Association between mercury levels and colorectal cancer: a systematic review, meta-analysis

Abstract

Background

This study examines how mercury (Hg) levels may be linked to colorectal cancer (CRC) by reviewing and analyzing existing research. It is important to understand this connection because mercury may cause cancer.

Methods

We searched PubMed, Scopus, and Web of Science for studies published up to January 2026. We included observational studies evaluating mercury levels in patients with CRC compared with healthy controls or non-cancerous colorectal tissue, as well as studies assessing the association between mercury exposure and colorectal cancer. We analyzed the data using a random-effects model. Mercury levels were measured in subgroups within both the case and control groups to address differences in measurement methods. We assessed the quality of the included studies using the Newcastle-Ottawa Scale. All data were analyzed with STATA version 17.

Results

This meta-analysis reviewed 8 studies selected from 1,099 database records. Four studies with 773 participants (355 cases and 418 controls) found no significant difference in mercury levels between people with colorectal cancer and healthy controls (SMD: 0.16 [-0.22, 0.54]; p = 0.42; I²: 82.25%). Subgroup analysis showed differences depending on study design, average age, percentage of male participants, and risk of bias score. Meta-regression found that participant age affected the results, while other factors did not. In two studies with 71 participants, there was also no significant difference in mercury levels between colorectal and healthy tissues in patients with colorectal cancer (SMD: -0.24 [-1.11, 0.64]; p = 0.60; I²: 76.68%), despite high variability across studies.

Conclusion

Our review found no significant link between mercury levels and CRC risk, underscoring the complexity of environmental toxins in cancer etiology and the need for further research on influencing factors.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12876-026-05099-4.

Introduction

Toxic metals can build up in the body over time, leading to long-term exposure and serious health risks. Recent studies have found a link between this exposure and cancer development [1–3]. Mercury (Hg) is a harmful metal, and the World Health Organization (WHO) lists it among the top 10 contaminants threatening human health. Mercury is also ranked third on the Substance Priority List by the Agency for Toxic Substances and Disease Registry [4]. The health effects of mercury depend on its chemical form, which influences how easily the body absorbs and processes it.

There are multiple forms of mercury with various mechanisms of exposure, modes of absorption, and biological effects. Mercury in its elemental form is primarily inhaled, enters the body via dental fillings, and can lead to work-related exposure, while inorganic mercury may bioaccumulate within tissues and bind sulfhydryl groups in proteins and antioxidants [4–7]. In contrast, organic mercury, such as methylmercury, enters the body via fish or seafood consumption and is readily absorbed in the digestive system [5]. There are differences in toxicokinetics, tissue distribution, oxidative stress, and DNA damage associated with various mercury types, which might be biologically relevant in cancer development [6–8]. Although no studies have confirmed a direct link between specific forms of mercury and colorectal cancer, it is crucial to distinguish between forms of mercury while interpreting results obtained from diverse samples.

People are mainly exposed to elemental mercury, a liquid metal that evaporates easily, through dental fillings and small-scale gold mining [5]. Concerns about mercury exposure have grown because it is linked to heart disease and stomach cancers [9, 10]. Mercury is also toxic to the reproductive and nervous systems [11]. The International Agency for Research on Cancer [12] has labeled mercury as a possible human carcinogen. While there is a link between methylmercury exposure and leukemia, there is no strong connection to rectal cancer deaths [13]. Ingestion of methylmercury is one of the important ways through which exposure to this substance occurs, especially via consumption of fish and other contaminated foods. Nonetheless, the role of dietary mercury in development of CRC is yet to be conclusively established. According to results from a Korean study involving 2,769 patients, increased dietary exposure to mercury is a risk factor for developing colorectal cancer, especially rectal cancer [14]. To reduce the health risks from mercury pollution, strong measures are needed. For example, a Korean case-control study with 2,769 participants found that higher mercury intake from food was linked to a greater risk of colorectal cancer [15].

CRC accounted for roughly 9.6% of all new cases of cancer in 2022, after lung and breast cancers. According to projections, by 2040, colorectal cancer might reach an incidence of 3.2 million new cases each year, while mortality might reach 1.6 million [16]. The rate of incidence has been declining over time in western countries because of the use of screening techniques such as colonoscopy. However, its incidence in young individuals is on the increase [17].

This article presents a systematic review and meta-analysis to elucidate the potential link between mercury exposure and CRC. We hope to contribute valuable insights into the complex interplay between environmental toxins and cancer development, ultimately advancing our knowledge of colorectal cancer etiology.

Methods

This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 2020 [18].

Search strategy

The search was done by using keywords such as “colorectal cancer” and “mercury.” We also used the corresponding synonyms and MeSH terms. The databases searched were PubMed, Scopus, and Web of Science. The time period involved searching for articles in the English language published up until January 2026. The detailed, reproducible search strategies are presented in SM 1. Additionally, Google Scholar was searched with the keywords “(Hg OR mercury) AND colorectal,” with the first ten pages reviewed. Furthermore, reference sections of selected articles as well as systematic reviews were examined.

Eligibility criteria

We conducted a systematic review of observational studies examining the relationship between mercury levels and colorectal cancer. Our inclusion criteria followed the PECO framework, focusing on colorectal cancer patients who had mercury measured in blood, tissue, urine, or other biological samples. We compared these patients to healthy individuals, normal tissue from colorectal cancer patients, or included studies without a control group. Studies without an external healthy control group were considered eligible only for qualitative synthesis when they provided relevant exposure or tissue-based information; however, they were not included in the quantitative meta-analysis unless comparative data were available. The main outcome we looked at was mercury levels. As a secondary outcome, we included any data showing an association between mercury and colorectal cancer. To ensure we included only relevant, high-quality studies, we set several exclusion criteria. We excluded studies that did not provide sufficient data to compare mercury levels between groups. We also excluded duplicate studies, studies with overlapping participant data, and any that were reviews, editorials, conference papers, case series or reports, secondary analyses, or animal experiments. Studies using qualitative research methods were also excluded. Case reports and case series were excluded because they do not provide sufficient comparative or population-level evidence for either meta-analysis or meaningful qualitative interpretation.

Study selection and data extraction

Each article title and abstract was independently screened by two independent authors, and subsequently the full texts of potentially eligible articles were independently evaluated by both reviewers. In case of any dispute, consensus was finally obtained through discussion. The extraction of data was done by two independent investigators, with the data from each investigator cross-checked for any discrepancy. The extracted data included author name, country, methodology used, sample size, mean age, male/female proportion, method used to measure mercury, source of biological samples, units of mercury measure, mean mercury values, study characteristics of diseases, and risk of bias assessment. General characteristics of all selected articles were provided in Table 1. Quantitative extraction data with information on the type of comparison, sample source, units of mercury measure, mean/standard deviation, sample size, and inclusion in the pooled meta-analysis were presented in Table 2.

Table 1.

General characteristics of all included studies

Author	Country	Study design	Study population / comparison	Sample size	Mean age	Male/Female	Mercury exposure or sample source	Main finding	NOS score / Risk of bias
Arriola et al. 1999* [20]	Mexico	Cross-sectional / tissue-comparison	CRC tissue compared with adjacent normal tissue	12	30–86	NR	Colorectal tissue	No significant difference in mercury levels between normal and cancerous tissue	5 / Moderate
Huynh et al. 2021* [21]	Vietnam	Case-control	CRC patients compared with healthy controls	102 cases; 60 controls	Cases: 52.7 ± 14.3; controls: 54.2 ± 8.5	Cases: 52/50; controls: 30/30	Nails	No significant difference in mercury levels between CRC patients and controls	7 / Low
Juloski et al. 2020* [22]	Serbia	Cross-sectional / tissue-comparison	CRC tissue compared with adjacent normal tissue	59	67 ± 10	29/30	Colorectal tissue	Mercury levels were significantly higher in normal tissue than cancerous tissue	7 / Low
Kim et al. 2020~[15]	South Korea	Case-control	CRC patients compared with controls	923 cases; 1856 controls	Cases: 56.6 ± 8.7; controls: 56.1 ± 9.1	Cases: 625/298; controls: 1250/596	Dietary mercury intake	Higher dietary mercury intake was associated with increased CRC risk	6 / Moderate
Li et al. 2023* [23]	China	Case-control	CRC patients compared with healthy controls	101 cases; 60 controls	Cases: 60.77 ± 10.28; controls: 45.13 ± 16.76	Cases: 54/47; controls: 16/44	Blood	No significant difference in mercury levels between CRC patients and controls	7 / Low
Mahmood et al. 2022* [24]	Pakistan	Case-control	CRC patients compared with healthy controls	185 cases; 151 controls	Cases: 42.85 ± 13.02; controls: 41.93 ± 11.01	Cases: 86/79; controls: 70/81	Blood	No significant difference in mercury levels between CRC patients and controls	8 / Low
Nawi et al. 2020* [25]	Malaysia	Cohort	CRC patients compared with controls	102 cases; 102 controls	NR	Cases: 63/39; controls: 61/41	Blood	No significant difference in mercury levels between CRC patients and controls	5 / Moderate
Yorifuji et al. 2007# [13]	Japan	Cohort / ecological exposure-based analysis	High methylmercury-exposed areas compared with reference population	Minamata and Ashikita: 33,733; Amakusa: 56,575; reference population: 28,434,159	NA	48%, 48.1%, and 49.1% male	High methylmercury exposure	No strong evidence of increased rectal cancer mortality	7 / Low

Abbreviations: CRC Colorectal cancer, NOS Newcastle-Ottawa Scale, NR Not reported, NA Not applicable

Table 2.

Data extraction for mercury-related outcomes in the included studies

Analysis group	Author	Comparison / outcome	Sample source or exposure type	Unit reported in original study	Case / cancerous tissue, mean ± SD	Control / normal tissue, mean ± SD	N case / cancerous tissue	N control / normal tissue	Included in pooled meta-analysis	Notes
A. CRC patients vs. healthy/non-CRC controls	Huynh et al., 2021 [21]	Colon cancer patients vs. healthy controls	Nails	µg/g	0.47 ± 0.21	0.37 ± 0.15	22	60	Yes	Untreated colon cancer subgroup compared with healthy controls
A. CRC patients vs. healthy/non-CRC controls	Huynh et al., 2021 [21]	Rectal cancer patients vs. healthy controls	Nails	µg/g	0.46 ± 0.26	0.37 ± 0.15	20	60	Yes	Untreated rectal cancer subgroup compared with the same healthy control group; interpreted cautiously because the control group was shared
A. CRC patients vs. healthy/non-CRC controls	Li et al., 2023 [23]	CRC patients vs. healthy controls	Serum	µg/L	0.26 ± 0.95	0.48 ± 2.64	101	60	Yes	Case-control comparison
A. CRC patients vs. healthy/non-CRC controls	Mahmood et al., 2022 [24]	CRC patients vs. healthy controls	Serum	µg/g, wet weight	5.50 ± 2.81	6.37 ± 2.41	165	151	Yes	Case-control comparison
A. CRC patients vs. healthy/non-CRC controls	Nawi et al., 2020 [25]	CRC patients vs. non-CRC controls	Serum	µg/L	0.90 ± 0.74	0.59 ± 0.33	102	102	Yes	Case-control/cohort-based comparison
B. Cancerous colorectal tissue vs. adjacent normal colorectal tissue	Arriola et al., 1999 [20]	Cancerous colorectal tissue vs. adjacent normal tissue	Tissue	NR	0.018 ± 0.016	0.028 ± 0.045	12	12	Yes	NR
B. Cancerous colorectal tissue vs. adjacent normal colorectal tissue	Juloski et al., 2020 [22]	Cancerous colorectal tissue vs. adjacent normal tissue	Tissue	µg/kg	0.78 ± 0.70	2.73 ± 4.35	59	59	Yes	Paired tissue comparison; analyzed separately from case-control studies
Qualitative synthesis only	Kim et al., 2020 [15]	Association between dietary mercury intake and CRC risk	Dietary mercury intake	µg/day	15.0 ± 2.0	14.1 ± 2.2	923	1846	No	Not pooled because it evaluated dietary mercury intake and CRC risk rather than mercury concentration in biological samples or tissues

Two separate quantitative analyses were conducted. Analysis A compared mercury levels between CRC patients and healthy/non-CRC controls using systemic biological samples, including nails and serum. Analysis B compared mercury levels between cancerous colorectal tissue and adjacent normal colorectal tissue. Tissue-based comparisons were analyzed separately because tissue mercury concentrations may reflect local deposition and are not directly comparable with systemic exposure matrices. Because mercury concentrations were reported in different units and biological matrices across studies, standardized mean differences were used for pooling

Abbreviations: CRC Colorectal cancer, SD Standard deviation

Quality assessment

We used the Newcastle-Ottawa Scale (NOS) to evaluate the risk of bias in the included observational studies. The original NOS was applied for case-control and cohort studies, while the adapted NOS version for cross-sectional studies was used for cross-sectional/tissue-comparison studies. The scale assesses study quality across selection, comparability, and outcome/exposure domains, with a maximum score of 9. Studies were rated as low risk of bias (scores of 7 or higher), moderate risk of bias (scores of 5 to 6), or high risk of bias (scores of 4 or lower). Two independent investigators assessed the quality, and any disagreements were resolved through discussion with a third investigator.

Quantitative analysis

In our meta-analysis, we collected two main types of data: mercury levels in healthy people compared to those with colorectal conditions, and mercury levels in normal versus colorectal tissue from cancer patients. Tissue-based comparisons were analyzed separately from case-control comparisons because tissue mercury levels may reflect local deposition rather than systemic exposure. We used standardized mean differences (SMDs) with 95% confidence intervals (CIs) to compare mercury levels between case and control groups. For the case-control analysis, different biological matrices, including nails, blood, and serum, were pooled because the number of eligible studies within each matrix was limited. However, we acknowledge that these matrices reflect different exposure windows; therefore, subgroup analyses by sample source were considered exploratory, and the overall pooled estimate was interpreted cautiously rather than as a definitive matrix-independent conclusion. When only the median, range, or interquartile range (IQR) was available, we calculated the mean and standard deviation (SD) using the method described by Wan et al. [19]. We checked for differences between studies using the Cochrane Q-test and the I² index, considering a P value below 0.05 as significant. If heterogeneity was present, we used a random-effects model for the analysis. Publication bias and small-study effects were assessed descriptively using funnel plots and Egger’s test; however, because the number of included effect sizes was limited, these assessments were interpreted cautiously and were not used to draw definitive conclusions about publication bias.

Subgroup analyses based on study design, proportion of males, mean age, biological sample type, and risk-of-bias assessment category were conducted. Univariable meta-regression was also conducted using mean age, proportion of males, smoking rate, and body mass index to explore the extent to which any of these factors could contribute to the heterogeneity between studies. Subgroup analysis and meta-regression covariates were chosen based on their potential clinical importance and the availability of such data for all included studies. Since the review protocol was not registered prospectively, all analyses conducted herein should be regarded as exploratory. The use of multivariable meta-regression was not considered due to the small number of studies and effect sizes available, and no such analysis was conducted. Meta-regression was considered an exploratory analysis since there was a small number of studies and effect sizes available for inclusion. All statistical analyses were carried out using Stata version 17.0, and p < 0.05 was deemed statistically significant unless specified otherwise.

Results

Selection of studies

From the database search, 1,099 records were identified: 310 from PubMed, 440 from Scopus, and 349 from Web of Science. After removing 465 duplicates, 634 records were screened, and 601 were excluded based on title and abstract screening. Of the remaining 33 full-text articles, 25 were excluded after eligibility assessment, resulting in eight studies [13, 15, 20–25] included in the qualitative synthesis. Among these, six studies [20–25] were appropriate for meta-analysis. The PRISMA flow diagram and supplementary search details were cross-checked to ensure that the same record counts were reported consistently across the manuscript (Fig. 1).

Fig. 1.

The PRISMA flow diagram of search results

Study characteristics

Eight observational studies were analyzed, conducted in countries including Mexico, Vietnam, Serbia, South Korea, China, Pakistan, Malaysia, and Japan between 1999 and 2023. Five studies used a case-control design, two used a cross-sectional design, and one used a cohort design. Two of these studies compared mercury levels in both normal and cancerous tissues from patients with cancer. The other six studies included control groups for comparison. Sample sizes ranged from 12 to 90,308. Details comes in Tables 1 and 2.

Risk of bias within studies

We employed the NOS scale to assess the risk of bias in all eight studies included, and five of those studies demonstrated a low risk and three of them moderate score (Table 1).

Synthesis of results

All in all, six studies were selected for the quantitative meta-analysis. Specifically, two separate outcomes were evaluated: mercury concentrations among CRC patients relative to the controls without cancer, and the level of mercury within the malignant tissue relative to its adjacent normal tissue among patients with CRC. Concerning the former outcome, there were four studies providing five effect sizes. If a single study provided several effect sizes, the common reference group was taken into account not to include them twice. In regard to the latter outcome, there were two studies with two effect sizes.

Colorectal vs. healthy

In this analysis, four [21, 23–25] studies comprising five effect sizes were included, with a total of 773 participants, 355 cases and 418 controls. Huynh et al. contributed two anatomical-site-specific effect sizes for colon and rectal cancer using a shared control group. Therefore, this analysis should be interpreted with caution because these two effect sizes are not fully independent and may increase the apparent precision of the pooled estimate. The results indicated no significant difference in mercury levels between colorectal cancer patients and healthy controls (SMD: 0.16 [-0.22, 0.54]; p = 0.42), and there was a high level of heterogeneity observed (I²: 82.25%) (see Fig. 2). Sensitivity analysis demonstrated that excluding any of the studies did not alter the pooled effect size (Fig. 3). As illustrated in Fig. 4, the funnel plot was retained as a descriptive assessment of small-study effects. Egger’s test (p = 0.25) and Begg’s test (p = 0.81) were also performed (Fig. 1); however, because this analysis included only five effect sizes, these tests were considered exploratory and were not used to make a definitive conclusion regarding the presence or absence of publication bias.

Fig. 2.

Meta-analysis comparing mercury levels in colorectal cancer patients and a control group

Fig. 3.

Sensitivity analysis comparing mercury levels in colorectal cancer patients and a control group

Fig. 4.

Funnel plot of studies comparing Hg levels in colorectal cancer patients and a control group

The subgroup meta-analysis was conducted as an exploratory analysis to describe possible sources of heterogeneity (Fig. 5). Subgroups were examined according to study design, proportion of male participants, mean age, sample source, and risk-of-bias category. Because several subgroups included only one to three effect sizes, these findings should be interpreted descriptively and should not be considered confirmatory evidence of effect modification. The pooled estimates appeared to vary across some subgroup categories; however, no firm statistical inference was made for subgroups represented by only a single study or effect size. Therefore, these subgroup findings were used only to describe patterns across the limited available evidence and to guide interpretation of heterogeneity. Descriptively, case-control studies showed a smaller pooled estimate (SMD: 0.02 [-0.36, 0.41]) than the cohort subgroup (SMD: 0.54 [0.26, 0.82]); however, the cohort subgroup included only one study, and this comparison should not be interpreted as evidence of a study-design effect. In the subgroup analysis based on male proportion, studies with male participants > 50% showed a pooled estimate of 0.52 [0.28, 0.76], while studies with male participants ≤ 50% showed a pooled estimate of -0.25 [-0.47, -0.03]. The analysis based on mean age showed that studies with participants aged ≤ 45 years had a pooled estimate of -0.33 [-0.55, -0.12], whereas studies with participants aged > 45 years had a pooled estimate of 0.31 [-0.05, 0.67]. Studies using nail samples showed a pooled estimate of 0.46 [-0.00, 0.92], while studies using serum samples showed a pooled estimate of 0.03 [-0.48, 0.55]. Finally, studies with low risk of bias showed a pooled estimate of 0.31 [-0.05, 0.67], while studies with moderate risk of bias showed a pooled estimate of -0.33 [-0.55, -0.12]. Given the limited number of studies and effect sizes in these subgroups, all subgroup findings should be interpreted as exploratory and hypothesis-generating only, rather than confirmatory.

Fig. 5.

Subgroup meta-analysis comparing mercury levels in colorectal cancer patients and a control group

Moreover, exploratory univariate meta-regression analyses were conducted to determine whether some factors related to each study could account for the heterogeneity found in the data. This was done by using the few effect sizes that were available. Thus, any conclusions from this analysis should be drawn with caution since the data are not stable. For mean age, five effect sizes, taken from four studies, yielded a coefficient of 0.042 [0.009, 0.076] (p = 0.012). Male percentage was assessed using five effect sizes from four studies, and smoking percentage and BMI were assessed only among studies with available data. The variables of male percentage, smoking percentage, and BMI were not statistically significant (Table 3). Given the small number of included effect sizes, these findings were retained as exploratory analyses only and were not used to draw firm conclusions regarding effect modification.

Table 3.

Meta regression of the studies based on variable

Variables	Coefficient	Standard error	95% Confidence interval	P-value
Mean age	0.042	0.017	0.009, 0.076	0.012
Male (%)	0.063	0.047	-0.030, 0.157	0.183
Smoking (%)	-0.031	0.050	-0.129, 0.067	0.536
BMI	0.161	0.432	-0.687, 1.009	0.702

Normal vs. cancerous tissue

For the analysis, two [20, 22] studies with two effect size included in analysis with 71 participants. Analysis showed no differences of mercury level between colorectal and healthy tissue in patients with colorectal cancer (SMD: -0.24 [-1.11, 0.64]; p = 0.60), which showed high heterogeneity (I2: 76.68%) (Fig. 6).

Fig. 6.

Meta-analysis comparing mercury levels in normal and cancer tissue of colorectal cancer patients. Positive SMD values indicate higher mercury levels in cancerous tissue, whereas negative values indicate higher mercury levels in adjacent normal tissue

Discussion

This systematic review and meta-analysis examined six studies to examine how mercury levels relate to CRC. We focused on two main questions: whether mercury levels differ between CRC patients and healthy people, and whether mercury levels vary between normal and cancerous colorectal tissues. Four separate studies [21, 23–25] involving 773 individuals (355 patients with CRC and 418 control subjects) yielded no significant difference in mercury levels between the groups under investigation. The result must be interpreted with caution since the heterogeneity is very high (I² = 82.25%). It means that there is considerable variation across studies due to different aspects like design, the source of biological samples, participants’ characteristics, location of study, mercury measurements methods, etc. Despite the sensitivity analysis indicated that the pooled estimates were stable enough even without some of the included studies, it should be understood as a descriptive stability rather than as a reliable and robust estimate. That is, one must not conclude that the pooled estimates suggest the absence of any difference in mercury levels between the studied groups for any other populations and sample matrices.

When we looked at subgroups, which were conducted as exploratory analyses, we found that case-control studies [21, 23, 24] showed smaller differences in mercury levels than cohort studies [25]. However, because some subgroup categories included only one study or a very small number of effect sizes, these findings should be interpreted descriptively rather than as confirmatory evidence that study design affects mercury levels. Similarly, the age-related findings from subgroup analysis and meta-regression should be considered hypothesis-generating only and should not be interpreted as evidence that age directly influences mercury accumulation.

The analysis found no significant difference in mercury levels between studies using nail samples [21] and those using serum [23–25], though mercury levels tended to be higher in nail samples. This finding raises questions about which biological samples are best for measuring mercury exposure and its link to cancer risk. However, subgroup analyses by sample source were limited by the small number of studies and should not be used to draw firm conclusions regarding the comparative validity of different biological matrices. The ROB analysis showed that studies with low ROB [21, 23, 24] had greater differences in mercury levels than those with moderate ROB [25], although this pattern should also be interpreted cautiously because each risk-of-bias subgroup contained few effect sizes. Meta-regression showed that participant age was a significant factor affecting mercury levels, while male percentage, smoking percentage, and BMI were not, but given the small number of included studies, these findings should be viewed as exploratory and not as definitive evidence of effect modification.

We studied normal and cancerous colorectal tissues from two studies [20, 22] with 71 participants and did not find a significant difference in mercury levels. This tissue-based result should also be interpreted with caution because heterogeneity was high (I²=76.68%) and only two studies were available. Differences in tissue sampling site, adjacent normal tissue definition, analytical methods, tumor characteristics, and local metal deposition may have contributed to the variability between studies. ecause the included tissue studies were based on cross-sectional concentration measurements, accumulation rates cannot be inferred. Therefore, the tissue comparison should be considered exploratory, and no firm conclusion can be made regarding mercury accumulation in cancerous versus adjacent non-cancerous colorectal tissue.

In the primary studies included in this research that evaluated mercury levels for comparison between cancerous and healthy tissues, only Juloski [22] found a significant difference, reporting that normal tissue had higher mercury levels than cancer tissue. Other studies did not report any differences.

Two studies not included in the meta-analysis also explore the link between mercury and colorectal cancer. Yorifuji [13] found significant associations between Me mercury exposure and some cancers, like leukemia, but did not find strong evidence of a major link between Me mercury exposure and deaths from rectal cancer. These results show that more individual-level research is needed to better understand these connections, since ecological studies have limits in proving direct causes. Kim’s [15] study found a clear link between eating mercury and a higher risk of rectal cancer, especially in people who eat less fish and shellfish. This means that what people eat can affect how mercury from seafood influences cancer risk.

The findings of some of these studies were notable. For instance, Juloski et al. [22] demonstrated significant differences in trace element concentrations between malignant and adjacent healthy tissues. They found that Zinc (Zn) and Sodium (Na) were significantly lower in malignant tissue, and Copper (Cu), Cadmium (Cd), Magnesium (Mg), Selenium (Se), Calcium (Ca), and Potassium (K) were significantly higher in malignant tissue. In addition, the Cu/Zn ratio was significantly higher in CRC than in healthy tissue, particularly in patients with advanced CRC (III and IV). However, this ratio did not show significant differences between sexes or age groups. Also, no significant correlation was found between elemental concentrations and the patients’ age. Furthermore, in the context of the correlation between toxic metals and CRC, Mahmood et al. [24] evaluated the concentrations of specific essential and toxic elements in the serum of 165 newly diagnosed CRC patients compared to 151 healthy controls. The correlation between elemental levels in cancer patients differed significantly from that in healthy subjects. Notably, variations in elemental concentrations were observed by CRC type—such as primary colorectal lymphoma, gastrointestinal stromal tumor, and adenocarcinoma—as well as across disease stages (I-IV).

Another study, which didn’t enter the meta-analysis, showed that high dietary mercury intake was associated with an increased risk of CRC [15]. In another study in which the role of heavy metals in development of CRC has been studied, the showed that the level of Se was lower in the CRC group compared with the control group, while vanadium (V), arsenic (As), tin (Sn), barium (Ba) and lead (Pb) were higher, chromium (Cr) and copper (Cu) were significantly higher in the CRC group than those in the control group [23]. In a study, the potential mutations that can play a role in developing CRC in patients with a history of heavy metal exposure has investigated. The study examined the links between heavy metal exposure and the types and frequencies of gene mutations in patients with lung adenocarcinoma (LC) and CRC. The researchers found a correlation between the combined concentrations of heavy metals and the number of gene mutations, particularly insertions/deletions (indels). Lead, arsenic, and cadmium were identified as the most significant contributors to increased mutation rates. The study also revealed that established mutational signatures showed significant correlations with metal exposure. These results indicate that heavy metal exposure can affect genomic stability in cancer-related genes, highlighting its role in cancer development [26].

Unlike earlier research, one study looked at how methyl-mercury (MeHg) exposure relates to cancer deaths among people living near the Shiranui Sea in Kumamoto, Japan, with a focus on leukemia and gastric cancer. The study found that the age-standardized mortality ratio (ASMR) for leukemia was higher in both groups, especially early on in the exposure group. In contrast, the ASMR for gastric cancer was lower, possibly related to salt intake. These results show that more individual-level studies are needed to better understand how MeHg affects cancer [13].

A number of biological processes have been proposed that would support the link between mercury exposure and carcinogenesis, although there is no specific link between mercury and CRC. Mercury can induce oxidative stress through the disruption of the antioxidant defense system both enzymatically and non-enzymatically causing the production of free radicals, genetic instability, and mutations [6, 7, 27–32]. Mercury might also inhibit DNA repair processes by targeting enzymes involved in single-strand DNA repair mechanisms, as well as causing alterations in the epigenetics, for example in DNA methylation [8, 33–37]. Such biological processes could play a role in cancer development, including CRC, but the majority of scientific data supporting this statement come from in vitro studies, animal studies, and non-CRC models.

As mentioned above, there is no strong evidence for a role of mercury in colorectal cancer. The role of Selenium has been investigated in several studies [38–40], but the role of mercury has not been considered to date as one of our best-known. Although the relationship between mercury and thyroid cancer has been studied in a systematic review by Webster et al. [41], the lack of significant associations between mercury levels and colorectal cancer risk challenges previous assumptions that environmental toxins are direct contributors to CRC development. It underscores the need for further research to explore other potential risk factors, including dietary habits, genetic predispositions, and additional environmental exposures that may interact with mercury to influence cancer risk.

Future studies should aim for larger sample sizes and standardized methodologies for measuring mercury levels to enhance the reliability of findings. Longitudinal studies tracking mercury exposure over time in diverse populations could provide valuable insights into the cumulative effects of mercury on colorectal cancer risk.

Despite the valuable information presented, there are certain limitations associated with this meta-analysis. Firstly, significant heterogeneity was present among the studied groups in terms of methodology used, population studied, source of biological samples, periods of exposure and mercury measurement methods. Some subgroups analysis was based on few studies or effect sizes that met eligibility criteria. None of the included articles used experimental methodology; hence no causal association can be stated. The use of observational data also increases chances of potential selection bias, measurement errors and uncontrolled confounders that can affect the associations found. Additionally, all three factors, including CRC, mercury exposure, and matrix in which mercury was measured, were inconsistently analyzed throughout the studies. Assessment of publication bias was also constrained by the few numbers of effect sizes used in the study. Accordingly, even though the funnel plot analysis and Egger’s and Begg’s tests were conducted and reported, these results cannot rule out publication bias, hence should be cautiously interpreted. There was an imbalance in geographical representation since some areas did not provide any data for the study. This might have reduced the generalizability of this study. Furthermore, there was a limitation in language since the literature search involved only studies that were written in English language. Language bias may have occurred during data collection as the study involved studies from only one language. Lastly, the study did not include its protocol in the PROSPERO database.

Conclusion

In summary, the present systematic review and meta-analysis have not found a significant difference in the level of mercury between colorectal cancer patients and healthy controls or between cancerous and non-cancerous colorectal tissues from the same individuals. Nonetheless, the lack of sufficient data, high levels of heterogeneity, and diverse ways of estimating mercury exposure do not allow concluding about no association between the exposure to mercury and the development of colorectal cancer. The results of small-to-moderate effect, matrix-, population-, or exposure-specific association cannot be excluded. It is necessary to conduct further research that will allow revealing the relationship between mercury exposure and colorectal cancer.

Authors’ contributions

RH, MM, and MV developed the title. MN, AS, RH and AM were responsible for data extraction. MV, MM, and MH collaborated on writing the initial draft. RH and MV finalized the manuscript. All authors have consented to the publication.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

Data can be obtained by reaching out to the corresponding author of the study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.