Abstract
Background
Whether absolute temperature or day-to-day temperature fluctuation more strongly precipitates acute type A aortic dissection (ATAAD) remains unresolved, particularly in East Asian populations with distinct temperate-monsoon climates. This study investigated the association between meteorological factors and ATAAD occurrence in a Korean population over a 12-year period.
Methods
Clinical data from 522 patients who underwent open surgical repair for ATAAD at a single tertiary center between January 2008 and December 2019 were retrospectively analyzed alongside meteorological data obtained from the Korea Meteorological Administration. Primary analyses focused on temperature variables—mean, lowest, and perceived (apparent) temperature on the day of onset and the previous day, as well as day-to-day temperature change—using Pearson correlation coefficients and Poisson regression models at both monthly and daily temporal scales.
Results
ATAAD incidence demonstrated a marked winter predominance, peaking in December (n=65, 12.5%) and reaching its nadir in July (n=24, 4.6%); cold-season months (November–March) collectively accounted for 273 cases (52.3%). A strong negative correlation was observed between monthly ATAAD occurrence and all absolute temperature variables (r=−0.83 to −0.85, all P<0.001). Every 10 ℃ decrease in mean onset-day temperature was associated with a 26.8% increase in monthly incidence rate [incidence rate ratio (IRR): 1.268, 95% CI: 1.151–1.395, P<0.001] and a 7.4% increase in daily incidence rate (IRR 1.074, 95% CI: 1.012–1.139, P=0.02). Day-to-day temperature change was not significantly associated with ATAAD occurrence at either temporal scale (all P≥0.63).
Conclusions
In a Korean population, lower absolute ambient temperature was significantly associated with increased ATAAD occurrence on both monthly and daily analyses, whereas day-to-day temperature fluctuation showed no statistically significant association. Heightened clinical awareness during cold seasons may contribute to earlier recognition and improved outcomes in high-risk patients.
Keywords: Aortic dissection, temperature, meteorological factors, seasonality, Korea
Highlight box.
Key findings
• In a Korean cohort of 522 acute type A aortic dissection (ATAAD) patients over 12 years, lower absolute ambient temperature was strongly associated with increased ATAAD occurrence at both monthly [Pearson r=−0.83 to −0.85; incidence rate ratio (IRR) per 10 ℃ decrease: 1.268] and daily (IRR per 10 ℃ decrease: 1.074) temporal scales. Day-to-day temperature change showed no statistically significant association at either scale (all P≥0.63). Cold-season months (November–March) accounted for 52.3% of all events, peaking in December.
What is known and what is new?
• Prior studies from Western populations have reported a winter predominance in aortic dissection, with both absolute temperature and day-to-day fluctuation proposed as triggers. However, East Asian-specific evidence is limited, and the relative contributions of these two variables remain unresolved.
• This study simultaneously evaluated absolute temperature and temperature variability using dual-scale Poisson regression in a Korean population, providing the first systematic country-specific evidence from a temperate-monsoon climate.
What is the implication, and what should change now?
• Health systems in cold-temperate East Asian countries should prioritize surgical preparedness during prolonged cold periods. High-risk patients with known aortic pathology may benefit from heightened clinical awareness and blood pressure monitoring during cold seasons.
Introduction
Whether sustained cold exposure or day-to-day temperature fluctuation more strongly precipitates acute type A aortic dissection (ATAAD) remains a clinically important and unresolved question. ATAAD carries an in-hospital mortality of approximately 15–30% even after prompt surgical intervention (1,2), and identifying the environmental triggers that precipitate acute events may inform clinical preparedness and preventive strategies.
A growing body of evidence suggests that meteorological factors—particularly ambient temperature—may influence the timing of ATAAD onset (3). Several studies from Western Europe, North America, and East Asia have reported a consistent winter peak in aortic dissection incidence, with low ambient temperature proposed as a key environmental trigger (4-7). Cold exposure activates the sympathetic nervous system, raising blood pressure and aortic wall stress, thereby potentially precipitating intimal rupture in susceptible individuals (8). However, the relative contributions of absolute temperature versus acute temperature change remain debated (4,9).
Prior landmark studies, including the International Registry of Acute Aortic Dissection (IRAD) analysis by Mehta et al., reported that day-to-day temperature variability—rather than absolute temperature—was the primary meteorological driver of ATAAD occurrence (10,11). In contrast, subsequent single-center European and Asian studies have suggested that absolute ambient temperature is the more consistent predictor (4,12,13). This discrepancy highlights the potential role of regional climatic differences and warrants country-specific investigation.
Korea experiences a temperate monsoon climate with pronounced seasonal variation: cold, dry winters (mean January temperature below 0 ℃ in Seoul) and hot, humid summers. Whether the meteorological patterning of ATAAD in this specific climatic context mirrors Western cohorts, and whether temperature change adds independent predictive value beyond absolute temperature, has not been systematically examined.
This study aimed to investigate the association between meteorological variables and the monthly and daily occurrence of ATAAD in a large single-center Korean cohort over a 12-year period. The originality of this study lies in: (I) providing the first systematic dual-scale (monthly + daily) Poisson regression evidence from a Korean temperate-monsoon population; (II) simultaneously testing ΔT versus absolute temperature at two temporal scales using both Pearson and Poisson methods; and (III) employing multiple sensitivity analyses to confirm model robustness. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1194/rc).
Methods
Study population
This study retrospectively analyzed 522 patients who underwent open surgical repair for ATAAD at Gangnam Severance Hospital between January 2008 and December 2019. All diagnoses were confirmed by computed tomography. Of the initial 550 patients, 28 were excluded: 27 for incomplete medical records and 1 for a diagnosis made abroad. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Gangnam Severance Hospital (IRB No. 3-2023-0234), and the requirement for informed consent was waived given the retrospective design. The patient selection process is summarized in Figure 1.
Figure 1.
Patient selection flow diagram. Meteorological data were obtained from the KMA station nearest to each patient’s location at symptom onset. ATAAD, acute type A aortic dissection; KMA, Korea Meteorological Administration.
Meteorological data
Meteorological data corresponding to each patient’s date of symptom onset were obtained from the Korea Meteorological Administration (KMA). Meteorological variables were matched to the KMA automated weather station nearest to the patient’s location at the time of symptom onset.
Three categories of variables were collected for both the onset day (D) and the previous day (d), with day-to-day differences (Δ = onset day minus previous day) also calculated. Onset-day mean temperature (D mean T) was pre-specified as the primary exposure variable, as it is the most widely used metric in comparable prior studies.
Temperature variables
The following temperature variables were collected: onset-day and previous-day mean temperature (D mean T / d mean T), lowest temperature (D low T / d low T), and perceived (apparent) temperature (D per T / d per T), together with their day-to-day change counterparts (Δ mean T / Δ low T / Δ per T). Perceived temperature was calculated using the formula: Per T = 13.12 + 0.6215T − 11.37V0.16 + 0.3965V0.16T, where T is ambient temperature (℃) and V is wind speed (km/h).
Humidity variables
Mean and lowest relative humidity on the onset and previous day (D mean H / d mean H / D low H / d low H), and their day-to-day changes (Δ mean H / Δ low H) were collected.
Atmospheric pressure variables
Both local station pressure and sea-level pressure were collected. As the two were nearly perfectly collinear (r>0.99, P<0.001), sea-level pressure was selected as the representative variable to reduce redundancy. Mean, highest, and lowest sea-level pressure on the onset and previous day (D mean ocean P / d mean ocean P, etc.), and their day-to-day changes, were also collected.
Statistical analysis
Differences in ATAAD occurrence across months and seasons were assessed using one-way analysis of variance (ANOVA). Pearson correlation coefficients were calculated to quantify the association between the monthly number of ATAAD cases (total cases per month, averaged across 12 years) and the monthly mean values of each meteorological variable. Correlation coefficients were considered very strong if |r|≥0.9, strong if 0.7≤|r|<0.9, moderate if 0.5≤|r|<0.7, and weak if |r|<0.5.
Poisson regression models were constructed at two temporal scales. (I) Monthly model: the dependent variable was the total number of ATAAD cases per calendar month aggregated across 12 years; the predictor was the monthly mean of each meteorological variable; this is the pre-specified primary analysis. (II) Daily model: the dependent variable was the number of ATAAD cases per calendar day across all 12 years; the predictor was the meteorological value on that date; this is the pre-specified confirmatory secondary analysis.
Three sensitivity analyses were performed to assess robustness (results in Table S1): (I) negative binomial regression for all primary temperature variables at both temporal scales; (II) season-adjusted monthly Poisson model with categorical season as a covariate alongside D mean T; (III) multivariable daily Poisson model including D mean T, a restricted cubic spline for calendar time, and day of the week. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A two-sided P value <0.05 was considered statistically significant.
Results
Patient characteristics
A total of 522 patients were included in the final analysis. The mean age was 61.3±14.7 years, and 255 patients (48.9%) were male. Hypertension was the most common comorbidity (n=317, 60.7%), followed by smoking (n=208, 39.8%), diabetes (n=46, 8.8%), dyslipidemia (n=44, 8.4%), prior open-heart surgery (n=11, 2.1%), and Marfan syndrome or connective tissue disorder (n=31, 5.9%). Detailed baseline characteristics are presented in Table 1.
Table 1. Characteristics of patients (N=522).
| Variables | Value |
|---|---|
| Age (years) | 61.3±14.7 |
| Male sex | 255 (48.9) |
| Smoking | 208 (39.8) |
| Hypertension | 317 (60.7) |
| Dyslipidemia | 44 (8.4) |
| CAOD | 24 (4.6) |
| DM | 46 (8.8) |
| COPD | 8 (1.5) |
| Old CVA | 35 (6.7) |
| CKD | 19 (3.6) |
| Marfan syndrome | 31 (5.9) |
| Prior OHS | 11 (2.1) |
| Preoperative CPR | 38 (7.3) |
| Hemopericardium | 221 (42.3) |
| Malperfusion | 175 (33.5) |
Values are presented as mean ± standard deviation or number (%). CAOD, coronary artery occlusive disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CPR, cardiopulmonary resuscitation; CVA, cerebrovascular accident; DM, diabetes mellitus; OHS, open-heart surgery.
Seasonal and monthly distribution
ATAAD occurrence demonstrated a modest but consistent winter predominance over the 12-year study period (Figure 2). Cold-season months (November through March) collectively accounted for 273 events—52.3% of the total cohort. The peak occurred in December (n=65, 12.5%), followed by February (n=59, 11.3%) and March (n=55, 10.5%). Conversely, summer months recorded the lowest incidence, with July yielding only 24 cases (4.6%) and August 34 cases (6.5%). South Korea officially defines four meteorological seasons as follows: spring (March–May), summer (June–August), autumn (September–November), and winter (December–February). The cold-season grouping used for the 52.3% statistic (November–March) was separately defined based on the observed temperature profile. Monthly case counts inversely tracked mean and lowest ambient temperatures throughout the year (Figure 2). Detailed monthly meteorological data are summarized in Table 2.
Figure 2.
Monthly distribution of ATAAD cases and onset-day temperature across the 12-year study period (2008–2019). Bars represent the total number of ATAAD cases per calendar month, color-coded by standard Korean meteorological season. Overlaid line graphs depict monthly mean onset-day temperature and monthly mean lowest onset-day temperature in degrees Celsius (right y-axis). The solid gray horizontal line indicates the mean monthly ATAAD case count over the 12-year period (43.5 cases), and the dotted gray horizontal line indicates 0 ℃. ATAAD, acute type A aortic dissection.
Table 2. Meteorological factors by month.
| Factors | Total | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | P value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N (%) | 522 | 42 (8.0) | 59 (11.3) | 55 (10.5) | 46 (8.8) | 44 (8.4) | 29 (5.6) | 24 (4.6) | 34 (6.5) | 28 (5.3) | 44 (8.4) | 52 (10.0) | 65 (12.5) | |
| D mean T (℃) | 10.1 (10.4) | −2.7 (3.8) | −0.2 (4.9) | 5.9 (3.4) | 11.0 (3.8) | 17.9 (2.9) | 22.7 (1.6) | 27.0 (2.8) | 26.2 (2.2) | 21.6 (3.2) | 13.8 (4.3) | 7.0 (4.5) | −0.1 (4.8) | <0.0001 |
| d mean T (℃) | 10.2 (10.4) | −2.3 (3.5) | −0.5 (4.5) | 6.1 (3.5) | 11.1 (3.9) | 18.0 (3.0) | 23.6 (1.5) | 26.0 (3.8) | 26.9 (2.2) | 21.9 (2.8) | 13.4 (3.8) | 7.4 (4.7) | 0.3 (4.9) | <0.0001 |
| D low T (℃) | 5.8 (10.5) | −6.5 (3.9) | −4.7 (4.7) | 1.3 (3.2) | 6.3 (3.7) | 12.7 (3.2) | 18.9 (2.3) | 23.5 (2.7) | 23.1 (1.9) | 17.4 (3.8) | 8.8 (4.9) | 2.8 (4.7) | −4.1 (5.1) | <0.0001 |
| d low T (℃) | 5.8 (10.7) | −6.5 (3.7) | −5.0 (4.7) | 1.1 (3.7) | 6.6 (3.4) | 13.1 (3.0) | 18.9 (2.1) | 23.1 (2.5) | 23.7 (1.9) | 17.4 (3.5) | 8.8 (4.3) | 3.0 (4.5) | −4.3 (6.4) | <0.0001 |
| D per T (℃) | 9.2 (11.3) | −4.8 (4.7) | −2.2 (6.0) | 4.7 (4.1) | 10.3 (4.3) | 17.9 (2.9) | 22.8 (1.7) | 27.0 (2.8) | 26.3 (2.1) | 21.7 (3.1) | 13.5 (4.5) | 6.0 (4.9) | −1.5 (5.4) | <0.0001 |
| d per T (℃) | 9.4 (11.2) | −4.1 (4.1) | −2.5 (5.5) | 4.9 (4.1) | 10.2 (4.5) | 17.7 (2.9) | 23.4 (1.6) | 26.4 (2.8) | 26.9 (2.2) | 21.9 (2.8) | 13.3 (3.9) | 6.7 (5.2) | −1.2 (5.6) | <0.0001 |
| Δ mean T (℃) | −0.2 (2.7) | −0.4 (3.4) | 0.3 (3.2) | −0.3 (2.7) | −0.0 (2.6) | −0.0 (2.8) | −0.9 (1.5) | 1.0 (2.7) | −0.7 (1.5) | −0.3 (1.5) | 0.4 (1.7) | −0.4 (3.3) | −0.4 (3.2) | 0.36 |
| Δ low T (℃) | −0.0 (3.1) | 0.1 (3.6) | 0.3 (3.3) | 0.2 (3.1) | −0.3 (2.4) | −0.4 (2.5) | 0.1 (1.2) | 0.4 (1.4) | −0.6 (1.5) | −0.1 (2.8) | −0.1 (2.4) | −0.2 (3.0) | 0.1 (5.4) | 0.98 |
| Δ per T (℃) | −0.2 (3.1) | −0.7 (4.4) | 0.3 (3.9) | −0.2 (3.2) | 0.1 (3.0) | 0.2 (2.6) | −0.6 (1.3) | 0.6 (1.5) | −0.6 (1.5) | −0.2 (1.5) | 0.3 (1.8) | −0.7 (3.9) | −0.3 (3.7) | 0.66 |
Values are presented as mean (standard deviation) unless otherwise indicated. Δ, day-to-day change (onset minus previous day); d, previous day; D, onset day; low T, lowest temperature; mean T, mean temperature; per T, perceived temperature.
Monthly analysis
Pearson correlation analysis using the monthly aggregated data revealed strong and statistically significant negative associations between ATAAD occurrence and all absolute temperature variables (Table 3, Figure 3). Among onset-day variables, D mean T showed r=−0.841 (P=0.0006) and D low T showed r=−0.850 (P=0.0005). Day-to-day temperature change variables (Δ mean T, Δ low T, Δ per T) showed no significant correlation (r=−0.065 to 0.107, all P>0.35). In the monthly Poisson regression model, every 10 ℃ decrease in D mean T was associated with a 26.8% increase in monthly incidence rate [incidence rate ratio (IRR) 1.268, 95% CI: 1.151–1.395, P<0.001]. Full results for all temperature variables are presented in Table 4.
Table 3. Correlation between ATAAD occurrence and temperature.
| Variables | ATAAD occurrence | D mean T | d mean T | D low T | d low T | D per T | d per T | Δ mean T | Δ low T | Δ per T |
|---|---|---|---|---|---|---|---|---|---|---|
| ATAAD occurrence | 1 | −0.841 (0.0006) | −0.840 (0.0006) | −0.850 (0.0005) | −0.846 (0.0005) | −0.835 (0.0007) | −0.838 (0.0007) | −0.065 (0.84) | 0.107 (0.74) | −0.022 (0.95) |
| D mean T | −0.841 (0.0006) | 1 | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.086 (0.79) | −0.295 (0.35) | 0.160 (0.62) |
| d mean T | −0.840 (0.0006) | 0.999 (<0.0001) | 1 | 0.998 (<0.0001) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.039 (0.90) | −0.317 (0.32) | 0.117 (0.72) |
| D low T | −0.850 (0.0005) | 0.999 (<0.0001) | 0.998 (<0.0001) | 1 | 0.999 (<0.0001) | 0.997 (<0.0001) | 0.998 (<0.0001) | 0.069 (0.83) | −0.285 (0.37) | 0.132 (0.68) |
| d low T | −0.846 (0.0005) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.998 (<0.0001) | 1 | 0.998 (<0.0001) | 0.998 (<0.0001) | 0.057 (0.86) | −0.309 (0.33) | 0.122 (0.71) |
| D per T | −0.835 (0.0007) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.997 (<0.0001) | 0.998 (<0.0001) | 1 | 0.999 (<0.0001) | 0.082 (0.80) | −0.303 (0.34) | 0.160 (0.62) |
| d per T | −0.838 (0.0007) | 0.999 (<0.0001) | 0.999 (<0.0001) | 0.998 (<0.0001) | 0.998 (<0.0001) | 0.999 (<0.0001) | 1 | 0.046 (0.89) | −0.317 (0.32) | 0.122 (0.71) |
| Δ mean T | −0.065 (0.84) | 0.086 (0.79) | 0.039 (0.90) | 0.069 (0.83) | 0.057 (0.86) | 0.082 (0.80) | 0.046 (0.89) | 1 | 0.447 (0.15) | 0.921 (<0.0001) |
| Δ low T | 0.107 (0.74) | −0.295 (0.35) | −0.317 (0.32) | −0.285 (0.37) | −0.309 (0.33) | −0.303 (0.34) | −0.317(0.32) | 0.447 (0.15) | 1 | 0.324 (0.31) |
| Δ per T | −0.022 (0.95) | 0.160 (0.62) | 0.117 (0.72) | 0.132 (0.68) | 0.122 (0.71) | 0.160 (0.62) | 0.122 (0.71) | 0.921 (<0.0001) | 0.324 (0.31) | 1 |
Data are presented as Pearson r (P value). Δ, day-to-day change (onset minus previous day); ATAAD, acute type A aortic dissection; d, previous day; D, onset day; low T, lowest temperature; mean T, mean temperature; per T, perceived temperature.
Figure 3.
Scatter plot of monthly ATAAD case counts vs. mean onset-day temperature. Each data point represents the mean monthly ATAAD case count averaged across the 12 annual observations for each calendar month, plotted against the 12-year mean onset-day temperature for that month, color-coded by season; error bars represent ±1 standard deviation. The dashed red line indicates the linear regression fit. A strong inverse correlation was observed (Pearson r=−0.84, P<0.001). ATAAD, acute type A aortic dissection.
Table 4. Poisson regression model of monthly occurrence.
| Variables | Univariable model | ||
|---|---|---|---|
| IRR (95% CI) (per 1 ℃ decrease) | IRR (95% CI) (per 10 ℃ decrease) | P value | |
| D mean T | 1.024 (1.014, 1.034) | 1.268 (1.151, 1.395) | <0.0001 |
| d mean T | 1.024 (1.014, 1.034) | 1.268 (1.152, 1.397) | <0.0001 |
| D low T | 1.024 (1.015, 1.034) | 1.269 (1.157, 1.392) | <0.0001 |
| d low T | 1.024 (1.014, 1.033) | 1.265 (1.153, 1.388) | <0.0001 |
| D per T | 1.022 (1.013, 1.031) | 1.244 (1.136, 1.362) | <0.0001 |
| d per T | 1.022 (1.013, 1.032) | 1.247 (1.139, 1.365) | <0.0001 |
| Δ mean T | 1.040 (0.717, 1.507) | 1.475 (0.036, 60.422) | 0.84 |
| Δ low T | 0.893 (0.468, 1.704) | 0.322 (0.001, 206.073) | 0.73 |
| Δ per T | 1.015 (0.668, 1.542) | 1.158 (0.018, 76.119) | 0.95 |
Δ, day-to-day change (onset minus previous day); CI, confidence interval; d, previous day; D, onset day; IRR, incidence rate ratio; low T, lowest temperature; mean T, mean temperature; per T, perceived temperature.
Daily analysis
The daily Poisson regression model, constructed across the full 4,383-day study period, confirmed the temperature–ATAAD relationship at the individual-case level, with attenuated effect sizes as expected given greater day-to-day variability. A 10 ℃ decrease in D mean T was associated with a 7.4% increase in daily incidence rate (IRR 1.074, 95% CI: 1.012–1.139, P=0.02). Day-to-day temperature change variables were not significantly associated with daily ATAAD occurrence (all P>0.05), consistent with the monthly analysis (Table 5). The distribution of day-to-day temperature change on ATAAD event days was virtually identical to that on non-event days across the full study period (Figure S1).
Table 5. Poisson regression model of daily occurrence.
| Variables | Univariable model | ||
|---|---|---|---|
| IRR (95% CI) (per 1 ℃ decrease) | IRR (95% CI) (per 10 ℃ decrease) | P value | |
| D mean T | 1.007 (1.001, 1.013) | 1.074 (1.012, 1.139) | 0.02 |
| d mean T | 1.007 (1.001, 1.014) | 1.078 (1.015, 1.144) | 0.01 |
| D low T | 1.007 (1.001, 1.013) | 1.076 (1.015, 1.141) | 0.01 |
| d low T | 1.007 (1.001, 1.013) | 1.069 (1.009, 1.132) | 0.02 |
| D per T | 1.006 (1.001, 1.012) | 1.065 (1.008, 1.125) | 0.02 |
| d per T | 1.007 (1.001, 1.012) | 1.069 (1.011, 1.130) | 0.02 |
| Δ mean T | 0.994 (0.965, 1.023) | 0.937 (0.702, 1.251) | 0.66 |
| Δ low T | 1.006 (0.982, 1.030) | 1.061 (0.834, 1.350) | 0.63 |
| Δ per T | 0.995 (0.971, 1.020) | 0.952 (0.743, 1.219) | 0.70 |
Δ, day-to-day change (onset minus previous day); CI, confidence interval; d, previous day; D, onset day; IRR, incidence rate ratio; low T, lowest temperature; mean T, mean temperature; per T, perceived temperature.
Humidity, atmospheric pressure, and collinearity
Mean relative humidity on the onset day showed a moderate negative monthly correlation (D mean H: r=−0.674, P=0.02), and sea-level pressure showed a strong positive correlation (D mean ocean P: r=0.776, P=0.003). However, all humidity and pressure variables were highly collinear with temperature (r>0.72, P<0.001), indicating that these apparent associations most likely reflect seasonal confounding by temperature rather than independent meteorological effects. Full results are presented in Table 6.
Table 6. Pearson correlation and monthly Poisson regression: humidity, local pressure, and sea-level pressure vs. ATAAD occurrence.
| Variable | Monthly correlation | Monthly Poisson regression (univariable model) | ||||
|---|---|---|---|---|---|---|
| Pearson r | P value | IRR (95% CI) (per 1 unit) | IRR (95% CI) (per 10 unit) | P value | ||
| Humidity (%) | ||||||
| D mean H | −0.674 | 0.02* | 1.029 (1.009–1.049) | 1.324 (1.089–1.610) | 0.005* | |
| d mean H | −0.674 | 0.02* | 1.029 (1.009–1.049) | 1.324 (1.089–1.610) | 0.005* | |
| D low H | −0.690 | 0.01* | 1.026 (1.009–1.043) | 1.290 (1.089–1.528) | 0.003* | |
| d low H | −0.690 | 0.01* | 1.026 (1.009–1.043) | 1.290 (1.089–1.528) | 0.003* | |
| Δ mean H | −0.208 | 0.52 | 1.034 (0.936–1.143) | 1.402 (0.518–3.797) | 0.51 | |
| Δ low H | −0.139 | 0.67 | 1.017 (0.943–1.098) | 1.188 (0.555–2.543) | 0.66 | |
| Local atmospheric pressure (hPa) | ||||||
| D mean local P | +0.737 | 0.01* | 1.031 (1.013–1.049) | 1.356 (1.139–1.616) | 0.001* | |
| d mean local P | +0.737 | 0.01* | 1.031 (1.013–1.049) | 1.356 (1.139–1.616) | 0.001* | |
| Δ mean local P | −0.067 | 0.84 | 1.039 (0.727–1.485) | 1.470 (0.041–52.275) | 0.83 | |
| Sea-level atmospheric pressure (hPa) | ||||||
| D mean ocean P | +0.776 | 0.003* | 1.032 (1.016–1.049) | 1.376 (1.170–1.618) | 0.0001* | |
| d mean ocean P | +0.776 | 0.003* | 1.032 (1.016–1.049) | 1.376 (1.170–1.618) | 0.0001* | |
| D high ocean P | +0.797 | 0.002* | 1.032 (1.016–1.047) | 1.364 (1.178–1.579) | <0.0001* | |
| d high ocean P | +0.797 | 0.002* | 1.032 (1.016–1.047) | 1.364 (1.178–1.579) | <0.0001* | |
| D low ocean P | +0.750 | 0.005* | 1.033 (1.015–1.052) | 1.383 (1.156–1.654) | 0.0004* | |
| d low ocean P | +0.770 | 0.003* | 1.033 (1.016–1.051) | 1.384 (1.168–1.640) | 0.0002* | |
| Δ mean ocean P | −0.067 | 0.84 | 1.040 (0.726–1.489) | 1.475 (0.041–53.404) | 0.83 | |
*, statistically significant. Δ, day-to-day change (onset minus previous day); ATAAD, acute type A aortic dissection; CI, confidence interval; d, previous day; D, onset day; IRR, incidence rate ratio; low H, lowest humidity; local P, station-level atmospheric pressure; mean H, mean humidity; ocean P, sea-level atmospheric pressure.
Sensitivity analyses
All three sensitivity analyses confirmed the robustness of the primary findings (Table S1). The negative binomial model could not be fitted to the 12-point monthly aggregated data, in which the Poisson dispersion index of 1.26 confirmed that Poisson regression was appropriate. In the year-month sensitivity analysis (144 year-month strata), negative binomial estimates were virtually identical to Poisson estimates (D mean T per 10 ℃ decrease: NB IRR 1.209, 95% CI: 0.994–1.468 vs. Poisson IRR 1.205, 95% CI: 1.106–1.314). Season-adjusted monthly Poisson regression confirmed the temperature–ATAAD association after categorical seasonal adjustment (IRR 1.198, 95% CI: 1.062–1.351, P=0.003). Multivariable daily Poisson regression including calendar time spline and day-of-week also confirmed the association (IRR 1.068, 95% CI: 1.006–1.134, P=0.03).
Discussion
The present study demonstrated that cold ambient temperature is significantly associated with increased ATAAD occurrence in a Korean tertiary referral center over a 12-year period, consistent with a growing body of international evidence from both Western and East Asian populations (4-7,12,13).
Our findings are directly concordant with two large-scale East Asian studies. Kato et al. (Japan, 96,812 hospitalizations, 2012–2020) used a nationwide time-stratified case-crossover design with distributed lag non-linear models and found low ambient temperature associated with acute aortic dissection (AAD) hospitalization, peaking at −10 ℃ [odds ratio (OR) 2.28 vs. 20 ℃ reference] (14). Yu et al. (single-centre 8-year time-series, Wuhan, China, 2014–2018, 2,120 patients) found that extremely cold temperature increased type A AAD risk [risk ratio (RR) 3.951] while temperature variability within 1 week did not affect onset risk (12)—directly corroborating our ΔT null finding. The slightly smaller effect size in our study (IRR 1.268 per 10 ℃ monthly; IRR 1.074 per 10 ℃ daily) compared to these studies may reflect differences in analytical method [Poisson vs. distributed lag non-linear model (DLNM)/case-crossover] and study scale (single-center vs. nationwide). Consistent findings have also been reported from other East Asian single-center studies in China (15-17).
The larger monthly vs. daily effect size (IRR 1.268 vs. 1.074) is an expected methodological feature: monthly means smooth out within-season variability, reducing noise and amplifying the temperature signal in ecological aggregation. This pattern has been observed consistently across the temperature–ATAAD literature and does not represent a contradiction.
Korea’s temperate monsoon climate (Seoul mean January: −2 ℃ to −3 ℃; mean August: ~26 ℃; annual range: ~29 ℃) provides a substantially wider thermal amplitude than Western European study sites: Toulouse, France [annual range ~17 ℃; Benouaich et al. (9)] and Germany [~18 ℃; Kurz et al. (18), Kletzer et al. (19), Boening et al. (20)]. This larger absolute temperature range likely amplifies the seasonal ATAAD signal relative to milder European climates, while the broadly comparable East Asian temperate-monsoon climate of Japan accounts for the strong concordance between our findings and the Japanese nationwide data (14).
The pathophysiological mechanism most widely proposed involves sympathetic nervous system activation in response to cold exposure, triggering systemic vasoconstriction, elevated heart rate, and increased systolic blood pressure (8,21). The resultant rise in aortic wall stress may precipitate intimal rupture in susceptible individuals with pre-existing hypertension, aortic dilation, or medial degeneration (1,2,22). This mechanism aligns with our finding that lower absolute temperatures—whether mean, lowest, or perceived—were uniformly associated with higher ATAAD incidence.
The absence of any significant ΔT association at either temporal scale stands in contrast to the hypothesis proposed by Mehta et al. (10) and the findings of Chen et al. (13), who found ΔT independently associated with AAD in non-heating seasons after mutual adjustment. The discrepancy with Chen et al. may reflect: (I) mutual adjustment for absolute temperature vs. our separate univariable ΔT models; (II) case-crossover vs. Poisson regression design; and (III) non-heating season stratification in Chen et al. Furthermore, our ΔT operationalization (onset day minus previous day) is simplified: it does not account for physiological context relative to seasonal norms or cumulative cold stress, as a 2 ℃ decrease from 30 ℃ and one from 15 ℃ are treated identically. Kletzer et al. (19), using DLNM with multi-day lag structures, found peak effects at 3–4 days prior to onset, supporting the view that more advanced thermal stress operationalizations may yield different ΔT conclusions. Our null ΔT finding should therefore be interpreted as ‘no statistically significant association was observed with this simplified 24-hour ΔT metric’ rather than definitive evidence of no ΔT effect. To explore this further, we performed extended analyses using multi-day rolling mean temperature (3- and 7-day windows), multi-day ΔT (over 3 and 7 days), and deviation from the 12-year monthly climatological mean; all extended operationalizations remained statistically non-significant (Table S2), reinforcing that the null ΔT signal is not an artifact of the 24-hour metric alone.
Regarding humidity and pressure: both co-vary with ambient temperature across seasons, and their high collinearity with temperature (r>0.72) suggests that their apparent monthly correlations reflect seasonal confounding rather than independent meteorological effects. The marginal advantage of sea-level over local station pressure (r=0.776 vs. 0.737) is consistent with the predominantly Seoul metropolitan catchment area (mean elevation ~38 m), which minimizes the difference between local and sea-level pressure, and with sea-level pressure’s greater sensitivity to the broad synoptic-scale high-pressure patterns associated with cold, dry winter weather. The winter peak in ATAAD is also consistent with the known cold-season predominance of other cardiovascular events, including myocardial infarction, stroke, and sudden cardiac death (8).
Limitations
Several limitations should be acknowledged. First, this is a retrospective, single-center surgical series; patients who died before hospital arrival, those managed conservatively, and those referred to other centers are not represented. In Korea, an estimated 5–10% of ATAAD presentations at tertiary centers may be managed non-operatively, and cold-induced hemodynamic stress could preferentially cause more severe dissections during winter—potentially amplifying or attenuating the observed seasonal signal through differential selection.
Second, meteorological variables were matched to the KMA automated weather station nearest to the patient’s location at the time of symptom onset; however, the recorded outdoor ambient temperature does not fully capture individual-level cold exposure at the precise moment of symptom onset. Factors such as indoor heating practices, building insulation, physical activity, and microclimate variation are not accounted for. Future studies using wearable environmental sensors or high-resolution gridded meteorological data would provide more precise individual-level exposure estimation.
Third, the ΔT operationalization (onset day minus previous day) does not account for physiological context relative to seasonal norms, cumulative cold stress, or multi-day lag effects—a limitation that may explain the null ΔT finding and warrants investigation using more advanced analytical frameworks such as DLNM. Fourth, patients with known pre-existing aortic dissection on prior imaging were not separately excluded as a formal criterion in this retrospective dataset. Fifth, other seasonally varying confounders (physical activity, dietary habits, respiratory infection rates) were not adjusted for, and residual seasonal confounding cannot be entirely excluded.
Conclusions
Cold ambient temperature was significantly associated with increased ATAAD occurrence in a Korean population, with more than half of all events occurring during the cold-season months. Both monthly and daily Poisson regression analyses confirmed that lower absolute temperatures—including mean, lowest, and perceived temperature on the onset day and the preceding day—were independently associated with higher incidence rates, while day-to-day temperature fluctuation showed no statistically significant association in the present study; this finding does not exclude the possibility of a small or context-dependent ΔT effect, and methodologically advanced studies incorporating lag structures and seasonal context are needed. Heightened clinical awareness and blood pressure monitoring during cold seasons may contribute to earlier recognition and improved outcomes in high-risk patients.
Supplementary
The article’s supplementary files as
Acknowledgments
This work was partially based on Myeong Su Kim’s thesis for the Master of Medical Science degree at Yeungnam University College of Medicine (2024).
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Gangnam Severance Hospital (IRB No. 3-2023-0234). The requirement for individual consent was waived due to the retrospective nature of the study.
Footnotes
Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1194/rc
Funding: None.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1194/coif). S.W.S. serves as an unpaid editorial board member of Journal of Thoracic Disease from February 2025 to January 2027. The other authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1194/dss
References
- 1.Hagan PG, Nienaber CA, Isselbacher EM, et al. The International Registry of Acute Aortic Dissection (IRAD): new insights into an old disease. JAMA 2000;283:897-903. 10.1001/jama.283.7.897 [DOI] [PubMed] [Google Scholar]
- 2.Golledge J, Eagle KA. Acute aortic dissection. Lancet 2008;372:55-66. 10.1016/S0140-6736(08)60994-0 [DOI] [PubMed] [Google Scholar]
- 3.Nienaber CA, Clough RE. Management of acute aortic dissection. Lancet 2015;385:800-11. 10.1016/S0140-6736(14)61005-9 [DOI] [PubMed] [Google Scholar]
- 4.DeAnda A, Jr, Grossi EA, Balsam LB, et al. The Chronobiology of Stanford Type A Aortic Dissections: A Comparison of Northern versus Southern Hemispheres. Aorta (Stamford) 2015;3:182-6. 10.12945/j.aorta.2015.15.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nienaber CA, Eagle KA. Aortic dissection: new frontiers in diagnosis and management: Part I: from etiology to diagnostic strategies. Circulation 2003;108:628-35. 10.1161/01.CIR.0000087009.16755.E4 [DOI] [PubMed] [Google Scholar]
- 6.Ma WG, Li B, Zhang W, et al. Chronologic and Climatic Factors of Acute Aortic Dissection: Study of 1642 Patients in Two Continents. Ann Thorac Surg 2020;110:575-81. 10.1016/j.athoracsur.2019.11.013 [DOI] [PubMed] [Google Scholar]
- 7.Wu H, Wang Z, Li M, et al. A Systematic Review and Meta-Analysis of Seasonal and Monthly Variability in the Incidence of Acute Aortic Dissection. Ann Vasc Surg 2022;85:383-94. 10.1016/j.avsg.2022.04.008 [DOI] [PubMed] [Google Scholar]
- 8.Ikäheimo TM. Cardiovascular diseases, cold exposure and exercise. Temperature (Austin) 2018;5:123-46. 10.1080/23328940.2017.1414014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Benouaich V, Soler P, Gourraud PA, et al. Impact of meteorological conditions on the occurrence of acute type A aortic dissections. Interact Cardiovasc Thorac Surg 2010;10:403-6. 10.1510/icvts.2009.219873 [DOI] [PubMed] [Google Scholar]
- 10.Mehta RH, Manfredini R, Hassan F, et al. Chronobiological patterns of acute aortic dissection. Circulation 2002;106:1110-5. 10.1161/01.cir.0000027568.39540.4b [DOI] [PubMed] [Google Scholar]
- 11.Mehta RH, Manfredini R, Bossone E, et al. The winter peak in the occurrence of acute aortic dissection is independent of climate. Chronobiol Int 2005;22:723-9. 10.1080/07420520500179605 [DOI] [PubMed] [Google Scholar]
- 12.Yu X, Xia L, Xiao J, et al. Association of Daily Mean Temperature and Temperature Variability With Onset Risks of Acute Aortic Dissection. J Am Heart Assoc 2021;10:e020190. 10.1161/JAHA.120.020190 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chen J, Gao Y, Jiang Y, et al. Low ambient temperature and temperature drop between neighbouring days and acute aortic dissection: a case-crossover study. Eur Heart J 2022;43:228-35. 10.1093/eurheartj/ehab803 [DOI] [PubMed] [Google Scholar]
- 14.Kato K, Nishino T, Otsuka T, et al. Nationwide analysis of the relationship between low ambient temperature and acute aortic dissection-related hospitalizations. Eur J Prev Cardiol 2025;32:317-24. 10.1093/eurjpc/zwae278 [DOI] [PubMed] [Google Scholar]
- 15.Zhang H, Yin L, Zhang Y, et al. Short-term effects of air pollution and weather changes on the occurrence of acute aortic dissection in a cold region. Front Public Health 2023;11:1172532. 10.3389/fpubh.2023.1172532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lin Q, Wu Q, Chen X, et al. Seasonal and daily variations in the occurence and outcomes of acute Stanford type A dissections: a retrospective single-center study. J Cardiothorac Surg 2023;18:108. 10.1186/s13019-023-02222-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Luo ZR, Yu LL, Huang ST, et al. Impact of meteorological factors on the occurrence of acute aortic dissection in Fujian Province, China: a single-center seven-year retrospective study. J Cardiothorac Surg 2020;15:178. 10.1186/s13019-020-01227-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kurz SD, Mahlke H, Graw K, et al. Patterns in acute aortic dissection and a connection to meteorological conditions in Germany. PLoS One 2024;19:e0296794. 10.1371/journal.pone.0296794 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kletzer J, van Kampen A, Peterss S, et al. Weathering the Aorta: The Impact of Environmental Factors on Acute Aortic Dissection. Eur J Cardiothorac Surg 2025;67:ezaf403. 10.1093/ejcts/ezaf403 [DOI] [PubMed] [Google Scholar]
- 20.Boening A, Brühl M, Büsch C, et al. Do the Frequency of Aortic Dissection Surgery and 30-day Survival Depend on the Day of the Week, Climate Parameters, or Moon Phase? Thorac Cardiovasc Surg 2026;74:274-82. 10.1055/a-2547-5604 [DOI] [PubMed] [Google Scholar]
- 21.Li Y, Wu J, Xu Y, et al. Cold exposure and the cardiovascular system: from physiological adaptation to pathological risk. Front Physiol 2025;16:1740919. 10.3389/fphys.2025.1740919 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dong N, Piao H, Li B, et al. Poor management of hypertension is an important precipitating factor for the development of acute aortic dissection. J Clin Hypertens (Greenwich) 2019;21:804-12. 10.1111/jch.13556 [DOI] [PMC free article] [PubMed] [Google Scholar]