Introduction
Globally, acute appendicitis affects 1.17 individuals per 1,000 population annually, with a lifetime risk of 8.6% for men and 6.7% for women [1]. In Korea, the incidence rate is 2.27 per 1,000 population [2]. Although numerous studies have explored the pathogenetic roles of various infectious agents in appendicitis, including viral, bacterial, fungal, and parasitic organisms, there is still no consensus on specific causes [3]. A recent study examining the pathological evidence of appendicitis suggests that pressure in the appendix lumen increases due to the proliferation of intestinal bacteria following lumen obstruction and the accumulation of secreted mucus, leading to pain around the navel [4]. However, this does not account for the initial surge in intestinal bacteria that triggers acute appendicitis. It is believed that this increase may be due to immunological changes or environmental factors, rather than the onset of any specific disease state. The incidence of appendicitis in Western countries rose from the 19th to the early 20th century and then declined after the mid to late 20th century [5]. To explain these historical fluctuations, particulate matter ≤10 μm in diameter (PM10) has been hypothesized as a potential risk factor associated with an increased incidence of appendicitis. Previous studies have also investigated the link between air pollution and appendicitis [6,7].
PM10 is defined as fine dust composed of particles ≤10 μm in diameter and is one of the most well-known air pollutants, along with fine particulate matter (PM2.5), sulfur dioxide (SO2), nitrogen dioxide (NO2), ozone (O3), and carbon monoxide (CO). Since the Korean peninsula is exposed to relatively high levels of PM10 due to geopolitical reasons, the health effects of PM10 on the population have become particularly apparent in recent years [8]. Previous studies have shown that short-term exposure to pollutants can trigger inflammatory processes, potentially contributing to the development of appendicitis [9].
Methods
This study was reviewed and approved by the Ewha Womans University Mokdong Hospital Institutional Review Board (IRB File No: SEUMC 2020-08-026). The requirement for informed consent was waived.
We conducted a time-stratified case-crossover study design by linking the PM10 level during the case event (the date of hospitalization) to each appendicitis case. The study was described according to the STROBE statement, available at https://www.strobe-statement.org/.
The electronic medical records of all patients diagnosed with acute appendicitis and hospitalized at Ewha Womans University Mokdong Hospital over an 18-year period were collected. Additionally, national Korean air pollution data for the same period were gathered.
To derive time-stratified matched control events (when no admission occurred), we selected control period dates using the same year, month, and day of the week as the appendicitis hospitalization date, but from different weeks. These lags in exposure were referred to as same-day exposure and exposure lagged by "n days" before the event. The control events were matched with case events on the same day of the week to avoid time trend bias associated with specific weekdays.
Since this study employed a case-crossover analysis, intra-individual comparisons were conducted without accounting for the confounding effects of other risk factors, such as a patient's lifestyle [10]. Instead, we focused on weather variables as potential confounding factors, including daily mean temperature and relative humidity.
We collected data from 9,886 patients treated for appendicitis at Ewha Womans University Mokdong Hospital, a tertiary medical center in the western part of Seoul, Korea, from January 1, 2001, to December 31, 2018. The information gathered included registration number, gender, age, number of hospitalization days, admission and discharge dates, and residential address. Appendicitis is classified under the International Classification of Disease (ICD) ninth revision (ICD-9) codes 540.9, 540.0, 540.1, or 10th revision (ICD-10) codes K35.0, K35.1, K35.9, along with the in-hospital surgery code 470.
Patients were excluded if they had two or more duplicate records of surgical treatments or if their only surgical records were from the first visit (n=1,107). Additionally, patients were excluded if they had missing data on sex, age, hospitalization date, and/or address (n=1,100). Finally, those missing data on the case period, specifically the date of hospitalization due to appendicitis, were also excluded (n=1,153). This resulted in a total of 6,526 patients being included in the final analysis (Fig. 1).
The primary outcome was the PM10 exposure level at the time of appendicitis diagnosis.
Korea's metropolitan area spans 11,861 km² and includes Seoul, Incheon, and Gyeonggi-do Provinces. In 2018, 133 monitoring stations were established throughout this region. These stations are part of a national air pollution surveillance network known as "AirKorea," which oversees 240 measurement branches nationwide (https://www.airkorea.or.kr/). The AirKorea database records hourly mean concentrations of air pollutants from continuous monitoring stations, and calculates daily mean and maximum values. PM10 measurements were conducted using the beta-ray absorption method, as detailed in previous studies [11]. We linked the exposure data from these monitoring stations to the nearest administrative area corresponding to each patient's residential location. For the two-pollutant model, data were also collected on four additional air pollutants: SO2, NO2, CO, and O3. To assess the effect size and perform sensitivity analysis, we gathered PM2.5 measurement data from 2015 to 2018.
Meteorological data, including the daily mean values of PM10, temperature, and humidity, were obtained from the Korea Meteorological Administration (http://www.kma.go.kr).
Sample size estimation was not performed because this study included all target patients who met the exclusion criteria.
We analyzed the association between PM10 exposure and acute appendicitis using conditional logistic regression (CLR), which is an expanded logistic regression method that accounts for several control periods. This model is particularly effective for case-crossover studies as it accommodates the matched case and control periods within each subject. The CLR model is beneficial in this context because it extends the logistic regression framework to accommodate matched case-control data. This allows for the estimation of exposure-outcome associations while considering the matching structure of the data. Specifically, the CLR model calculates the OR for the occurrence of an event following exposure, taking into account the individual matching factors [10]. For each case of appendicitis, we matched the day of the appendicitis event with four control periods at weekly intervals, ranging from one to four weeks prior to the event. We then calculated ORs with 95% CIs to assess the relationship between an increase in the interquartile range of PM10 levels and the incidence of appendicitis.
We analyzed both the single lag effects (from lag 0 to lag 14) and the moving-average effects (from lag 0−1 to lag 0−14). Subgroup analyses were conducted based on sex (boys/girls), season (warm season: April‒September; cold season: October‒December), and age groups (under 10, 10–19, 20–29, 30–39, 40–49, 50–59, and 60 years or older). The seasonal categorization reflects the climatic patterns typical of Korea. The selection of specific lag periods was informed by the biological likelihood that the inflammatory response to air pollution exposure could manifest within a few days.
We calculated Spearman’s correlation between air pollutants before conducting the two-pollutants model. Pairs of exposure variables that demonstrated a high correlation coefficient (greater than 0.7) were excluded from this analysis. In the two-pollutant model, PM10 was set as the main exposure while other air pollutants (SO2, NO2, CO, and O3) were adjusted respectively.
When considering the delay between the actual case date and the date of hospitalization, we calculated the PM10 concentrations on the hospital admission day and for 7 and 14 days before admission to evaluate the cumulative effect over several days. The ORs for temperature and humidity were also adjusted and we included the daily average humidity and temperature as confounding variables. The lag effect of exposure considered the moving-average effect. The main results were for those on the The main results were for those on the same-day, 3 day, 7 day, and 14 day moving averages. All the data preprocessing and statistical analysis were performed using R statistical software (Ver. 4.0.0, R Development Core Team, Vienna, Austria) and SAS 9.4 (SAS Institute, Cary, NC, USA), and the α level for statistical significance was 0.05.
Results
A total of 6,526 appendicitis patients were included in the analysis after excluding those with insufficient essential information, such as duplicated, missing, or unmatched data (Fig. 1). The majority of the study participants resided in Seoul, with the administrative areas detailed in Supplement 1. Overall, 51.09% of the study subjects were male (Table 1); however, gender did not significantly influence the risk of appendicitis (P=0.7141). Analysis of age groups revealed that individuals aged 10–19 years (20.85%) were most susceptible to appendicitis, while those aged 60 years and older (8.99%) were least susceptible. Nonetheless, no significant association was found between age and the risk of appendicitis. The number of appendicitis patients hospitalized during the cold season (52.94%) was higher compared to those admitted in the warm season (47.06%). However, no significant association was observed between the season and the risk of appendicitis (P=0.7408).
Table 2 presents the summary statistics for the exposure data collected during the 2001–2018 study period. We assessed the level of environmental exposure in the case and control groups using the t-test and observed no significant difference in PM10 levels between the case and control periods. Supplement 2 contains the summary statistics for daily air pollutant exposure levels measured throughout the study period.
Table 3 summarizes the results of the case-crossover analysis examining the association between PM10 exposure and the risk of appendicitis. The analysis found an association between the risk of appendicitis and the average PM10 concentrations 3 days (OR, 1.045; 95% CI, 1.007–1.084) and 7 days (OR, 1.053; 95% CI, 1.005–1.103) prior to hospital admission. However, no association was observed with PM10 concentrations on the day of admission or 14 days before admission.
† PM10 exposure level between current hospital admission day and two days before hospital admission (lag 0‒2).
‡ PM10 exposure level between current hospital admission day and 6 days before hospital admission (lag 0‒6).
Sex differences: In male patients, the risk of appendicitis was significantly associated with the mean PM10 concentrations 3 days (OR, 1.076; 95% CI, 1.023–1.132) and 7 days (OR, 1.103; 95% CI, 1.035–1.176) prior to admission. However, no significant association was observed in female patients at any lag time.
Age differences: Only patients under the age of 10 demonstrated an association between the risk of appendicitis and the PM10 concentrations at the time of admission (OR, 1.129; 95% CI, 1.016–1.255), 3 days before admission (OR, 1.140; 95% CI, 1.033–1.258), and 7 days before admission (OR, 1.235; 95% CI, 1.087–1.402).
Seasonal differences: The risk of appendicitis during the cold season was significantly associated with the mean PM10 concentration at various lag times, including at the time of admission (OR, 1.252; 95% CI, 1.192–1.316), three days prior (OR, 1.424; 95% CI, 1.343–1.509), seven days prior (OR, 1.716; 95% CI, 1.593–1.849), and 14 days prior (OR, 2.632; 95% CI, 2.334–2.967). Conversely, no significant association was found between PM10 concentration and increased appendicitis risk during the warm season.
Fig. 2 depicts the risk of appendicitis across various reference time intervals, highlighting the associated interquartile ranges of PM10 through case-crossover analysis. This analysis includes data from the total patient population, patients aged 0–9, males, and those observed during the cold season over various exposure time intervals. The risk of appendicitis was linked to PM10 concentrations when the exposure period ranged from 3 to 7 days prior to hospital admission. In subgroup analyses, a significant association was found between the risk of appendicitis and PM10 concentrations during an exposure interval of two to nine days for male patients and those under 10 years of age. Additionally, during the cold season, the risk of appendicitis increased as the exposure interval extended from 1 day to 14 days.
The overall risk of appendicitis during the various reference time intervals associated with increases in the interquartile ranges of PM10, stratified by sex, season, and age are shown in Tables 4, 5. In males, the risk of appendicitis consistently increased over the case period from one to six days. During the cold season, the longer the case period, the higher the risk of appendicitis, with the greatest risk observed at a 14-day interval (OR, 2.632; 95% CI, 2.334–2.967). The correlation between PM10 and other air pollutants is summarized in Supplement 3. Since no air pollutant demonstrated a high correlation with PM10, no variable was excluded from the two-pollutant model. We observed varying trends in the association between PM10 and appendicitis across two-pollutant models. While the association remained statistically significant when adjusted for SO2, CO, and O3, it was attenuated upon adjustment for NO2 (Table 6). Since the AirKorea database began recording PM2.5 measurements in 2015, we included a sensitivity analysis to assess the impact of PM2.5 on the study results using a two-pollutant model with both PM10 and PM2.5, detailed in Supplement 4.
Discussion
The present study suggests that short-term exposure to PM10 is significantly associated with an increased risk of appendicitis in boys under the age of 10 who were hospitalized during the cold season. To the best of our knowledge, this study is the first to examine the impacts of PM10 exposure on acute appendicitis across diverse subgroups and various lag intervals within the Korean population.
Ewha Womans University Mokdong Hospital, affiliated with Ewha Medical School, serves as the sole community-based tertiary medical institution for a population of 1.5 million in the Yangcheon-gu district of Seoul. Each year, the hospital performs approximately 500 to 600 appendectomies, with the majority of these cases being emergency surgeries for acute appendicitis. Over the 18-year study period, about 72% of the appendicitis patients came from three neighboring administrative districts. In terms of exposure validity, we accessed national air pollution data from AirKorea for analysis. We then correlated this data with our patient records, matching it to the inpatients' home addresses. As for the validity of our outcome definition, all study participants underwent an appendectomy following their diagnosis. This procedure followed a standardized method for critical pathways and was verified by biopsy, ensuring the accuracy and consistency of our outcome definitions.
Regarding sex differences, biological differences, such as variations in immune response and hormonal influences, may make males more susceptible to inflammatory triggers caused by air pollution. Additionally, behavioral factors, including differences in outdoor activities and exposure levels, could contribute to the observed disparity [12].
Children under 10 may be more vulnerable to PM10 exposure due to their developing respiratory and immune systems, which are more susceptible to inflammatory agents. Their higher breathing rates relative to body size and increased time spent outdoors can lead to greater exposure to air pollutants. Additionally, because their immune systems are not fully developed, they are less capable of handling environmental insults such as air pollution [13]. Cold weather can exacerbate the effects of PM10 on respiratory and immune systems, potentially triggering inflammatory responses more readily [14].
This study observed that seasonal variations in air pollution exposure levels significantly impacted the risk of appendicitis. Interestingly, our findings contradict those of previous studies, which suggested that appendicitis incidence was higher in the summer due to increased outdoor activities [15]. Our results lead us to hypothesize that climate factors, such as temperature and humidity, play a more significant role in increasing appendicitis risk than does outdoor activity. Additionally, the distinct weather conditions associated with Korea's four seasons may also influence the results observed in our study.
Our study has several key strengths and novelties. First, it specifically targeted the Korean population. Despite Korea's relatively high air pollution levels compared to Western countries, to our knowledge, this is the first study utilizing Korean data to explore the relationship between PM10 and appendicitis. Second, our methodology is distinctive as it aimed to categorize cases across various dimensions, including age, sex, and season. While appendicitis is recognized as an acute condition linked to short-term effects of PM10, the critical exposure period remains unclear. Therefore, diverging from previous research that primarily examined a lag effect of fewer than 3 days, our study investigated the impact of PM10 over a 14-day period prior to hospital admission.
In Linfen City, China, an increase of 10 μg/m3 in pollutant levels, considering a 1-day lag, was associated with heightened health risks from January to December 2018. Specifically, the relative risks and their 95% CIs were as follows: PM10: 1.0179 (1.0129–1.0230), SO2: 1.0236 (1.0184–1.0288), and NO2: 1.0979 (1.0704–1.1262). The study indicated that men and young adults aged 21–39 years were particularly susceptible to the effects of air pollution. Furthermore, the impact of air pollutants was more pronounced during the colder months, though the seasonal variation was not statistically significant [16]. In Italy, factors predicting perforated appendicitis included consultation delay (OR, 1.621; 95% CI, 1.288–2.039; P<0.001) and the 2-day lag mean concentration of PM10 (OR, 1.066; 95% CI, 0.007–1.130; P=0.029) during the period from January 1 to December 31, 2014 [17]. In Taiwan, when temperatures fell below 23℃, higher levels of PM10 were linked to a significant increase in hospital admissions for appendicitis between 2009 and 2013 [7]. In New Zealand, no correlation was found between PM10 levels and admissions for appendicitis from January to December 2018 [18]. While the above-mentioned three studies found associations between PM10 and appendicitis admissions, other results concerning age and climate varied.
Our study has several limitations. First, since nationwide air pollutant data are only available by district, we could not account for variations between smaller regional units. Consequently, we cannot ensure a precise match between the pollution data from our database and the actual pollution levels in the patients' neighborhoods.
Second, we did not take into account the socioeconomic and lifestyle data of the patients; therefore, we did not categorize our patients as outdoor or indoor workers to explore any potential association between outdoor activity and appendicitis. However, it is important to note that appendicitis is an acute condition, and the data were gathered from a single institution located near where most patients reside. We assumed that the levels of air pollution exposure were varied across different neighborhoods. Additionally, the use of a case-crossover data analysis helped to eliminate bias from personal confounders.
Third, since our AirKorea database only began recording PM2.5 measurements in 2015, we were unable to assess the effects of PM2.5 across all participants. Instead, we conducted a two-pollutant model analysis using PM10 and PM2.5 data from cases recruited between 2015 and 2018. Our findings indicate that the overall effect size of PM2.5 was greater than that of PM10, and it was associated with smaller P-values. Therefore, further studies that focus on PM2.5 as the primary exposure are warranted.
These comprehensive analyses may offer additional insights into preventive measures for appendicitis that are typically overlooked in clinical practice. Further research is essential to enhance our understanding of appendicitis epidemiology and to help decrease the incidence of the condition.
Our study revealed a positive association between PM10 concentration and the incidence of appendicitis, suggesting that short-term exposure to PM10 may trigger appendicitis. The risk was notably higher in boys under the age of 10 and during the colder seasons. However, these findings should not be interpreted as direct evidence that PM10 directly causes appendicitis. To further investigate these relationships, additional ecological and large-scale epidemiological studies are necessary.
