hg / − 1°C mean 24 h outdoor temperature measured by the local meteorological office [35]. In summary, the expected temperature-dependent and seasonal difference in blood pressure contributes to lower intraocular pressure values in the summer months but is too weak to explain the seasonal differences in intraocular pressure that we have seen.
We hypothesized that older patients might be more susceptible to seasonal weather changes, especially because of higher likelihood of dehydration, and thus show a greater variation in IOP [39–42]. Low hydration has been linked to decreased IOP [42]. Indeed, we observed greater IOP variations in patients older than 30 years. However, the young study population was considerably smaller (n = 14,389) compared to the study population of 31–60 year old patients (n = 60,553) and 61 years or older patients (n = 124,805). The less clear yearly variation in IOP in the young study population might be explained by the weaker statistical power and is not necessarily caused by lower IOP variation. This finding might still be in contrast to the fact that glaucoma and ocular hypertension are mostly diseases of the elder.
Even though the magnitude of time-of-year IOP variation was modest, it should still remain of scientific interest. Potentially, some clinical consideration can also be drawn from this study. It should be kept in mind that the measured seasonal change represents a mean variation of hundreds to thousands of measurements, which could be larger in some individual patients and for these patients be a crucial factor for diagnostic and therapeutic decisions. It has been demonstrated that low summer IOP is beneficial for the progression of retinal nerve fiber layer thinning [14]. Higher IOP fluctuation has also been described as a risk factor of glaucoma progression [43]. On an individual level, potential seasonal effects on IOP should be taken into account, when evaluating IOP data and treatment effects over time. Exemplarily, a patient with high seasonal IOP fluctuations might show normal IOP at yearly check-ups in summer, but show hypertensive IOP during winter. On the other hand, patients starting treatment for hypertensive IOP during spring might show a good treatment effect that is caused by the seasonal IOP change and not by the prescribed medication. Also deprescribing of antihypertensive eye drops during summer months in patients with proved relevant seasonal IOP-changes might be considered. Because of the small magnitude of seasonal changes in IOP of far less than 1 mm Hg we assume that these changes are not of high relevance in most clinical cases. The widely studied 24 hour IOP fluctuation has higher absolute changes in IOP and is likely of greater relevance in the diagnostic and therapeutic process in glaucoma care than the comparatively small seasonal IOP changes [44]. Nevertheless, this study offers further insight into physiological IOP regulation and factors influencing this process that need to be evaluated with greater precision in future studies.
Future studies should document, on the individual level, sunlight exposure and the temperatures to which a person is exposed. From this, the mode of action of sunlight and temperature could possibly be better distinguished. Further insights into the mechanism of action of seasonal IOP variations would potentially also allow new therapeutic approaches. For example, Melatonin and its analogues have already shown promising IOP-lowering effects in small studies [45–47].
This study has several limitations. Reports on mean daily sunshine duration, mean daily temperature and mean daily atmospheric pressure of the weather station located closest to the clinic were correlated to the IOP of the patients. Patients are likely to be exposed to at least partially differing weather because they live in different areas and spend more or less time outdoors. Thus, the weather reports are only an approximation to the weather experienced by the patients. Future prospective studies should track the individual sunshine, temperature and atmospheric pressure the patients are exposed to. Nevertheless, in the retrospective and anonymized setting of our analysis the presented weather data are the best available approximation of weather exposure.
Another limitation of this study is the single facility patient population, which might not be representative of other populations. Other locations with differing climatic conditions might present different or no cyclical yearly IOP variations. To our knowledge, all studies published regarding seasonal IOP variations have been conducted at climatically similar locations in the northern hemisphere. Therefore, it would be interesting to examine the correlation of weather variables and IOP in climatic differing locations and in the southern hemisphere.
Due to the retrospective cross-sectional anonymous design a selection bias could potentially confound our findings. Repetitive measurements of the same eye at different time points cannot be analyzed, which weakens the conclusions that can be drawn from this study. No information about the proportion of patients who had repetitive IOP-measurements over the course of the investigated 6-year-period is available. Therefore, IOP variation of individual subjects over time cannot be analyzed and no statements about seasonal variations on an individual level can be made. However, this does not invalidate the observation of seasonal IOP changes. Furthermore, our approach including a comparably high number of subjects reduces the risk of diurnal IOP changes of several mm Hg overlaying the subtle seasonal IOP changes.
The spectrum of patients should be identical all over the year because we have prespecified recurrent appointment lots for all subspecialities that are always fully booked. We do not believe that varying percentages of glaucoma patients contribute to our findings. As stated in the methods section, glaucoma patients at our clinic are mostly solely measured by GAT and are therefore not included in our study to any relevant extent. Furthermore, values above 30 mm Hg were excluded in our analysis. However, numbers proofing our assumptions can not be retrieved from our dataset.
Interpretation of the reported absolute IOP should take into account that IOP was not measured by GAT but by non-contact tonometry, which has been reported to differ from GAT [48]. While absolute values differ, a significant positive correlation between NCT and GAT measurements has been reported [49]. Although GAT beeing the gold standard of tonometry, it is potentially more susceptible to inter-rater-variance than automated NCT [50]. As seen in the Fourier Analysis and Figure S1 the seasonal IOP pattern is recurring yearly and the results are highly significant. Unreliable measurements would rather disguise these minimal seasonal IOP changes. However, it should be noted that the absolute values of the NCT were not compared with GAT and therefore the absolute values may potentially show a systematic deviation from IOP values determined by GAT [51].
Our dataset lacks information regarding eye disease, previous ocular surgeries, use of eye drops, use of medication and general diseases. Data regarding central corneal thickness are missing, too. These factors are known to potentially affect IOP and measurement of IOP at a significant level. In our study, these factors contribute to high noise in IOP-measurements in our patient group, especially as we included patients with ocular diseases and not a healthy population. However, we have no reason to assume that the other factors would cause a seasonal IOP variation, which could be clearly demonstrated in our Fourier analysis, except potentially increased use of inhaled glucocorticoids during winter months with possible effects on IOP [52]. Increased use of glucocorticoid eye drops during spring and early summer for treatment of allergic conjunctivitis could also potentially influence IOP [53].