This nationwide population-based cohort study demonstrated an 18.8% increase in total cancer risk among ND patients compared to control patients without ND over a 10-year follow-up period. The adjusted HR for cancer in patients of the ND group was 1.25 (95% CI, 1.19 ~ 1.31; P < 0.01) compared with the HR in patients of control group. During the 10-year follow-up period, compared with the pure control group where no ND or T2D occurred, the cancer risk ratio of the ND group additionally diagnosed with T2D and other ND during the follow-up period increased to 1.87. These results suggest that the cancer risk is further increased as new ND and T2D were additionally diagnosed during the 10-year follow-up period. These findings are contrary to the results of previous epidemiologic studies examining the association between ND and cancer. This difference between the results of this study and previous ones is likely due to methodological differences. In many previous studies on the association between a particular ND (AD or PD) and cancer, the control group for ND was simply selected as subjects not having AD or PD. In such cases, NDs other than AD or PD may be included in the control group, which may affect cancer development in the control group. To confirm the selection bias in the process of control group selection, we separately analyzed only data of AD patients, and cancer incidence rates were compared with two types of control groups for AD (non-ND control group and non-AD control group). The number of cancer patients in AD patients was 632/3,141 (20.1%). When the control group was selected as patients without any ND (non-ND control group), 8,628 (18.5%) out of 46,742 patients in the non-ND control group were diagnosed with cancer (χ2 = 5.38, p = 0.02, compared with AD patients), but when the control group was selected as patients only without AD (non-AD control group; in this case, ND other than AD may be included in the control group), 10,036 (19.0%) out of 52,893 patients in the non-AD control group were diagnosed with cancer (χ2 = 2.53, p = 0.11, compared with AD patients). Therefore, this finding shows that cancer incidence slightly increases when NDs other than AD were included in the control group. These results indicate that the lower cancer incidence in the AD group compared to the control group in previous epidemiologic studies could not completely rule out the effect on cancer incidence of included NDs other than AD in the control group. In this study, patients without any ND were selected as a control group, and as a result, the incidence of cancer was higher in the ND group over a 10-year follow-up period compared to the control group.
During the 10-year follow-up period of this study, there were new cases of ND and/or T2D not present at the beginning of the study. These new ND and/or T2D cases may affect the development of subsequent cancers; therefore, we also investigated the impact of these new cases on the development of subsequent cancers. In both the control and ND groups, newly diagnosed ND and T2D showed an additive effect on the incidence of cancer during the follow-up period. The additive effect on cancer incidence was higher in T2D than in ND. Moreover, the additive effects of the newly diagnosed ND and T2D on cancer incidence were synergistic. This study showed a positive association between ND and cancer and that T2D had an additive effect on this association. These results are consistent with the fact that, unlike previously thought only in the aspect of apoptosis, recent molecular biology studies have suggested that the pathophysiology of cancer and NDs have a common mechanism. Signaling pathways, such as those arising from DNA damage, deviation from the normal cell cycle, inflammation, and oxidative stress that affect cell death and survival, have been studied in connection with cancer development, but recent studies have also linked them to NDs (1–3). p53, cyclin D, cyclin E, cyclin F, Pin1, and protein phosphatase 2A are commonly involved in the pathophysiology of cancer and NDs and are involved in cell cycle regulation [30, 31]. Among these, p53 is the most widely studied tumor suppressor gene and known to be associated with 50% of cancer . It is well known that p53 plays an important role in apoptosis, and in most cancers, mutations at the gene level prevent it from functioning properly, resulting in poor activity of p53, which allows cancer cells to escape from apoptosis, leading to cancer. However, not only changes at the gene level, but the protein folding process is also an important factor in preventing p53 from normal functioning [33–35]. Cancer occurs due to not only a mutation in the p53 protein but also a problem in the p53 folding process . In addition to the association between ND and cancer, many studies have reported that T2D is also associated with ND and cancer. In the relationship between T2D and ND, amylin’s (or IAPP) functions in the periphery and its impact on pancreatic cell function and T2D progression have been described [6, 37]. The presence of AD-related proteins (β-amyloid protein and tau protein) in the pancreas and insulin-sensitive tissues and their roles in inducing peripheral insulin resistance or disruptions in insulin secretion are a potential mechanistic link among these AD-related proteins promoting T2D. Similarly, amylin accumulation in the brain, its ability to induce neurotoxicity, and form “cross-seeding” aggregates with β-amyloid protein illustrates the role for this pancreatic amyloidogenic protein in neurodegeneration [38, 39]. The association between diabetes and cancer is well established [40, 41]. Epidemiologic studies suggest that people with diabetes (predominantly T2D) are at a significantly higher risk for many forms of cancer [42, 43]. However, the risk of prostate cancer is lower in men with T2D . The relationship between T2D and cancer varies depending on the cancer site, and more rigorous and collaborative studies are needed to understand the relationship between the two diseases.
Our study has some limitations. First, this study was a retrospective observational design based on claims data, and the diagnoses for NDs identified using KCD-6 codes in the claims database, which may have been inaccurate compared with diagnoses obtained from a medical chart, neuroimaging tests, or neuropsychological tests. Second, this study did not analyze the association between NDs and various types of cancer. Third, patients with NDs tend to have less frequent contact with the healthcare system, including cancer examinations, which may have caused surveillance bias. We investigated the utilization rate of medical institutions in patients of the ND and control groups to examine the possibility of this surveillance bias. The annual average number of visits to medical institutions before cancer onset was investigated. There was no significant difference in the annual average number of visits to medical institutions between the two groups, with an average of 26.8 ± 19.8 in all patients in the ND group and an average of 26.6 ± 18.7 in all patients in the control group (t = −0.75, p = 0.45). Even when separately analyzing patients with cancer, there was no significant difference in the annual average number of visits to medical institutions between the two groups, with an average of 25.0 ± 18.7 in all patients in the ND group and an average of 24.8 ± 16.7 in all patients in the control group (t = −0.55, p = 0.58).