Digestive cancers represent a significant global health burden, accounting for a substantial proportion of cancer incidence and mortality worldwide (1). Risk factors for digestive cancers are multifactorial and include genetic predisposition, chronic inflammation, infection (e.g., Helicobacter pylori or hepatitis viruses), lifestyle factors such as smoking and alcohol consumption, and particularly, diet quality (2, 3). Growing evidence suggests that certain dietary patterns as well as the consumption of ultra-processed foods (UPF) and low intake of fibre-rich foods, may play a key role in the development of cancer (4, 5).

According to Monteiro et al. (6), UPFs are industrial formulations typically made from substances derived from foods and additives, with little if any whole food added. These products are often characterized nutritionally by elevated levels of total, trans, and saturated fatty acids, salt, and simple sugars, as well as a diminished presence of fibre and micronutrients (6, 7). The growth consumption of these products can result in excessive calorie intake and an imbalance in micro- and macronutrient intake (8). UPF consumption has increased significantly and now constitutes approximately 50 to 60% of the total daily energy intake in medium and high-income countries, replacing traditional eating patterns based on regular freshly prepared meals (9–11). Particularly in Spain, UPFs represented more than 30% of daily energy purchases by 2010 in Spanish households (12). Epidemiological evidence suggests that this extensive consumption and distribution of UPF could be responsible for the global burden and the increasing prevalence of non-communicable diseases such as obesity, cardiovascular disease (CVD), diabetes and various cancers (7, 10, 13–15), which together currently account for approximately 74% of all deaths worldwide (16). In the Spanish population, higher UPF consumption has been associated with increased risk of abdominal obesity (17) and a poorer cardiometabolic profile (18), conditions which have been proposed as potential pathways that may mediate the relationship between UPF consumption and digestive cancer development.

Increasing evidence from prospective cohorts and meta-analyses suggests that high consumption of UPF may be associated with an elevated risk of overall cancer (19). For example, results from large-scale studies such as the NutriNet-Santé cohort (20) and the UK Biobank (21), as well as findings from case–control studies in non-Western populations like Morocco (22), have consistently pointed toward a positive association between UPF consumption and cancer risk. However, most research has focused on overall cancer incidence, and few studies have looked into how UPF consumption relates to specific different types of cancers, particularly within Mediterranean populations.

In this regard, a previous meta-analysis showed that the highest categories of processed food consumption were associated with a 78% increased risk of oesophageal cancer compared with the lowest categories (23). Similar results were observed in the EPIC study, where higher UPF consumption was associated with 24% greater risk of oesophageal cancer (24). In the case of stomach cancer there is limited information, but a case–control study showed an increased risk from higher UPF consumption (25). Finally, in a recent study on pancreatic cancer, a high UPF consumption was found to be associated with an increased risk (14). Nevertheless, the present evidence on the association between UPF consumption and specific digestive cancers (like oesophagus, stomach and pancreas), is very scarce. Thus, we explored the association between UPF consumption and cancers of the oesophagus, stomach, and pancreas in a multi-case control study carried out in a Mediterranean area of Spain.

The distribution of cases and controls according to sociodemographic and lifestyle characteristics is shown in Table 3. Educational level was similar for both cases and controls. Alcohol consumption and tobacco smoking were more prevalent in oesophageal cancer than in pancreatic and stomach cancers and controls. In addition, we observed higher energy intake and UPF consumption in oesophageal cancer cases.

The association between total consumption of UPF and risk of oesophageal, stomach, and pancreatic cancers is presented in Table 4. In multivariate analysis, a higher UPF consumption was associated with higher risk of oesophagus and stomach cancer. Compared with the lowest tertile of UPF consumption, the medium and highest tertiles of UPF consumption were associated with higher risk of oesophagus cancer, RRR = 1.85 (1.13–3.04) and RRR = 2.29 (1.37–3.82), respectively. For every 100 grams of UPF consumption, the risk of oesophagus cancer increased by a 25%. Similarly, compared with the lowest tertile of UPF consumption, the medium and highest tertiles were associated with higher risk of stomach cancer, RRR = 1.58 (1.13–2.21) and RRR = 1.56 (1.08–2.27), respectively. In addition, a 100-gram increase in UPF consumption was associated with 13% higher risk of stomach cancer. No associations were observed for pancreatic cancer.

When we evaluated the association between UPF consumption as percentage of total food consumption, we also observed a higher risk for oesophageal and stomach cancers (Supplementary Table 1). Compared with the lowest tertile of UPF consumption, the highest tertile was linked to a 68% higher risk of oesophageal cancer. For every 2% increment of UPF consumption, the risk of oesophagus cancer increased by an 11% and the risk of stomach cancer increased by a 6%. No association was observed for pancreatic cancer (Supplementary Table 1).

The associations between the different subgroups of UPF in total grams per day and oesophageal, stomach, and pancreatic cancers are presented in Table 5. For UP dairy products, the consumption of 100g/day was significantly associated with a higher risk of oesophageal cancer, RRR = 3.08 (1.16–8.13), although no evidence of trend was observed by tertiles of consumption. For stomach cancer, a significant trend was observed by tertiles of consumption (p-trend <0.001), medium and highest tertiles showed higher risk of stomach cancer, RRR = 1.44 (1.03–2.03) and RRR = 2.10 (1.43–2.82), respectively. Moreover, each additional 100g/day of consumption was significantly associated with a higher risk, RRR = 4.13 (1.85–9.20). Similarly, the consumption of sweets and pastries was positively associated with stomach cancer, compared to lowest tertile, those in the medium and highest tertiles showed higher risks, RRR = 1.79 (1.27–2.51) and RRR = 1.76 (1.24–2.50), respectively (p-trend = 0.002); the consumption of 100 g/day of sweets and pastries was associated with higher risk of stomach cancer, RRR = 1.76 (1.14–2.72). Regarding UP drinks, the highest tertile of consumption was associated with a higher risk of oesophageal cancer, RRR = 1.81 (1.13–2.90), with a significant positive trend (p-trend = 0.021); in addition, every 100 g/day of UP drinks increased in consumption was associated with a 21% higher risk of oesophageal cancer. Finally, consumption of pre-cooked foods also increased the risk of oesophageal cancer, every 100 g/day of consumption was associated with higher risk, RRR = 2.83 (1.54–5.19). No association was observed for processed meat consumption and any type of digestive cancer. Similar results were observed when we assessed the effect of UPF subgroups as percentage of total food consumption (Supplementary Table 2).

In this multi-case control study, we observed that higher consumption of UPF was associated with an increased risk of oesophageal and stomach cancers. Regarding subgroups of UPF, the consumption of UP dairy products was associated with greater risk of oesophageal and stomach cancers; the consumption of sweets and pastries was associated with stomach cancer and the consumption of UP drinks and pre-cooked foods was linked to an increased risk of oesophageal cancer. No associations were found between UPF consumption and the risk of pancreatic cancer.

A previous systematic umbrella review based on existing meta-analyses concluded that higher UPF consumption was associated with an increased risk of overall cancer (34). The most recent evidence based on observational studies have shown a positive association between UPF consumption and overall cancer mortality (21, 35), although others studies have found no association (36–38). The association found in our study for oesophageal and stomach cancer would be in line with the present general evidence although it is based on two cancer sites only.

However, our results for oesophageal cancer are consistent with those reported in other studies. In a population-based case–control study in China (39), the highest quartile of UPF consumption was linked to an almost 3-fold increased risk of developing oesophageal cancer. Similarly, in the EPIC multi-centre prospective cohort study in Europe, the increment of 10% g/day of UPF consumption was associated with a 24% increase in the risk of oesophageal cancer (24). In the same line, Chang et al. (21) in the UK Biobank prospective cohort study showed that per 10% increment in UPF consumption in the total diet, the risk of oesophageal cancer increased by 8%, although this association was not statistically significant.

When we explored the association between the subgroups of UPF and oesophageal cancer risk, we found significant positive associations with UP dairy products, drinks and pre-cooked foods. In the study by Navarro Silvera et al. (40), the consumption of high-fat dairy, mostly UP dairy foods, including cheese or cheese spreads, ice cream, milkshakes, puddings, custard, flan or milk-based beverages, was associated with an increased risk of both oesophageal adenocarcinoma and squamous cell carcinoma; however, other epidemiological studies did not find associations between specific categories of dairy products and the risk of developing oesophageal cancer (41, 42). In contrast, recent epidemiological evidence suggests a positive association between the consumption of UP drinks like sugar-sweetened beverages (SSB) or alcoholic drinks and oesophageal cancer. In a pooled analysis of two case–control studies in US (43), the SSB consumption was associated with a 55% increase in oesophageal cancer risk. In addition, in a dose–response meta-analysis by Yu et al. (44), each additional 12.5 g/day of ethanol consumption was associated with a 33% increase in oesophageal cancer risk, with higher risks for spirits and liquors. Similarly, in a meta-analysis of 106 cohort studies (45), the association between alcohol intake and oesophageal cancer was evident even for low levels of intake. The positive association between pre-cooked foods and oesophageal cancer we found is consistent with that reported by Yan et al. in a recent meta-analysis (23).

Previous studies have shown an increased risk for gastrointestinal cancer related to UPF consumption (21, 46), but few studies have specifically explored stomach cancer. In our study, we observed that high UPF consumption was associated with a 56% higher risk of stomach cancer. In a recent meta-analysis of five prospective cohort studies (4) the highest UPF consumption was significantly associated with an increased risk of non-cardia gastric cancer, while no association was found with gastric cardia cancer. Similarly, a prospective cohort study within the UK Biobank (21) found that the highest level of UPF consumption was linked to a greater risk of non-cardia gastric cancer, while an inverse effect was found between high UPF consumption and gastric cardia cancer, although none of these associations were statistically significant. When we evaluated specific associations between UPF subgroups, we found that the consumption of UP dairy products and sweets and pastries were associated with a higher risk of stomach cancer. These results are consistent with those reported by Tayyem et al. (47), who observed an increased risk of stomach cancer among participants with higher consumption of UP dairy products. Furthermore, in a previous analysis of the PANESOES study population (48), a high consumption of UP dairy products was associated with an 85% higher risk of stomach cancer. In addition, Peres et al. (25), carried out a study in Brazilian population and found that medium and high intake of sweets increased the risk of stomach cancer by 74 and 125%, respectively. Similarly, Almahri et al. (49), showed that the intake of sweets such as candies and biscuits was significantly associated with an 87% increased risk of stomach cancer.

Finally, regarding the effect of UPF consumption on pancreatic cancer, we did not observe any association. In the EPIC study, the UPF consumption showed evidence of a certain trend, although it did not reach significance (50). However, in a population-based cohort study of 98,265 American adults (14), a high consumption of UPF was associated with an increased risk of pancreatic cancer; participants in the highest quartile of UPF consumption had 47% higher risk of pancreatic cancer compared with the lowest quartile. To the best of our knowledge, no prior studies have specifically examined the relationship between subgroups of UPF and pancreatic cancer. Unlike our findings for the other cancers, no significant associations were found between the consumption of UPF subgroups and the risk of pancreatic cancer.

With respect to potential mechanisms that may underlie the observed positive association between UPF and cancer we should consider the obesogenic properties of some UPF, as well as the exposure to food additives and contaminants (11). In addition, other factors may play a role, such as the relationship between UPF consumption and a poor overall diet quality with excessive calorie intake, primarily due to the high energy density and low satiety often associated with these foods (8). Some studies have shown that diets high in UPF tend to contain significant levels of trans fatty acids, salt, and/or sugars as well as lower content in potassium, protein and fibre, which have been associated with higher prevalence of non-communicable diseases in observational studies (51). Another potential mechanism may relate to the link between higher consumption of UPF and obesity (52), a known risk factor for overall and specific cancers like breast, colorectal and pancreatic cancer (19). Food contaminants found in UPF like neo-formed compounds that arise during their chemical, biological, and physical transformations, could also represent a potential mechanism for increasing the risk of cancer; in this sense, UPF typically contains neo-formed processing contaminants (e.g., trans fatty acids) that may be positively associated with cancer risk due to high-heat processing (53, 54). It has also been suggested that several contaminants from food packaging may be present in UPF. For example, increased urinary concentrations of bisphenol and phthalates were observed in a young U. S. population with high UPF consumption, which could suggest that UPF packaging might be one of the links potentially contributing to adverse health outcomes, including cancer (55). However, further research is required to confirm whether additives, packaging compounds, or heat-produced neo-formed contaminants can contribute to cancer development, since the observational nature of our study cannot establish causality.

We acknowledge limitations in our study. As in other case–control studies, our study is more susceptible to certain types of bias, such as selection bias. However, in this regard, it is important to note that the participation rate of cases and controls was similar. Another limitation inherent to the case–control design is recall bias. In this sense, although dietary intake data were collected during hospital interviews, they referred to the five years prior to diagnosis making more unlikely that diet be a consequence of the disease when the interview was done. In addition, the study sample was limited (especially for oesophageal and pancreatic cancers), which hinder the possibility of exploring associations by histological subtype. The observational nature of our study limits the possibility of inferring causal relationships, making it necessary to conduct further studies with larger samples with a long follow-up period, although these studies may be less feasible. Moreover, despite performing multivariate analyses that included relevant known risk factors such as tobacco and alcohol consumption, the possibility of residual confounding remains.

Our study also has several strengths. Firstly, we used the NOVA classification system, which is a well-established and widely used system in the literature and allowed us to specifically and accurately examine the association between UPF and digestive cancer. Secondly, our dietary assessment tool has been previously validated (30). Thirdly, our findings align with those reported in the existing literature regarding the harmful effects of UPF consumption. Nevertheless, our results are novel because we explore the above-mentioned effect on three different gastrointestinal cancers which have been understudied despite being highly relevant in public health. Finally, although our study was hospital-based, our findings may be generalizable to other populations with similar characteristics, as cases and controls were recruited from public referral hospitals within the Spanish National Health System, which provides universal and free access to healthcare for the entire population.

Our study suggests that higher UPF consumption is associated with higher risk of oesophageal and stomach cancers, and that the UPF subgroups associated were UP dairy products, sweets and pastries, alcoholic drinks and pre-cooked foods. No association was observed for pancreatic cancer. Further studies are needed to confirm these results.