The association between apparent temperature and psoriasis outpatient visits: a time-series study in Hefei, China

Temperature may be an important environmental factor affecting psoriasis. This study aimed to determine the potential association between apparent temperature (AT) and outpatient visits for psoriasis in Hefei, China. Daily psoriasis clinic visits in Hefei were collected from January 1, 2016 to December 31, 2020. A Poisson generalized linear regression model (PGLM) combined with a distributional lagged nonlinear model (DLNM) was used to analyze the impact of AT on psoriasis outpatient visits; the model was adjusted for relative humidity, wind speed, precipitation, PM2.5, NO2, SO2, time trends, Sundays, and holidays. The analyses were stratified by age and sex. A total of 24,351 patients with psoriasis were included in this study. Only a low AT showed a delayed and strong risk effect. Referring to the median AT (16.8 °C), the effect of low AT started at lag 2 days and showed an increasing and then decreasing trend for approximately 5 days; the effect of extreme cold (1st percentile) and cold (5th percentile) peaked at lag 4 days with a relative risk (RR) of 1.117 (95% CI:1.065 to 1.171) and 1.081 (95% CI:1.044 to 1.119), respectively. The effect of mild cold (25th percentile) reached a maximum RR of 1.033 (95% CI:1.017 to 1.048) at lag 6 days. Subgroup analysis showed that low AT risk was more pronounced and longer-lasting in men and individuals aged <45 years. Our study provides evidence that a low AT increases the risk of psoriasis. Men and young people are vulnerable to potential adverse effects. There is a need for enhanced health interventions, medical care, and early warnings for patients.


Introduction
Psoriasis is a common chronic inflammatory disease mediated by systemic immunity that affects 2%-3% of the general population [1]. Patients with this disease show overgrowth and abnormal differentiation of keratinforming cells, and the skin appears as well-defined red circular plaques covered with silvery scales [2][3][4]. Psoriasis causes recurrent itching and pain, which can lead to a reduced quality of life and psychological disorders [5]. Globally, the prevalence of psoriasis increased between 1990 and 2017, and the psychological and social problems associated with psoriasis generate high medical costs [6]. Given the large population of China, psoriasis is becoming a major public health concern.
The development of psoriasis is often thought to result from a combination of genetic susceptibility, immune system disorders, and environmental factors [7], with environmental factors playing a vital role in the development of psoriasis [8]. Previous studies have shown that the environmental risk factors affecting psoriasis are ultraviolet (UV) light, smoking, obesity, infections, alcohol consumption, vitamin D deficiency, and stress [8,9]. In addition, extensive research has shown that outdoor air pollutants are harmful to patients with psoriasis [10][11][12]. In recent years, a growing number of clinical and epidemiological studies have shown an apparent seasonal variation in the onset of psoriasis, with a trend toward winter flaring and summer clearing [13][14][15]. Given this seasonality, we presume that the risk posed by ambient temperature is not negligible.
However, there is insufficient evidence demonstrating an association between ambient temperature and psoriasis. Although there is growing evidence that environmental temperatures, including high and low temperatures, play an important role in the incidence of skin diseases, most of these studies have focused on eczema [16][17][18], acne [19], and atopic dermatitis [20,21], and few have focused on psoriasis. The effect of ambient temperature on health is influenced by other weather factors including humidity, wind speed, and barometric pressure. In addition, both the World Meteorological Organization and Health Organization recommend the use of biometeorological indicators to assess the effects of temperature on human health [22]. The apparent temperature (AT) is a composite indicator that combines ambient temperature, wind speed, and humidity and better describes the actual human perception of temperature in the environment [23]. Therefore, we used the AT metric in this study to assess the effect of temperature on psoriasis.
Considering that climate change will lead to increasingly frequent and severe extreme weather events, climate change over a relatively short period can pose a serious threat to human health. Therefore, this study aimed to investigate the relationship between AT and psoriasis outpatient visits in Hefei, China through a timeseries analysis to provide a scientific basis for psoriasis prevention and intervention.

Study area
Hefei is located in East China (31°52′N, 117°17′E) and is a suburban city of the Yangtze River Delta urban agglomeration in China. The city has a typical subtropical, monsoonal, humid climate with four distinct seasons and an average temperature of 15.7°C. The resident population At the end of 2020 was 9.37 million.

Information on psoriasis visits
We retrieved psoriasis visits to the First Affiliated Hospital of Anhui Medical University from January 1, 2016 to December 31, 2020, using the outpatient electronic medical record system, including the visit date, patient age, sex, and address of residence. The hospital's dermatology department has an annual outpatient volume of over 500,000, making it the largest dermatology center in the Anhui Province in terms of outpatient volume. Our physicians completed the psoriasis diagnosis and coded it through the International Classification of Diseases, 10th edition (ICD-10 codes: L40-L45) and made notes on the diagnostic information through the outpatient electronic medical record system to distinguish whether the diagnosis was suspected or not. Based on outpatient medical record information, we determined whether the condition was a short-term exacerbation of symptoms or a new onset of symptoms. We also determined whether repeated outpatient visits were based on the chief complaint and the current medical history in the outpatient medical records. The specific criteria were as follows: The inclusion criteria included (1) confirmed diagnosis of psoriasis, (2) those who visited the clinic for new onset of related symptoms or aggravation of related symptoms, (3) complete basic information, and (4) those whose current place of residence was Hefei City. The exclusion criteria were (1) patients with suspected diagnosis in the remarks, (2) patients with missing, important, personal information, (3) those who only visited outpatient clinics for medication or repeat visits as recommended by physicians, and (4) those whose place of residence is outside Hefei City.
The hospital outpatient department regularly conducts random checks on outpatient electronic medical records to ensure the authenticity of outpatient medical records to a certain extent. In addition, after we collected the medical record information, we double-checked the screened data. For first-time patients, we screened patients with a confirmed diagnosis. For repeat patients, information on symptoms noted by the physician was used to determine whether the condition had worsened in the short term. The number of patients with missing, important, personal information was very small, and the impact on the overall results was negligible based on the large sample size. This study was reviewed and approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University (PJ2022-11-31).

Meteorological and air pollutants information
The Anhui Provincial Meteorological Bureau provided data on ambient temperature (°C), wind speed (m/s), relative humidity (%), 24-hour rainfall (mm), and Sunshine duration (h) for the study period. Daily air pollutant concentrations were obtained from 10 Hefei fixed air quality monitoring stations, including nitrogen dioxide (NO 2 ), sulfur dioxide (SO 2 ), carbon monoxide (CO), ozone (O 3 ), and fine particulate matter (PM) with a median aerometric diameter of < 2.5 μm (PM 2.5 ) and < 10 μm (PM 10 ). Missing variable information was replaced with the monthly averages.

Calculation of AT
The formula for calculating AT is as follows [24,25]: We first calculated ρ (water vapor pressure) from equation (2), where T denotes the mean ambient temperature and RH denotes the relative humidity. Then, we calculated AT from equation (1), where WS denotes the wind speed.

Statistical analysis
Considering the Poisson distribution and overdispersion of the daily number of psoriasis visits, the influence of ambient temperature on health exhibited delayed and non-linear effects [26,27]. Therefore, we used a Poisson generalized linear regression model (PGLM) combined with a distributed lagged nonlinear model (DLNM) to examine the relationship between AT and psoriasis visits. To avoid multicollinearity between the variables, Spearman correlation analysis of meteorology and air pollutants was conducted, as shown in Supplementary figure S1 in the appendix. Covariates with Spearman correlation coefficients <0.7 were selected for inclusion in the model, including AT, relative humidity (RH), wind speed, 24 h rainfall, Sunshine duration (SD), PM 2.5 , NO 2 , and SO 2 . The specific model structure is as follows:  In the above equation, the subscripts t and l denote the date of visit and lag days, respectively; Y t represents the expected number of visits for psoriasis on day t; ns() is a natural spline function; cb(AT t,l ) is the crossed basis function obtained via DLNM; RH t,l , WS t,l , Rainfall t,l , SD t,l , PM 2.5t,l , NO 2t,l , and SO 2t,l are natural cubic spline functions of AT, RH, wind speed, 24h rainfall, Sunshine duration, PM 2.5 , NO 2 , and SO 2 , respectively; β, γ, and δ are Dow t , Holiday t , and AT t,l , respectively; Dow indicates the day of the week; and Holiday is a binary variable representing the holiday. The degrees of freedom for meteorological factors and air pollutants are 3 per year [28], and the degrees of freedom for seasonal and long-term trends are 7 per year [16]. Extreme cold (first percentile), cold (fifth percentile), and mild cold (twenty-fifth percentile) were defined according to the previous literature [29,30]. The relative risks (RR) and lagged effects of extreme cold and mild cold were assessed separately, using the median apparent temperature as a reference [31]. We did not evaluate the RR of high AT because the overall exposure-response analysis revealed that the impact of high AT on psoriasis was insignificant. In addition, we conducted sensitivity analyses by varying time (6-8 df/year), meteorological variables (3-5 df/year), and air pollutants (3-5 df/year).
The study was statistically analyzed using the DLNM and Splines packages in R software (version 4.1.1). Statistical significance was set at P < 0.05.

Descriptive statistics
During the 5-year study period, 24,351 patients with psoriasis presented with an average daily attendance of 13.3. A total of 14,341 (58.89%) patients with psoriasis were men. Of the patients with psoriasis, 64.69% were aged 0-44 years, which was significantly higher than those in the 45+ age group. The mean annual AT was 20.0°C, which was slightly higher than the mean ambient temperature (16.8°C). Table 1 shows the descriptive statistics for psoriasis visits, meteorological variables, and air pollutants. Figure 1 shows the overall exposure-response association and the lagged effect of all psoriasis visits with daily AT. Figure 1(a) shows the general trend in the association between AT and psoriasis at 7 lags (days) in the threedimensional model, with a lag effect as the risk effect of AT on psoriasis tends to increase and then decay as the number of lag days increases. Figure 1(b) shows the impact of different ATs on psoriasis risk. Low AT significantly increased the risk of psoriasis when the AT was below 16.8°C. However, high AT appeared to reduce the risk of psoriasis, an effect that varied relatively flatly and not significantly with AT.   Note: P1, P5, P25, P50, P75, P95, P99: the 5th percentile, the 25th percentile, the 50th percentile, the 75th percentile, the 95th percentile.

Overall effects of AT on psoriasis clinic visits
We analyzed the overall exposure-response association between AT and psoriasis risk for each subgroup stratified by sex and age, as shown in figure 2. The overall exposure-response relationship showed similar trends across subgroups. However, we observed a more pronounced effect of low AT on the risk of psoriasis development in the male group and 0-44 years age group. Table 2 and figure 3 show the single-day lag effects at the 1st (−5.7°C), 5th (−1.6°C), and 25th (6°C) percentile AT for the total population and each subgroup, using the median AT (16.8°C) as a reference. Low AT demonstrated a delayed and significant risk impact. The lower the AT, the greater the impact of the risk posed, and the longer it lasted. For the total population, the impact of low AT started at lag 2 days and lasted approximately 5 days; the effects of extreme cold (1st percentile) and cold (5th percentile) showed a gradually increasing and then decreasing trend, reaching a maximum RR at lag 4 days with 1.117 (95%CI:1.065 to 1.171) and 1.081 (95%CI:1.044 to 1.119), respectively, while mild cold (25th percentile ) reached a maximum RR of 1.040 (95%CI:1.010-1.070) at lag 5 days. Depending on sex and age, the effects of low AT manifested differently. For the male subgroup, the impacts of extreme cold, cold, and mild cold all started at lag 2 days, reached a maximum at lag 4 days, and lasted for approximately 5 days. In the female subgroup, the effects of extreme cold and cold occurred only on the two days of lag 5 and lag 6, and the impact of mild cold was not significant. In terms of age, the effects of extreme cold and cold for the 0-44 years age group showed the same trend as that for the male group; the effect of mild cold started at lag 2 days, reached its maximum at lag 4 days, and lasted for approximately 5 days. In the 45 years age group, the effects of extreme cold and cold started at a lag 4 days and reached their maximum RRs, lasting just 2 days. The effects of mild cold appeared at a lag 5 days.

Lagged effects of specific low ATs on psoriasis visits
The cumulative lag effects at the 1st (−5.7°C), 5th (−1.6°C), and 25th (6°C) percentile AT were significantly higher than the single-day lag effects for the total population and each subgroup and showed a progressive increase over time. For the total population, males, and the 0-44 years age group, the risk effects on specific ATs all occurred at lag 0-3 days and lasted until lag 0-7 days. In the female subgroup, the risk effects of extreme cold, cold, and mild cold were only present at lag 0-6 days and lag 0-7 days. In the 45 years age group, the cumulative risk effect of extreme cold and cold was insignificant, and the cumulative risk effect of mild cold appeared at lag 0-7 days. The details are presented in table 3 and figure 4.

Sensitivity analysis
Sensitivity analyses showed that when adjusting the model for temporal trends (6-8 df/year), meteorological variables (3-5 df/year), and degrees of freedom for air pollutants (3-5 df/year), the exposure-response curves  before and after adjustment were similar, so our analysis was robust. See Supplementary figure S2-9 in the appendix for further details.

Discussion
To our knowledge, this is the first study to examine the relationship between AT and psoriasis outpatient visits. We found a significantly increased risk of psoriasis when AT was below 16.8°C and a lag effect, whereas higher AT did not appear to affect psoriasis. Low AT had a significant single-day lag effect on psoriasis visits between lag 2 days and lag 6 days. The cumulative effects were significantly higher than single-day lag effects. Stratified analysis showed that men and patients aged 0-44 years with psoriasis were more sensitive to low AT levels. Our results provide evidence for the effects of temperature on psoriasis. Many previous studies have confirmed the existence of clear seasonality of psoriasis [32][33][34][35]. However, few experts have focused on the role of meteorological factors, ignoring the role of ambient temperature. Our study showed a positive association between low AT and the risk of psoriasis. Similar to our findings, previous studies found that exacerbations of psoriasis tend to occur in the winter months, when temperatures are cooler [13][14][15]. Several underlying physiological mechanisms may explain this phenomenon. On the one hand, the body's immune system exhibits a pro-inflammatory profile in cold conditions to prevent infection, including increased cytokine expression and C-reactive protein levels and reduced expression of glucocorticoid receptors, changes that exacerbate psoriasis [36]. On the other hand, low temperatures can increase the risk of infection [37]. Many microorganisms have been implicated in the development of psoriasis; for example, group A streptococcal hemolytic pharyngeal infections can contribute to the exacerbation of pitting psoriasis [38], while Streptococcus pyogenes infections cause exacerbation of plaque psoriasis [39]. Other microorganisms, including Staphylococcus aureus, Malassezia, and Candida, are also involved in the pathogenesis of psoriasis [40]. Moreover, it has been suggested that cold temperatures cause vasoconstriction of the skin, reducing transepidermal water loss and possibly affecting the skin barrier [41]. Additionally, studies have shown that lower ambient temperatures can cause additional mental stress in humans [42]. Stress plays an important role in the onset, development, and exacerbation of psoriasis [43]. Hester et al suggested that mental stress affects psoriasis through immune system regulation and abnormal T cell activation [44]. Petra et al proposed the gutbrain-skin axis hypothesis, in which emotionally induced psoriasis-related skin inflammation is associated with ecological dysbiosis of the intestinal microbiota [45]. However, the exact mechanisms that increase the risk of psoriasis in cold weather remain unclear.
Previous studies have also found that psoriasis is less prevalent and less active during the summer [32,33]. One of the main mechanisms is that high temperatures increase UV exposure, resulting in increased vitamin D production [46]. Vitamin D plays a crucial role in regulating dendritic and keratin-forming cells and T-cell proliferation [47]. In our study, a protective effect of high temperature on psoriasis was not observed. A possible  reason for this result is that we included the variable factor of hours of Sunlight in our model to control for its effects. Another reason may be that we used AT, a more representative indicator of human heat sensitivity [23], to assess its effect on skin, which may be more objective. In hot environments, the body may increase sebum excretion, which irritates the skin, affects the barrier function of the skin [16,48], and may have a detrimental effect on psoriasis.
The health effects of environmental stress are delayed, and usually occur within days of exposure to high levels of air pollution or extreme temperatures [49]. In general, there may be a 2-to 3-day delay in the health effects of heat and up to 2 weeks for cold [50]. Our results showed that the effect of low AT on psoriasis occurred with a 2-day delay and lasted for 5 days. The influence of air temperature on other skin diseases also showed hysteresis. For example, Wang et al found that the adverse effects of hypothermia on atopic dermatitis appeared after 11 days of exposure and then persisted for 10 days [21]. A Li et al showed that elevated temperatures have a complex protective effect on patients with eczema, with a lag of approximately 3 days [16]. In our study area, Bo Liu et al found that in Hefei the exposure to low ambient temperature decreased the risk of outpatient visits for warts on the same day and day 6, but low ambient temperature increased the risk of outpatient visits for patients with warts lagged 1-3 days, and high ambient temperature increased the risk of outpatient visits for warts only on the day of exposure [51]. The theoretical mechanism for the lag in the effect of temperature on health remains unclear. The lag time may reflect an accelerated phase of the disease process in which psoriasis causes discomfort in humans. In addition, because previous studies have found that the intensity of the health effects of temperature extremes is more important than their duration [52], we also quantified the effects of low AT. Our lagged effects analysis found that the effect of low AT showed increasing and decreasing trends. This trend is relatively similar to the harvesting effect, whereby when an external stressor predominantly affects a group of vulnerable individuals, events in these individuals are quickly brought forward by the effects of exposure [53].
Our stratified analysis revealed that younger groups appeared more vulnerable to the effects of cold temperatures. Similar to our findings, data from one study showed that younger people with psoriasis have more pronounced seasonal variations; that is, they are more likely to experience exacerbations in the colder winter months [54]. The working conditions of younger men, who are more likely to work outdoors in stressful conditions should be considered. On the other hand, people over 45 years are usually in higher positions and are less likely to be exposed to extremely cold conditions [55]. Additionally, air pollution is a risk factor for psoriasis, and younger age groups are more likely to be exposed to air pollution [56]. A previous study reported that younger people rated their discomfort higher, possibly prompting younger patients with psoriasis to seek medical attention when their symptoms worsen [57]. The high cost of medical care may prevent some seniors who cannot afford treatment from going to the hospital only when their symptoms are severe [58]. The younger cohort had a greater proportion of early onset psoriasis (age of onset <40 years), which was more frequently exacerbated by infection due to a higher incidence of titular and eruptive psoriasis [59]. Low temperatures may exacerbate psoriasis through infection. The prevalence of comorbid cardiovascular disease was significantly higher in patients with psoriasis aged >40 years [60], and older people were more likely to experience extreme weather conditions [61]. Therefore, the exacerbations of psoriasis due to exposure to cold temperatures may be overlooked, resulting in lower visit rates.
Additionally, we found that men with psoriasis were more sensitive to cold temperatures. This may be related to the hormone levels, duration of exposure to the environment, and physiological metabolism of the different sexes. For example, women have stronger innate and adaptive immune responses than men because of their hormone levels; therefore, men are usually more susceptible to infection [62]. Women are more concerned about skin health and moisturize their skin in the face of dry weather in the cooler months, thereby reducing the symptoms of psoriasis [63]. On the other hand, women are more sensitive to temperature and are aware of ambient temperature discomfort earlier than men at the onset of exposure to extreme weather and may take defensive measures earlier [64]. Smoking is positively associated with seasonal exacerbations of psoriasis [54], and this behavior in men may make them more likely to experience exacerbation of symptoms after exposure to cold temperatures.
There are several novel aspects of this study. First, we used AT, a metric more representative of actual human ambient temperature perception, to systematically determine the relationship between temperature and psoriasis, while controlling for the effects of other meteorological factors and air pollutants. These results are objective and credible. Given the previous absence of definitive evidence of a relationship between temperature and psoriasis, this gap is filled to some extent. Second, we used outpatient data rather than hospitalization data as an outcome of temperature effects, which are more representative and time sensitive. Finally, it provides a scientific basis for the seasonality of psoriasis and helps to further elucidate the mechanisms of environmental factors in psoriasis.
This study had some limitations that should be addressed in future work. First, our data were obtained from a single hospital in the same region. Although this hospital is the largest dermatology center in Anhui Province, with three widely distributed skin outpatient centers covering all the citizens of Hefei, a selection bias may still exist. Caution should be exercised in generalizing these findings to other regions. Future studies should include multiple regions with different climates. Second, this was an ecological study, and data from weather monitoring stations did not reflect individual exposure levels, thus creating an ecological fallacy. In addition, we could not control for potential individual confounders, such as stress or occupational conditions, and the results cannot be directly extrapolated to the individual level. Future studies should use environmental data from exposed individuals. Third, the exact mechanism of the effect of temperature on psoriasis is still uncertain, and more longitudinal clinical cohort and physiological studies are needed to determine this further.

Conclusions
In conclusion, our study provides quantitative evidence of an association between AT and psoriasis, suggesting that a lower AT may be a detrimental factor for psoriasis. Younger age groups and men were the more sensitive populations. Vulnerable people should be promptly reminded to take warm measures when extreme cold weather is imminent and remain vigilant for up to 5 days after exposure to cold conditions. Additionally, psoriasis clinics should increase physician scheduling during special periods and rationally allocate medical resources.