Published online May 14, 2022. doi: 10.3748/wjg.v28.i18.2008
Peer-review started: November 14, 2021
First decision: January 9, 2022
Revised: January 22, 2022
Accepted: March 26, 2022
Article in press: March 26, 2022
Published online: May 14, 2022
Limited information is available on the risk factors of gastrointestinal bleedings (GIB) in the general population. Previous research mainly focused on the population using specific medications such as antiplatelets, anticoagulants etc. or the critically ill population such as hospitalized patients or elderly. Therefore, we investigated the risk factors for major GIB using data representative of the general population having a high granularity of the data source and long-term follow-up of participants.
Many studies investigating risk factors of GIB lacked information on lifestyle factors, clinical measurements and laboratory parameters. We wanted to better understand the effect of these additional variables on GIB as well as how this affects established risk factors. Additionally, to better understand the factors that predict the risk of GIB in the presence of a multicollinear set of variables [e.g., Body mass index (BMI), Waist-hip ratio (WHR) or physical activity].
The overall objective of the study was to identify and investigate new risk factors of major GIB in the general population of Finland considering established risk factors as well as demographics and morbidities. We were able to identify new risk factors of major GIB together with established risk factors which needs to be evaluated and confirmed by further research.
We conducted a retrospective cohort study using record linkage of data from the FINRISK health examination surveys which are representative of the general population of Finland to the national electronic health registers with 10 years of follow-up. This linkage enabled us to include and investigate demographics, socioeconomic and lifestyle factors, clinical measurements, laboratory parameters, and comorbidities on the risk of major GIB. We further implemented Least Absolute Shrinkage and Selection Operator (LASSO) to select the most important predictor variables for model prediction and association of these predictor variables were evaluated using Cox regression. The novelty of using LASSO is that it helps in the variable selection and in excluding unnecessary/uninfluential variables from the model thus reducing the likelihood of overfitting a model. It also helps to address multicollinearity that can be problematic in the traditional forms of regression.
The main results of the study showed that baseline age, unemployment, and higher BMI, higher gamma-glutamyl transferase (GGT) levels, having ≥1 precursor of GIB, previous cancer, psychiatric disorders, heart failure and liver disorders were all associated with an increased risk of GIB. Systolic blood pressure, above average coffee consumption per day, and history of osteoarthritis were all associated with a decreased risk of GIB. This study adds to the scarce literature on risk factors on gastrointestinal bleeding in the general population. Additionally, results are hypothesis generating for the new risk factors identified in this study which must be confirmed by future mechanistic and epidemiological studies.
This study identified new risk factors associated with major GIB which are unemployment, BMI, GGT, SBP and coffee consumption. Accounting for physical activity and waist-hip ratio, our study suggests that BMI is a better predictor of major GIB. Above average coffee consumption per day, which seems to be more common in Finland with the highest per capita coffee consumption in the world, was associated with a decreased risk of major GIB. Our study suggests that the risk factors of major GIBs might be slightly different in the general population than the at-risk population.
Future mechanistic and epidemiological studies should evaluate these risk factors in different study populations or countries across the world to establish causal associations. This will further support in complementing and refining existing risk scores for major GIBs.