Prevalence and trigger factors of functional gastrointestinal disorders among male civil pilots in China

Functional gastrointestinal disorders (FGIDs) are common among the aircrew due to their arduous working environment. This study investigated the prevalence of FGIDs in Chinese male pilots and assessed the effects of trigger factors on the FGIDs. A cross-sectional study including 212 male pilots was performed in a Chinese large civil airline company. FGIDs were diagnosed according to the Rome IV diagnostic criteria. The psychological performance, dietary pattern, sleep situation, and physical activity of the respondents were assessed. Logistic regression analysis and structural equation modeling were used to explore the association between these trigger factors and FGIDs. FGIDs were observed in 83 (39.22%) respondents, of which 31 (37.35%) had overlap syndromes. Age, flight level, flight time, high-salt food pattern, anxiety, and sleep performance were found to be associated with FGIDs (all P < 0.05). Stepwise logistic regression analysis revealed that the flight level (OR 0.59, 95% CI 0.31–0.080), high-salt food pattern (OR 2.31, 95% CI 1.28–4.16), and sleep performance (OR 2.39, 95% CI 1.11–5.14) were the influencing factors associated with FGIDs. Structural equation modeling confirmed the correlations between FGIDs and the occupational, dietary, and psychological factors with a reasonable fit. The preventive strategies were necessitated according to occupational and psychological characteristics.

Diet. The semi-quantitative food frequency questionnaire was designed to collect the food list, frequency of consumption and the portion size consumed 30 . Referenced a lot of Chinese food frequency questionnaire [31][32][33] , we selected 27 food categories often eaten in Chinese dietary, which covered cereals, meat, vegetables, fruits, egg products, nuts, fish, and so on. There was also an open question to detect the subject's favor food which not contained in the given 27 food categories. The intake frequencies were classified into eight levels: almost never; less than one time per month; 1-3 times per month; one times per week; 2-4 times per week; 5-6 times per week; one time per day; 2-3 times per day and 4 or more times per day. There was no unified portion size in China, and therefore, food types and portion size of local food were surveyed with visual aids, which consisted of photographs of utensils and food portions to assist with description of amounts consumed. Options for the average portion sizes were 0.5, 1, 1.5, 2. Based on the food composition and food model 34,35 , it investigated the food intake of each subject in the recent half year. And a 3-day 24-h dietary recalls (24 h DRs) was also used to evaluate the reliability and validity of this SQFFQ.
The diet pattern was identified by exploration factor analysis 36 . The common factors with Eigenvalues > 1.0 were extracted based on the screen plot. An orthogonal rotation procedure was applied to simplify the factor structure and render it more easily interpretable. The derived factors were named as the different dietary patterns according to the food categories that loaded most strongly on the factor. For each participant, the factor score for each pattern was calculated by summing the quantity of each food grouping weighted by their loading on each factor. The dietary pattern scores were expressed as three quartiles for comparison across levels of intake.
Psychology. The evaluations related to psychology, sleep, and physical activity were performed according to the corresponding standard criteria. Specifically, the respondents with the SAS index scored above 50 and the SDS index scored above 0.50 were defined as having anxiety and depression, respectively 26,27 . A PSQI score of 1-5 was considered as sleep quality, 6-10 as sleep latency, 11-15 as sleep duration, and 15-21 as sleep disturbance 28 . The physical activity intensity was divided as low (the total physical activity score below 600), median (the total physical activity score between 600 and 3000), and high (the total physical activity score above 3000) 29 . , was applied to explore the associations between latent physical and mental factors and the developing of FGIDs. A two-tailed P value of less than 0.05 was considered to be statistically significant.

Results
Characteristics of participants. A total of 212 male pilots were sampled in this cross-sectional study.
All the participants completed their questionnaire with a response rate of 100%. As shown in Table 1 Table 2). The prevalence of individual FGID is as follow. Globus (7.08%), Functional dyspepsia (FD: 13.21%), and IBS (9.43%) were the most frequent diagnoses among bowel disorders, gastroduodenal disorders and esophageal disorders respectively, whereas FC (4.25%), functional chest pain (FCP: 3.77%), and functional abdominal bloating (FAB: 4.72%) also had a moderate prevalence. Only two aviators had rumination syndrome, and two other aviators had central-mediated abdominal pain syndrome.
As for the overlap syndrome, 31 pilots met the criteria for more than one FGID symptom. The categories of FGIDs occurred in different combinations were shown as two-way combination 8.49%, three-way combination 3.33% and four-way combination 2.83% (Table 2 and Fig. 1). Bowel disorders and gastroduodenal disorders were the most prevalent (4.72%) in the two-way combination, whereas central mediated disorders of gastrointestinal pain had no combination. For the detailed syndrome, globus + nausea and vomiting disorders (NVD), FD + IBS and FD + FAB led the two-way combination (1.89%). Among the seven pilots with three-way combined FGIDs (4.72%), four were diagnosed with FD + IBS + globus. While, the six pilots with four-way overlap syndromes all manifested IBS syndromes.
Diet pattern. The SQFFQ demonstrated acceptable relative validity and high reproducibility with intraclass correlation coefficients (ICCs) of 0.44-0.77, spearman's correlation coefficients of 0.35-0.62 and the three-point quartile agreement of 72.0-95.0% in the same or adjacent quartiles for the food categories referenced 24 h DRs. Via the three-point semi-quantitative food frequency questionnaire, four dietary patterns were retained for analysis as follow: the vegetarian pattern characterized by green vegetables and fruits, the high-salt food pattern characterized by Chinese pickles and bacon, the starch food pattern characterized by the Chinese traditional staple food of rice and noodles and the protein food pattern characterized by high-quality protein products such as fish, milk and mushrooms.
The dietary pattern scores were expressed as three quartiles for comparing the levels of intake between different FGID groups, as shown in Table 3. The vegetarian pattern, the starch food pattern and protein food pattern were not related to the distribution of FGIDs. However, respondents with high-salt food pattern scores above 66.7% were more likely to be identified with FGIDs than those with lower scores (P < 0.001). The odds ratio [OR 2.46, 95% confidence level (CI) 1.71-3.35] indicated that high-salt food pattern was positively associated with FGIDs; for every 33% increase in the high-salt food pattern score, the risk of FGIDs might double. a The reference is the low-level group.
Psychological situation. According to the evaluation criteria of the SAS 27 , the anxiety scores were different between the three groups, as shown in Table 4 (P = 0.006). The overlap syndrome group had the highest anxiety score (48.75 ± 13.30), followed by the single FGID syndrome group (44.48 ± 6.46). In the post hoc analysis, the anxiety score was significantly higher in the overlap syndrome group than in the healthy group, which meant that the more FGID syndromes the respondents had, the higher anxiety scores they got. The depression and physical activity situation showed no significant difference between the three groups. According to the criteria of PSQI (short version), the single syndrome group had the highest sleep score (11.78 ± 3.31), followed by the combined syndrome group (11.27 ± 2.82) and the healthy control group (10.26 ± 2.84) (P = 0.005). The score of the single syndrome group was significantly higher than that of the control group. www.nature.com/scientificreports/ Table 5 shows a number of variables significantly associated with the prevalence of FGIDs. The respondents over 35 years were more likely to have FGIDs. For the occupation factors, the flight levels were adversely affected by the FGIDs. The captain had the highest prevalence of FGIDs, followed by the first and second officer. The aircrew flying for more than 20 h every month had almost one times higher prevalence of FGIDs than those flying for less than 10 h. The high-salt food pattern was positively associated with the development of FGIDs. The prevalence of the upper three-quartile score group was almost three times of those in the lower three-quartile score group (56.34% vs 19.18%), which meant that the respondents with high-salt food pattern were more likely to have FGIDs.

Associations between predictor variables and FGIDs.
The psychological factors and the FGIDs were also related. The prevalence of FGIDs in anxious aircrews was two times than that in aircrews without anxiety (P = 0.002). The depression and physical strength did not correlate with the FGIDs, but a good sleep quality might positively prevent the occurrence of FGIDs. The pilots with sleep scores of 16-21 had the highest prevalence of FGIDs, which was approximately three times than the prevalence in those with good sleep.
The multivariate logistic regression analysis showed that three predictors were found to independently and significantly correlate with the FGIDs: flight level (OR 0.59; 95% CI 0.31-0.80; P = 0.001), high-salt food pattern Table 1. Demographic characteristic of the study participants. Continuous variables were expressed by mean ± standard deviation and categorical variables were expressed by numbers (percentages). We also explored a subgroup analysis of the risk factors by FGID types. Given the relatively small number for each specific FGID symptom, this analysis was mainly conducted on the sub-type FGIDs as the esophageal disorders, gastroduodenal disorders, bowel disorders according to Rome IV diagnostic criteria. As shown in Table 6, the trigger factors for the subgroup syndromes were quite similar to those for FGIDs. For example, the higher flight level pilots with anxiety symptoms were more likely to have esophageal disorders. While, age, flight level, flight time per month, and high-salt food pattern were found to associate with the gastroduodenal disorders significantly. The older age, the high flight level, the high-salt food pattern, the anxious psychology and the poor sleep performance might be the potential factors contributing to the bowel disorders.

Structural equation model.
A confirmatory factor analysis (CFA) using Structural equation model (SEM) was conducted to assess the linear association between the effects of the trigger factors on the FGIDs. Age, flight factors, food categories with a factor loading of more than 0.4 in the high-salt food pattern, and psychological factors were considered to be enrolled in the analysis. The resulting SEM is illustrated in Fig. 1, in which the chi-square test result was statistically significant (χ 2 = 223.72, P < 0.001), the comparative fit index was (0.97) > 0.90, the root-mean-square error of approximation was (0.02) < 0.06, and the degree of freedom ratio was (1.03) < 3.00. www.nature.com/scientificreports/ As Fig. 2

Discussion
As studies on the prevalence of FGIDs in Chinese aviators were very limited, this study was conducted to identify the effect of the trigger factors on the FGIDs for the civil pilots. We have just focused on the FGIDs of male pilots. That is because the prevalence of FGIDs varied substantially between female and male 37 , such as FD 22,37,38 and IBS 23,39,40 . While, the proportion of female in Chinese pilots was very low, which was reported as only 1.85% for the commercial certificate pilots and 0.33% for the aviation transport license pilots 24 . In this study, we had used the convenience sampling methodology to recruited the subjects. There were too few female pilots attending this research voluntarily in the recruiting phase. On the contrary, female mostly undertake the cabin crews in China. Although both of the pilots and cabin crew are exposed to the same working environment, the potential activity level and occupational variables are quite varied. Considering the gender and occupational difference, we do not include the female pilots in this study. Therefore, it should be cautious when generalizing the results to the characteristics of all the aircrews. In this study, 39.15% of the respondents had FGIDs. This value was high compared with that in some other studies, in which the prevalence of military aircrews was 7.2% in the USA 41 and 3.50% in China 14 . However, this was lower than those reported by Park (49.7%) 42 and Trivedi in their study including soldiers presenting to a home hospital (50%) 43 . The diversity in prevalence might be because the first survey on the Chinese commercial pilots used Rome IV criteria. It might also be because of the developments in the Chinese aviation and tourism industries, in which the aviator occupation required a highly efficient and error-free performance. The pilots undergo very stressful physical and psychiatric training, which may be the reason of the high prevalence of FGIDs among them. Although the pilots with organic esophageal or GI tract disease were excluded in this survey, the helicobacter pylori status were not evaluated among the participants, which were also an essential for differential diagnosis.
This study also found that FD was the most prevalent FGIDs overall (13.21%), followed by IBS (9.43%) and globus (7.08%). The main subtype rank was similar to that in other studies, such as Bang 44 , Wu 14 , and Dong 45 . However, the prevalence in this study was quite higher than these studies, which were reported as 5.85% and 4.04% for FD and IBS for the Chinese air forth pilots' prevalence in Wu's research 14 . It also showed a higher prevalence of FD and IBS than the general populations, in which the global prevalence of FD was reported as 6.9% (95 CI 5.7-8.2) from a meta-analysis in 2020 comprising 81,144 subjects from 4 studies with the ROME IV criteria 22 . Another meta-analysis comprising 82,476 individuals from 34 countries claimed the pooled IBS prevalence was 3.8% (95 CI 3.1-4.5) used the Rome IV criteria 23 . Although researches suggested the Rome IV criteria might be stricter for the diagnostic of FGIDs compare to the prior criterion 46,47 , the prevalence of FD and IBS in this study were higher than those in the systematic reviews. Therefore, cautious interpretations should be made regarding this aspect.
In terms of overlapping syndromes, the prevalence was estimated as 14.62% of the total population and 37.35% of the patients with FGIDs in this study. In previous studies, the prevalence was reported to range from 1 to 17% 3,48-51 , not much different from the results of this study. As for the most combination syndromes, this study showed that FD + IBS, FD + FAB, and NVD + globus (1.89%) were the most prevalent subtype constitutions, which was quite similar to the findings of studies on healthy military males from the USA 52 and Korea 53 .
However, it's worth noting that previous studies had found that there may be few difference for Rome IV criteria compared to the prior Rome criterion 46,54 . For example, a study recruited 1375 adults self-identified as having IBS found that the Rome IV criteria significantly under reported the prevalence of IBS in comparison to Rome III(59.1% vs. 78.9%) 47 . Some systematic review also reported that the pooled prevalence of uninvestigated dyspepsia and IBS were lower with Rome IV(dyspepsia: 6.9% vs. 11.5%; IBS: 3.8 vs. 9.2%) 22,23 . It seemed that Rome IV criteria were more restrictive. The agreement between Rome III and Rome IV is still needed to be www.nature.com/scientificreports/ detected further. Understanding the impact of these changes to the diagnostic classification system for FGIDs will be important in the future. In addition, the results also found several trigger factors related to FGIDs. They included occupational exhaustion, dietary pattern, psychological stress, and sleep performance. For the occupational environment, high-speed flying, constant exposure to rapid acceleration and deceleration, and noise and vibration might be the complicated mechanisms resulting in abdomen discomfort [55][56][57] . Therefore, the flight level and flight time were closely associated with the development of FGIDs.  57,58 , the effects of dietary patterns on the FGIDs were proved by the CFA and SEM results. Although carbohydrates, proteins, and individual fatty acids were the most frequently compounds that influenced the digestive functions, a positive association was found between the high-salt dietary pattern and the FGIDs in this study. The high-salt intake might be of particular importance for the composition and activity of intestinal microbiota 59 . Specific nutrients could change the microbial metabolic activity, leading to GI discomfort.
In agreement with the results of previous studies, psychological distress was found to be positive related to FGIDs and combined syndromes in our research. The anxiety factor was proved to be an independent risk factor for FGIDs, whereas no statistical correlation was observed between depression and FGIDs. This result was similar to a study of the brain-gut pathway 60 , but not coincident with some studies, which found that depression had a negative correlation with FGIDs 14,44 . The exact reason or mechanism was unclear. More studies are needed to elucidate the association between each psychological factor and FGIDs.
In conclusion, this is the first population-based survey using Rome IV criteria to evaluate FGIDs among Chinese pilots, which has found that there is a high overall prevalence of FGIDs among Chinese male pilots, especially on FD, IBS and Globus. The SEM analysis elaborated the effects of the flight factor, food pattern, psychology and sleep performance on the FGIDs. Therefore, further works can be conducted on more integrative prevention and treatment combining psychological and physical approaches for the developing and progression of FGIDs.
This study also had some limitations. First, it could not identify a causal relationship between some influence factors and FGIDs, because it was a cross-sectional study with subjective bias. Second, the information was obtained using self-administered questionnaires, in which may conceal a possibility of subjective bias. Third, only male aviators were included in this study, generalizing the results to the characteristics of all commercial pilots with FGIDs was difficult. More researches are required on a wide range of respondents and the various trigger factors in the future.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.