Abstract
The aim of this study was to apply a 2-sample Mendelian randomization approach, uncovering the causal link between artificial sweetener (AS) intake in cereals, coffee and tea and cancer. For this purpose, the research team obtained and utilized comprehensive data from published genome-wide association studies and selected genetic variants associated with AS as instrumental variables. Through 5 analytical techniques such as inverse variance weighting, MR-Egger, weighted median, simple multiple, and weighted multiple. In this study, the potential risk associations between AS and 15 cancers were comprehensively evaluated, and sensitivity analyses were used to assess the robustness of the results. Statistical results showed that there were causal associations between AS and breast cancer, oral cancer, esophageal cancer, gastric cancer, colorectal cancer, prostate cancer, ovarian cancer, kidney cancer, and lung cancer (inverse variance weighting method: P < .05). AS was a protective factor for breast cancer, gastric cancer, colorectal cancer, prostate cancer, and kidney cancer, and reduced their incidence. AS is a risk factor for oral cancer, esophageal cancer, ovarian cancer, and lung cancer, which increases the probability of occurrence. Sensitivity analysis showed that there was no significant heterogeneity or pleiotropy, and the results were relatively stable. AS intake was associated with a reduced risk of breast, stomach, colorectal, prostate, and kidney cancers; AS intake was associated with an increased risk of oral, esophageal, ovarian, and lung cancers.
Keywords: artificial sweeteners, cancers, causality, Mendelian randomization
1. Introduction
Artificial sweeteners (AS) are sweeteners used in food and beverages that provide sweetness but far fewer calories than conventional sugar. As people become more health conscious, the demand for sweet flavors is gradually increasing and therefore more emphasis is being placed on consuming low-calorie and sugar-free products.[1] Artificial sweeteners are becoming increasingly popular as a sugar substitute for reducing calorie intake. The most common AS include aspartame, saccharin, soda, erythritol, acesulfame and sucralose, which are commonly used in foods and beverages such as cereals, coffee and tea,[2] to satisfy the need for sweetness. The effect of consumption of artificial sweeteners on cancer risk has been debated in conjunction with the prolific use of AS; however, results of observational studies examining the association between AS and cancer risk have often demonstrated inconsistency. Certain studies report that the use of non-sugar sweeteners may increase the risk of cancer.[3] However, another study showed a significant inverse trend of increased risk in the total sweetener category for breast and ovarian cancer.[4] In addition to human studies, there have been studies in animals. The results of a meta-analysis on the carcinogenic effects of aspartame in rodents showed that consumption of aspartame did not produce significant carcinogenic effects in rodents.[5] Although artificial sweeteners used in food have been declared safe for use with no apparent health risk as long as they are below the acceptable daily intake, respectively,[6] the carcinogenic effects of artificial sweeteners remain controversial. This Mendelian randomization analysis summarized data on artificial sweeteners and various cancers to determine the association between artificial sweeteners and cancer risk.
According to the World Health Organization, cancer is one of the leading causes of death globally, with about 10 million deaths in 2020, and the most common cancers include breast, lung, colorectal, rectal, and prostate cancers.[7] In China, the incidence of and deaths from malignant tumors have also continued to rise, becoming a major public health problem. The exploration of tumor risk factors, such as tobacco use, alcohol consumption, unhealthy diet, lack of physical activity, and air pollution, are the main research directions. In addition, lifestyle choices or behavioral patterns remain the most important factor influencing cancer risk, but environmental exposure to certain chemicals, both man-made and natural, may also increase an individual’s likelihood of developing cancer.[8] The role of factors such as artificial sweeteners, coupled with an aging population and shifting lifestyle patterns, has led to an increasing prevalence of cancer in recent years. This paper focuses on the causal association between AS risk factors and the risk of tumorigenesis, and the prevention of tumorigenesis to enhance people’s quality of life by avoiding risk factors.
Mendelian randomization (MR) studies are a method of assessing the causal relationship between various exposures and disease outcomes using genetic variation as an instrumental variable (IV).[9] This analysis makes causal inferences by using genetic variation as a proxy for modifiable risk factors or health outcomes.[10] Since genes are randomly assigned from conception, genetic variation is largely independent of other factors.[11] MR studies need to fulfill 3 basic assumptions[12]: the assumption of relevance: genetic variation is associated with risk factors; the assumption of independence: genetic variation is independent of any known or unknown confounders; and the assumption of exclusivity: genetic variation affects outcomes only through risk factors.
Two-sample Mendelian randomization involves 2 different study populations. For example, data on AS are measured in 1 sample and data on malignant tumors are measured in the other sample. This design has 2 advantages: first, all MR studies do not need to collect risk factors and outcomes. Two, the results can be very large (usually >50,000), allowing the use of pooled results from genome-wide association studies, and therefore the results are precise and statistically very efficient. At the same time, it can also solve the problem of difficult or costly sample collection.[13] In this study, we used MR methods to explore the potential causal relationship between AS and cancer-related genetic variants.
2. Materials and methods
2.1. Data sources for MR analysis
Genome-wide association study (GWAS) data on intake of AS in cereals, coffee, and tea are from UK Biobank[14] including 64,949 European individuals with 9,851,867 single nucleotide polymorphisms (SNPs). Data for the 15 malignancies studied were derived from GWAS summary data, which included European individual. Summary data are publicly available from https://gwas.mrcieu.ac.uk/datasets/.
As the data used in this MR study were from publicly available databases, no additional ethical approval was required. Information on the exposure factors AS, outcome 15 malignancies used in this study is detailed in Table 1.
Table 1.
Details of samples included in MR study.
| Variant | Sample size (n) | Case group (n) | Control subjects (n) | Place of origin of the sample | Year of publication | PubMed ID | Website |
|---|---|---|---|---|---|---|---|
| AS intake in cereal | 64,949 | / | / | European | 2018 | 36402876 | https://gwas.mrcieu.ac.uk/datasets/ukb-b-3143/ |
| AS intake in coffee | 64,949 | / | / | European | 2018 | 36402876 | https://gwas.mrcieu.ac.uk/datasets/ukb-b-1338/ |
| AS intake in tea | 64,949 | / | / | European | 2018 | 36402876 | https://gwas.mrcieu.ac.uk/datasets/ukb-b-5867/ |
| Breast cancer | 337,159 | 7480 | 329,679 | European | 2017 | / | https://gwas.mrcieu.ac.uk/datasets/ukb-a-55/ |
| Oral cavity cancer | 3464 | 1135 | 2329 | European | 2016 | 27749845 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST012238/ |
| Esophageal cancer | 160,589 | 1388 | 159,201 | East Asian | 2021 | 34594039 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90018621/ |
| Gastric cancer | 476,116 | 1029 | 475,087 | European | 2021 | 34594039 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90018849/ |
| Colorectal cancer | 377,673 | 5657 | 372,016 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/ieu-b-4965/ |
| Bladder cancer | 373,295 | 1279 | 372,016 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/ieu-b-4874/ |
| Prostate cancer | 292,053 | 22,534 | 270,176 | European | 2017 | https://gwas.mrcieu.ac.uk/datasets/ukb-a-204/ | |
| Endometrial cancer | 54,884 | 8758 | 46,126 | European | 2018 | 30093612 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST006465/ |
| Cervical cancer | 199,086 | 563 | 198,523 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/ieu-b-4876/ |
| Ovarian cancer | 199,741 | 1218 | 198,523 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/ieu-b-4963/ |
| Kidney cancer | 174,006 | 971 | 174,006 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/finn-b-C3_KIDNEY_NOTRENALPELVIS_EXALLC/ |
| Lung cancer | 492,803 | 3791 | 489,012 | European | 2021 | 34594039 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90018875/ |
| Liver and bile duct cancer | 372,366 | 350 | 372,016 | European | 2021 | / | https://gwas.mrcieu.ac.uk/datasets/ieu-b-4915/ |
| Pancreatic cancer | 159,700 | 499 | 159,201 | East Asian | 2021 | 34594039 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90018673/ |
| Thyroid cancer | 407,746 | / | / | European | 2021 | 34017140 | https://gwas.mrcieu.ac.uk/datasets/ebi-a-GCST90013863/ |
AS = artificial sweeteners, MR = Mendelian randomization.
2.2. Study design
This study used a 2-sample Mendelian randomization approach with AS in cereals, coffee, and tea as exposure factors. Genetically variant SNPs were used as instrumental variables to validate the causal relationship between the genetically predicted exposure factor AS and the outcome malignancy, using malignancy as the outcome factor. MR analyses need to follow 3 core assumptions[15]: the relevance assumption requires that the SNPs used as instrumental variables must have a strong correlation with the exposure factors of interest to the study. The independence assumption assumes that the selected genetic variants remain independent without association with those confounders that may affect exposure and outcome. The exclusivity assumption specifies that the effect of SNPs on the outcome variable can only be realized through that exposure factor and not through any other route (see Fig. 1).
Figure 1.
Flow chart of 2-sample MR analysis. MR = Mendelian randomization.
2.3. Selection of instrumental variables
SNPs need to meet the following criteria: SNPs must have a significance threshold of <1.0 × 10‐6 at the genome-wide level, which indicates that they are significantly correlated with AS; Set the SNPs to have a degree of explained variation (R2) of at least 0.001 or a genetic distance of no more than 10,000 kb, and remove those that have an R2 value of >0.001 within a distance of 10,000 kb from the most significant SNPs. That is, exclude SNPs that are in a state of linkage disequilibrium. Exclude SNPs with palindromic structures that may affect the accuracy of the analysis. SNPs with F-statistics <10 were excluded to filter out potentially weak instrumental variables. Weak instrumental variables, which refer to genetic variants whose explanatory power for the exposure factor under study is weak, despite their correlation with the exposure factor under study. Because of their small effect on the exposure factor, these variants are virtually useless in providing the statistical power needed to test the study assumption, which may result in impaired precision of causal effect estimates. Weak instrumental variables can lead to problems such as increased risk of false positive errors (i.e., type I errors).
The strength of instrumental variables can be assessed by the F statistic, which can be calculated for individual SNPs using the following formula: The F statistic was calculated as F = R2 × (n ‐ k ‐ 1) ÷ [k × (1-R2)], where n denotes the sample size, k denotes the number of instrumental variables used, and R2 reflects the extent to which the instrumental variables explain the exposure. R2 = 2 × (1 ‐ MAF) × MAF × β2, where MAF is the minimum allele frequency, and β is the allele effect value. Less affected by weak instrumental bias when F > 10.[16] Based on the GWAS results for AS and malignancy, we can synergistically collate SNPs with the same allele so that the effect values for exposure and outcome correspond to each other.[17]
2.4. Statistical treatment
In this study, the inverse variance weighting (IVW) method was used as the primary research method to assess the potential causal association between AS and malignancy. This method is best suited for situations where genotypes are highly balanced with exposure factors and can produce unbiased estimates. This study utilized the intercept test in the MR-Egger method to detect the presence of horizontal multivariate bias in the instrumental variables, and if P > .05, then there is no horizontal multivariate.[18] In addition, the study conducted a leave-one-out analysis to assess the effect of individual SNPs on significant effects. Cochran Q test was used in this study to detect heterogeneity among instrumental variables. When the P-value is <.05, it indicates significant heterogeneity, at which point the IVW of the random effects model is recommended for analysis.[19] For sensitivity analysis, we used the leave-one-out method, that is, after removing each SNPs in turn, we applied the IVW method again to estimate the causal effect of the remaining SNPs. The aim of the method is to judge the reliability of these estimates by assessing the possible influence of some specific large effect level genetic variants (SNPs) on the results of MR causal inference. However, the IVW method may lead to bias when there is multiplicity or imbalance in the data. Therefore, the MR-Egger method, the weighted median method, the simple multitude method, and the weighted multitude method were used as supplementary means of analysis.
The above analyses were performed in the R software environment (version 4.4.1; The R Foundation for Statistical Computing, Vienna, Austria) with the help of the “Two Sample MR” (version 0.6.5) R package for statistical analysis and data visualization. Given that the IVW method is more efficient than the other 4 Mendelian randomization methods in testing causal effects, the IVW method was chosen to test causal effects in this study. Due to the relaxation of the threshold requirement, in order to avoid increasing the risk of 1 type of error and to verify whether the results of the study are affected by multiple testing.
3. Results
3.1. Determination of instrumental variables
SNPs loci with gene-wide significant (P < .05) correlation with AS were selected for pooling using R software, and SNPs were included as IV by excluding the control step of chain disequilibrium interference. As a single SNP for IV, the F statistic ranged from 19.58 to 79.93, indicating that there was a strong correlation between SNPs and exposure, and the results of MR analysis were unlikely to be biased by weak instrumental variables (see Table S1, Supplemental Digital Content, https://links.lww.com/MD/P902).
3.2. Causal relationship between AS and cancers
Given the greater efficiency of the IVW method in testing causal effects compared to the other 4 Mendelian randomization methods, the IVW method was chosen for this study to test for causal effects. Of the 15 malignancies studied, AS in cereals were associated with 6 malignancies; AS in coffee was associated with 3 malignancies; and AS in tea was associated with 3 malignancies. The results of IVW analysis in MR showed the following: cereal exposure breast cancer outcome IVW: odds ratio (OR) = 0.977, 95% confidence interval (CI) = 0.956–0.999, P = .040, showing that cereal intake decreases the risk of breast cancer; outcome gastric cancer IVW: OR = 0.321, 95% CI = 0.131–0.787, P = .013, showing that cereal intake decreases the risk of gastric cancer; outcome colorectal cancer IVW: OR = 0.980, 95% CI = 0.962–0.999, P = .036, showing that cereal intake decreases the risk of colorectal cancer; outcome prostate cancer IVW: OR = 0.972, 95% CI = 0.946–0.999, P = .039, showing that cereal intake decreases the risk of prostate cancer; outcome ovarian cancer IVW: OR = 1.015, 95% CI = 1.005–1.025, P = .004, showing that cereal intake increases the risk of ovarian cancer; outcome lung cancer IVW: OR = 2.567, 95% CI = 1.161–5.674, P = .020, showing that cereal intake increases the risk of lung cancer. Coffee exposure breast cancer outcome IVW: OR = 0.991, 95% CI = 0.983–0.999, P = .043, showing that coffee ingestion decreases the risk of breast cancer; outcome esophageal cancer IVW: OR = 4.236, 95% CI = 1.114–16.112, P = .034, showing that coffee ingestion increases the risk of esophageal cancer; outcome kidney cancer IVW: OR = 0.203, 95% CI = 0.052–0.784, P = .021, showing that coffee ingestion decreases the risk of kidney cancer. Tea exposure oral cancer outcome IVW: OR = 6.136, 95% CI = 1.406–26.775, P = .016, showing that ingestion of tea increases the risk of oral cancer; outcome esophageal cancer IVW: OR = 3.789, 95% CI = 1.035–13.875, P = .044, showing that ingestion of tea increases the risk of esophageal cancer; outcome gastric cancer IVW: OR = 0.478, 95% CI = 0.250–0.912, P = .025, showing that ingestion of tea increases the risk of gastric cancer. IVW showed statistical significance (P < .05, see Table 2 and Fig. 2). So there was a causal association between AS and breast, oral, esophageal, gastric, colorectal, prostate, ovarian, renal, and lung cancers, and the forest plot is shown in Figure 3.
Table 2.
Results of IVW analysis in MR of AS with significantly associated cancers.
| Exposure | Outcome | Sample size | nsnp | OR | 95% CI | P-value |
|---|---|---|---|---|---|---|
| AS intake in cereal | Breast cancer | 337,159 | 11 | 0.977 | 0.956–0.999 | .040 |
| Gastric cancer | 476,116 | 31 | 0.321 | 0.131–0.787 | .013 | |
| Colorectal cancer | 377,673 | 11 | 0.980 | 0.962–0.999 | .036 | |
| Prostate cancer | 292,053 | 30 | 0.972 | 0.946–0.999 | .039 | |
| Ovarian cancer | 199,741 | 29 | 1.015 | 1.005–1.025 | .004 | |
| Lung cancer | 492,803 | 31 | 2.567 | 1.161–5.674 | .020 | |
| AS intake in coffee | Breast cancer | 337,159 | 25 | 0.991 | 0.983–0.999 | .043 |
| Esophageal cancer | 160,589 | 11 | 4.236 | 1.114–16.112 | .034 | |
| Kidney cancer | 174,977 | 14 | 0.203 | 0.052–0.784 | .021 | |
| AS intake in tea | Oral cavity cancer | 3464 | 22 | 6.136 | 1.406–26.775 | .016 |
| Esophageal cancer | 160,589 | 12 | 3.789 | 1.035–13.875 | .044 | |
| Gastric cancer | 476,116 | 16 | 0.478 | 0.250–0.912 | .025 |
AS = artificial sweeteners, IVW = inverse variance weighting, MR = Mendelian randomization, OR = dominance ratio, sample size = sample size of the data, 95% CI = 95% confidence interval.
Figure 2.
Scatter plot of MR analysis results of artificial sweeteners and risk of 15 tumors. MR = Mendelian randomization.
Figure 3.
Expression of SNP forest effect in artificial sweeteners and risk of 15 tumors. SNP = single nucleotide polymorphism.
Consumption of cereals, coffee and tea showed no statistically significant (P > .05) IVW with bladder, endometrial, cervical, liver, pancreatic and thyroid cancers, so there is no causal association between AS and bladder, endometrial, cervical, liver, pancreatic and thyroid cancers.
3.3. Sensitivity analysis
3.3.1. Results of heterogeneity test and multiplicity test
Cochrane test indicated that significant heterogeneity between instrumental variables existed if the P-value was <.05. The heterogeneity result of IVW was P > .05, and that of MR-Egger regression was P > .05, suggesting that there was no heterogeneity among SNPs. See Table 3. MR-Egger regression analysis Intercept P > .05, thus indicating that there is a small risk of potential confounding bias in the analysis results.
Table 3.
Two-sample Mendelian randomization analysis SNP tool variable information table.
| Exposure | Outcome | Heterogeneity test | Multiplicity test | ||
|---|---|---|---|---|---|
| MR-Egger Q P-value | IVW Q P-value | MR-Egger intercept value | MR-Egger intercept P-value | ||
| AS intake in cereal | Breast cancer | .472 | .530 | ‐2.22E‐04 | .550 |
| Gastric cancer | .756 | .797 | 5.47E‐05 | .997 | |
| Colorectal cancer | .222 | .295 | 7.25E‐06 | .983 | |
| Prostate cancer | .781 | .816 | 1.28E‐04 | .751 | |
| Ovarian cancer | .550 | .596 | ‐5.34E‐05 | .720 | |
| Lung cancer | .307 | .320 | ‐9.80E‐03 | .418 | |
| AS intake in coffee | Breast cancer | .787 | .822 | 1.01E‐04 | .696 |
| Esophageal cancer | .620 | .674 | ‐3.19E‐02 | .557 | |
| Kidney cancer | .862 | .833 | 4.48E‐02 | .290 | |
| AS intake in tea | Oral cavity cancer | .858 | .893 | ‐2.79E‐03 | .956 |
| Esophageal cancer | .865 | .862 | ‐4.82E‐02 | .395 | |
| Gastric cancer | .675 | .669 | ‐2.26E‐02 | .336 | |
AS = artificial sweeteners, IVW = inverse variance weighting, MR = Mendelian randomization, SNP = single nucleotide polymorphism, Q = Cochran Q value.
The results of the leave-one-out sensitivity analysis showed that after eliminating 1 SNP locus in turn, the remaining SNP loci combined yielded OR values close to those of the IVW method, and no single SNP was found to significantly affect the results (see Fig. 4). This suggests that the effect ORs obtained by the IVW method are stable and that the results of the MR analysis are not dominated by a single SNP. Steiger directionality test showed that breast, esophageal, gastric, colorectal, prostate, ovarian, renal, and lung cancers were not inverse (all P < .05), reinforcing the exclusivity assumption of the 3 Mendelian assumptions.
Figure 4.
Artificial sweeteners and risk of 15 tumors by leave-one-out method.
4. Discussion
We conducted a 2-sample MR analysis using large-scale GWAS summary statistics on 64,949 individuals consuming AS in cereals, coffee, and tea and patients with 15 cancers with the aim of exploring causal associations between AS and the risk of multiple cancers. The causal relationship was further confirmed by the IVW method, MR-Egger regression analysis, weighted median method, simple multitude method and weighted multitude method. The results of the leave-one-out sensitivity analysis showed that the findings were not due to the abnormal effects of a single SNP, which enhanced the reliability of the conclusions. Based on this study, we found causal associations between AS and breast, oral cavity, esophageal, gastric, colorectal, prostate, ovarian, renal, and lung cancers, while no associations were found with bladder, endometrial, cervical, liver, pancreatic, and thyroid cancers. In this study, a comprehensive MR analysis was performed to examine in more depth the causal role of AS intake from cereals, coffee, and tea in relation to 15 cancers.
The results of the MR study confirm that the intake of AS from cereals, coffee, and tea can have an impact on human health either as a risk factor or as a protective factor. This study confirmed that elevated AS intake is a protective factor for breast, gastric, colorectal, prostate, and kidney cancers, and a risk factor for oral, esophageal, ovarian, and lung cancers. Chenglou Zhu et al meta-analysis reveals that low AS intake may be associated with reduced risk of colorectal cancer.[20] In diabetic subjects, high intake of other AS was associated with colorectal and gastric cancer, and high consumption of aspartame was associated with gastric cancer, while the risk of breast cancer was lower.[21] These points are consistent with the results of this experimental study and corroborate each other. Liping Liu team found a negative association between the use of artificial sweeteners and the risk of urological cancer in a case-control study.[22] ASs added to coffee were positively associated with high-grade and low-grade plasma ovarian cancer; ASs added to tea were positively associated with oral cavity and pharyngeal cancers but negatively associated with bronchial and lung malignancies.[23] Xia Ye and her team’s study finds no link between AS and breast cancer risk,[24] but there are also studies that show AS does not significantly increase the risk of breast cancer.[25] Sharon P Fowler et al, in their study of the relationship between consumption of artificially sweetened beverages and long-term weight gain, gave patients different artificial sweeteners such as saccharin, aspartame, acetylsulfonate-K and sucralose in order to control their diabetes. At the end of the experiment, it was found that the use of AS may contribute to the epidemic of obesity, which is a high risk factor for breast cancer.[26,27] Acetylsulfonylmethane (acetylsulfonate-K) and sucralose can also lead to a variety of health problems, such as the production of the toxic substance acetoacetamide, eye irritation in animals, thymus atrophy (up to 40%), reduced red blood cell counts, and liver and kidney disease.[28] Morando Soffritti, in a study of 2270 rats and 852 mice, found that aspartame was a carcinogenic agent in experimental animals, inducing a significant increase in the incidence of several malignant tumors, including hematological neoplasms and hepatocellular carcinoma, among others.[29]
Artificial sweeteners are used as sugar substitutes; they are many times sweeter than natural sugar and contain no calories, but extensive scientific research has shown that consuming large amounts of artificial sweeteners can increase the likelihood of several diseases. Mélanie Deschasaux-Tanguy et al found that there is uncertainty about the effects of long-term AS intake on the human body, but that it can lead to an increased risk of several chronic diseases and even cancers, such as cancer of the oral cavity, esophagus, ovary, lung, cardiovascular disease, diabetes mellitus,[30] and endometrial cancer.[31] A meta-analysis in 2021 showed no correlation between artificial sweeteners and the development of cancer, except for female urologic tumors. Therefore, additional data from large clinical trials are needed to validate this study to determine whether artificial sweeteners can increase or decrease the prevalence of tumors in humans.
This 2-sample MR randomization study is of extreme clinical importance. First, as the first MR study to use AS as an exposure to estimate its causal effect on malignancy risk. A major strength of this study is its 2-sample MR design, which prevents the influence of potential confounders. Second, the study’s sample was obtained from the GWAS database,[32] making the study a valid causal inference with high statistical power. Third, the results of the study were extremely robust through strict quality control conditions and a series of sensitivity analyses. After removing SNPs associated with potential confounders, the MR study demonstrated a causal relationship between genetically predicted AS and risk of cancers, with independent associations and no inverse correlation. The application of rigorous tools (e.g., F-statistics significantly >10) during this investigation helped to mitigate potential bias caused by sample overlap.
However, there are some limitations to this study. First, the SNPs in this study were mainly from European populations, so our causal estimates may not be fully generalizable to other populations. Second, because the article was an “exploratory study,” multiple correction tests were not performed, which would be needed to increase rigor. For this reason, we did not conduct the study again. Third, due to the lack of information, outcomes were not stratified according to the different clinical types of malignant tumors, which is important in clinical practice. Subsequent further studies on different clinically staged malignancies are needed to fully explore the association between AS and cancer, the results of this study should be considered as “exploratory evidence,” and its clinical application should be supported by future confirmatory studies.
In conclusion, AS intake was associated with a decreased risk of breast, gastric, colorectal, prostate, and kidney cancers; AS intake was associated with an increased risk of oral, esophageal, ovarian, and lung cancers.
Author contributions
Conceptualization: Lichao Zhu.
Data curation: Haiyuan Yang, Ye Liu.
Formal analysis: Ye Liu, Xuanming Mao.
Investigation: Haiyuan Yang, Jie Peng, Lichao Zhu.
Methodology: Ye Liu.
Software: Ye Liu, Yiheng Zhang.
Supervision: Guoxin Sun.
Validation: Shuai Chen.
Visualization: Haiyuan Yang.
Writing – original draft: Haiyuan Yang, Yating Zhao.
Writing – review & editing: Haiyuan Yang, Yating Zhao.
Supplementary Material
Abbreviations:
- AS
- artificial sweeteners
- CI
- confidence interval
- GWAS
- genome-wide association study
- IV
- instrumental variable
- IVW
- inverse variance weighting
- MR
- Mendelian randomization
- OR
- odds ratio
- SNP
- single nucleotide polymorphism.
There was no research on humans or animals in this study, and no relevant ethical approval was needed.
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Supplemental Digital Content is available for this article.
How to cite this article: Yang H, Liu Y, Peng J, Mao X, Zhang Y, Chen S, Sun G, Zhu L, Zhao Y. Causal association between artificial sweeteners and risk of 15 tumors: A 2-way 2-sample Mendelian randomization analysis. Medicine 2025;104:48(e44379).
HY, YL, and JP contributed to this article equally.
Contributor Information
Haiyuan Yang, Email: 1154001413@qq.com.
Ye Liu, Email: LiuYe0524@outlook.com.
Jie Peng, Email: Pengjie30@outlook.com.
Xuanming Mao, Email: 1303396760@qq.com.
Yiheng Zhang, Email: 244906573@qq.com.
Shuai Chen, Email: 1103866127@qq.com.
Guoxin Sun, Email: 2414537919@qq.com.
Lichao Zhu, Email: 1424763291@qq.com.
References
- [1].Mooradian AD. Sugar or sweetener? South Med J. 2024;117:513–6. [DOI] [PubMed] [Google Scholar]
- [2].Zhang Y, Tang Z, Shi Y, Li L. Associations between artificial sweetener intake from cereals, coffee, and tea and the risk of type 2 diabetes mellitus: a genetic correlation, mediation, and mendelian randomization analysis. PLoS One. 2024;19:e0287496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Bruyère O, Ahmed SH, Atlan C, et al. Erratum to: review of the nutritional benefits and risks related to intense sweeteners. Arch Public Health. 2015;73:49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Gallus S, Scotti L, Negri E, et al. Artificial sweeteners and cancer risk in a network of case-control studies. Ann Oncol. 2007;18:40–4. [DOI] [PubMed] [Google Scholar]
- [5].Mallikarjun S, Sieburth RM. Aspartame and risk of cancer: a meta-analytic review. Arch Environ Occup Health. 2015;70:133–41. [DOI] [PubMed] [Google Scholar]
- [6].Haalck I, Székely A, Ramne S, et al. Are we using more sugar substitutes? Wastewater analysis reveals differences and rising trends in artificial sweetener usage in Swedish urban catchments. Environ Int. 2024;190:108814. [DOI] [PubMed] [Google Scholar]
- [7].Li M, Liu Q, Shi M, Fu M, He G. Association between remnant cholesterol and the risk of 4 site-specific cancers: evidence from a cross-sectional and Mendelian randomization study. Lipids Health Dis. 2024;23:256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Sandhu APS, Tanvir, Singh K, et al. Decoding cancer risk: understanding gene–environment interactions in cancer development. Cureus. 2024;16:e64936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey Smith G. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med. 2008;27:1133–63. [DOI] [PubMed] [Google Scholar]
- [10].Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23:R89–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Emdin CA, Khera AV, Kathiresan S. Mendelian randomization. JAMA. 2017;318:1925–6. [DOI] [PubMed] [Google Scholar]
- [12].Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ. 2018;362:k601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Pierce BL, Burgess S. Efficient design for Mendelian randomization studies: subsample and 2-sample instrumental variable estimators. Am J Epidemiol. 2013;178:1177–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Sudlow C, Gallacher J, Allen N, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12:e1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Davey Smith G, Holmes MV, Davies NM, Ebrahim S. Mendel’s laws, Mendelian randomization and causal inference in observational data: substantive and nomenclatural issues. Eur J Epidemiol. 2020;35:99–111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Ji L, Liu Y, Wang Z, et al. Causal effect analysis of estrogen receptor associated breast cancer and clear cell ovarian cancer. Am J Transl Res. 2024;16:2699–710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Hartwig FP, Davies NM, Hemani G, Davey Smith G. Two-sample Mendelian randomization: avoiding the downsides of a powerful, widely applicable but potentially fallible technique. Int J Epidemiol. 2016;45:1717–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Chen X, Kong J, Diao X, et al. Depression and prostate cancer risk: a Mendelian randomization study. Cancer Med. 2020;9:9160–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Wang X, Wang X, Wang H, Yang M, Dong W, Shao D. Association between psoriasis and lung cancer: two-sample Mendelian randomization analyses. BMC Pulm Med. 2023;23:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Zhu C, Ji D, Ma J, Da M. Association between artificial sweetener-aspartame consumption and colorectal cancer risk: evidence-based strategies. Oncology (Huntingt). 2024;102:533–43. [DOI] [PubMed] [Google Scholar]
- [21].Palomar-Cros A, Straif K, Romaguera D, et al. Consumption of aspartame and other artificial sweeteners and risk of cancer in the Spanish multicase-control study (MCC-Spain). Int J Cancer. 2023;153:979–93. [DOI] [PubMed] [Google Scholar]
- [22].Liu L, Zhang P, Wang Y, Cui W, Li D. The relationship between the use of artificial sweeteners and cancer: a meta-analysis of case-control studies. Food Sci Nutr. 2021;9:4589–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Pan H, Feng C, Zhou Z, et al. The causal association between artificial sweeteners and the risk of cancer: a Mendelian randomization study. Food Funct. 2024;15:4527–37. [DOI] [PubMed] [Google Scholar]
- [24].Ye X, Zhang Y, He Y, Sheng M, Huang J, Lou W. Association between consumption of artificial sweeteners and breast cancer risk: a systematic review and meta-analysis of observational studies. Nutr Cancer. 2023;75:795–804. [DOI] [PubMed] [Google Scholar]
- [25].Romanos-Nanclares A, Collins LC, Hu FB, et al. Sugar-sweetened beverages, artificially sweetened beverages, and breast cancer risk: results from 2 prospective US cohorts. J Nutr. 2021;151:2768–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Fowler SP, Williams K, Resendez RG, Hunt KJ, Hazuda HP, Stern MP. Fueling the obesity epidemic? Artificially sweetened beverage use and long-term weight gain. Obesity (Silver Spring). 2008;16:1894–900. [DOI] [PubMed] [Google Scholar]
- [27].Milani I, Gaita C, Guarisco G, et al. The intricate relationship between obesity, type 2 diabetes and female breast cancer: a retrospective study of 335 women. Obes Sci Pract. 2024;10:e786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Belton K, Schaefer E, Guiney PD. A review of the environmental fate and effects of acesulfame-potassium. Integr Environ Assess Manag. 2020;16:421–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Soffritti M. Understanding the link between aspartame and cancer. Expert Rev Anticancer Ther. 2024;24:793–802. [DOI] [PubMed] [Google Scholar]
- [30].Deschasaux-Tanguy M, Srour B, Touvier M. Sweeteners: recent recommendations for health. Rev Prat. 2024;74:673–6. [PubMed] [Google Scholar]
- [31].Li H, Zhang Y, He Y, Huang J, Yao J, Zhuang X. Association between consumption of sweeteners and endometrial cancer risk: a systematic review and meta-analysis of observational studies. Br J Nutr. 2024;131:63–72. [DOI] [PubMed] [Google Scholar]
- [32].Ishigaki K, Akiyama M, Kanai M, et al. Large-scale genome-wide association study in a Japanese population identifies novel susceptibility loci across different diseases. Nat Genet. 2020;52:669–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.