Cancer, which is increasing around the world, is expected to rank as the leading cause of mortality and the main barrier to life expectancy (1). Iron plays a vital role in cell replication, metabolism, and growth and is becoming more relevant in the research of chronic disease and cancer (2). However, iron status can be a double-edged sword, for excess iron status has been determined to be harmful to human health in that it generates a lot of free radicals, such as reactive oxygen species, in reaction with hydrogen peroxide (3). Excessive free radicals are a well-known cause of cell and tissue organ damage, and these processes may affect the development of cancer (4). One study showed that patients with hereditary hemochromatosis, a genetic iron disorder characterized by elevated plasma iron (PI) levels, had an increased risk of liver cancer (5).

A large number of studies have shown that abnormal iron homeostasis is one of the markers of cancer. PI has traditionally been used as a short-term marker of body iron status (6). The results on the relationship between blood iron and cancer risk have not always been consistent and no statistically significant associations (7) or positive associations (8, 9), as well as inverse associations (10), have been indicated. A cohort of 309,443 adults in Taiwan showed that high PI (≥120 mg/dL) is a marker of increased risk for total cancers, especially liver cancer (8). In contrast, a meta-analysis in a Chinese population found that PI levels were lower in patients with cervical cancer than in controls (10). A clinical trial also found that a reduction of iron stores was associated with a modestly decreased cancer risk and mortality (7). However, a large prospective cohort from Sweden reported no association between PI and cancer risk except for postmenopausal breast cancer (11). Additionally, one prospective study reported that the relationship between PI and nonskin cancers was converse between the sexes with a negative association in males and a positive association in females (12). Taken together, the evidence that links iron status with cancer has yielded mixed results.

Different associations between nutrients and cancers may occur in individuals with or without hypertension. All the previous studies were conducted in general populations. However, to date, there is a lack of data on cancer incidence associated with PI among patients with hypertension. Given the conflicting findings from epidemiologic studies, we hypothesize that the association between iron status and cancer risk is site-dependent and varies by population characteristics. In the current study, we prospectively examined the association between PI and total cancer and its subgroups in Chinese adults with hypertension and evaluated possible effect modifiers on the iron-cancer association.

This nested case-control study suggests a nonlinear association between PI and the risk of total cancer, with higher cancer risk associated with both higher and lower iron levels in Chinese adults with hypertension. Moreover, we found that higher iron levels (≥150.4 ug/dL) were linked to an increased risk of digestive system cancers, whereas lower iron levels (<114.1 ug/dL) were linked to an increased risk of non-digestive system cancers. Our findings demonstrate that the association of iron with cancer might differ according to tumor location.

The mean PI level in our study was 144.2 ug/dL, which was remarkably higher than that observed in the United States (83.88 ug/dL), with data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2016 (14). To avoid reverse causation bias and potential confounding bias, we additionally excluded cancer cases that had occurred within the first year, as well as adjusted for potential confounders, and found that the positive association was still observed. Our results support the idea that body iron plays a key role in human carcinogenesis (2). Some studies have reported sex differences in the association of PI with cancer risk (12, 15). A prospective cohort study of US adults suggested that participants with PI above 140 µg/dl had a significantly increased risk of cancer mortality (15) and this association was stronger in females. However, no sex difference was observed in our study.

The relationship between PI concentrations and digestive system cancer risk observed in this study is in line with the prevailing hypothesis that higher body iron is associated with an increased risk of liver cancer (16). A systematic review showed that PI was positively correlated with liver cancer risk (HR: 2.47; 95%CI:1.31-4.63) (16). In a large cohort study in Taiwan, it was revealed that PI levels higher than 120 µg/dL or lower than 60 µg/dL were associated with a 25% increase and 18% increase, respectively, in all cancer incidence with 60 to 79 ug/dL as the reference level (8), primarily with liver cancer. A Mendelian randomization study in a European cohort suggested that a genetically high iron status was positively associated with liver cancer (17). For other digestive tract tumors, Milde et al. observed that adenocarcinoma colorectal patients had significantly higher levels of PI as compared to the age-matched control group (18). Results from the European Prospective Investigation into Cancer and Nutrition (EPIC) based on 481,419 individuals and 137 incident cases of esophageal adenocarcinoma, showed a statistically significant positive association of esophageal adenocarcinoma risk with heme iron and processed meat intake (19). Recent research suggests that red meat (high iron content) intake could be a risk factor for gastric cancer (20). However, one study from a large European population found that high body iron stores as measured by PI and ferritin decreased the risk of gastric cancer (21). Another study also found that high iron stores may increase the risk of colorectal cancer, whereas low iron stores may be an early sign of occult stomach cancer (22).

Data on the association between PI and non-digestive system cancers are limited and inconclusive. One study reported a dual role for iron in breast cancer where a proangiogenic environment mediated by iron deficiency and pro-oxidant conditions associated with iron accumulation could lead to a high incidence of breast cancer (23). A recent meta-analysis found that high levels of PI were associated with an increased risk of breast cancer (24). Nonetheless, some studies have reported results that were similar to ours (10, 25). A meta-analysis in a Chinese population found that PI levels were lower in patients with cervical cancer than in controls (10). A previous publication also reported that iron deficiency is known to be associated with bladder cancer (25). A study by Yuan et al. suggested that genetically high iron status was inversely associated with brain cancer based on 48,972 individuals of European-descent (17). However, a meta-analysis showed no association between PI levels and lung cancer risk (26). Epidemiological data on the association of iron status with other non-digestive system cancers are limited and scarce. More studies are needed to better elucidate this possible relation. Our study has provided some new insights. We demonstrated that the relationship between PI levels and cancer risk varies for different cancer subtypes, with a positive association for digestive system cancers and negative association for non-digestive system cancers.

Plausible reasons for the positive correlation between PI and digestive system cancers could be explained by the fact that the gastrointestinal tract is the major site of nutrient digestion and absorption, and it is prone to oxidative damage (27). Accumulated iron has been implicated in the risk of cancer through iron-catalyzed free radical-mediated oxidative stress and subsequent DNA damage, and iron also functions as a nutrient that fosters the growth and development of cancer cells. Additionally, it has been postulated that iron insufficiency could lead to oxidative deoxyribonucleic acid damage, thereby increasing the risk of cancer (28, 29). Elemental iron is crucial for the proper functioning of enzymes involved in cell respiration, energy metabolism, DNA synthesis and repair, signaling, and many more. Iron is an important trace element required for the formation of hemoglobin and myoglobin and is a key component in immune cell proliferation. Therefore, iron is both essential and potentially toxic. However, the specific mechanism of iron metabolism in different types of tumor cells is not yet fully understood and the interaction mechanism of retinol and iron should be explored further.

There are several strengths of this study. The data are directly derived from a large cohort study, with strict quality control measures. Furthermore, our study utilized a matched design to minimize potential data bias as much as possible. First, we focused our analysis on PI, one of the biomarkers of iron status, and the analysis did not include other measures that are thought to more closely reflect levels of iron in the body, such as ferritin levels. As PI is a relatively inexpensive and widely available test, tests for high PI status could be routinely conducted and interpreted in daily practice, making our results clinically relevant. Moreover, Chua et al. suggest that circulating iron may be more related to breast carcinogenesis than stored iron (ferritin) (12). A second study limitation was that the average follow-up time from the time of PI testing was 2.13 years, which may be too short a time for assessing cancer development. A third limitation is that the conclusions from our study may not apply to other populations as the participants were patients with hypertension. The differences in findings might be attributed to residual confounding by other confounders, such as unhealthy dietary patterns or other lifestyle behaviors.

In conclusion, our study provides evidence of a nonlinear relationship between iron status and overall cancer. Considering the small number of cancer cases, additional studies with larger sample sizes are warranted to verify our findings. Furthermore, the effects of iron on different cancer types need to be explored. The results of this study have important clinical implications with regard to the excessive supplement of iron and iron deficiency.

The datasets presented in this article are not readily available because Data described in the article, code book, and analytic code will be made available upon request pending approval. Requests to access the datasets should be directed to Zhu Hehao, zhuhehao@foxmail.com.

The studies involving humans were approved by The CSPPT was approved by the Ethics Committee of the Institute of Biomedicine, Anhui Medical University, Hefei, China (FW A assurance number: FW A00001263). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

The authors’ responsibilities were as follows— HeZ and YW: data curation, writing original draft; QH: supervision; YuS and LL: audited the data; XX: conception and conducted research; HG, YoS and HaZ: helped with copyediting; BW: had primary responsibility for the final content of the manuscript. All authors read and approved the final manuscript.