Ketone bodies and cancer risk in the general population: The prevention of renal and vascular end-stage disease (PREVEND) study

Ketone bodies (KBs) comprise acetoacetate (AcAc), acetone and β-hydroxybutyrate (β-OHB). In the state of limited glucose supply, KBs are produced from fatty acids in the liver and transported to the extrahepatic cells as an alternative source of energy. Research interest in the association between KBs and cancer is growing, largely because a ketogenic diet has been reported to have the potential to delay cancer development by starving proliferating tumour cells, which highly depend on aerobic glycolysis and thus cannot use KBs that are produced during fasting as an alternative source of energy (i.e., the Warburg effect).¹

Conversely, metabolomic studies have shown that higher KBs, as a sign of metabolic dysregulation, are associated with cancer development and progression.^{2, 3} One possible reason for these conflicting findings between KB levels and cancer incidence may be the lack of adjustment for important confounders.⁴ Another possible explanation can be differences in study populations of cancer patients versus the general population, as the effects of KBs on cancer may function through unique pathways subject to a specific pathophysiological context. For example, the effect of KBs in sensitising cancer cells to radiotherapy only applies to cancer patients, not to healthy individuals. Of note, available cohort studies on the association between KBs and cancer in the general population are scarce yet important, as KBs have been suggested to be promising markers to screen for cancer.⁵

Taking these considerations together, we aimed to examine whether and how KBs are associated with cancer incidence in a prospective population-based cohort study with extensive adjustment for confounders.

For this study, we analysed data from 6079 subjects in the Prevention of Renal and Vascular End-stage Disease (PREVEND) study, a prospective, population-based cohort.⁶ Total KBs (AcAc, β-OHB and acetone) were measured by the plasma-based NMR method. The primary outcome was the incidence of overall cancer. Secondary outcomes were the incidence of urinary tract, lung and colorectal cancer. Cox regression models were performed to calculate hazard ratios (HRs, 95% CIs), crude and adjusted. Details of the methods are provided in the (Tables S1 and S2, Figure S1).

While our study found no association between KBs and cancer development, several studies suggested potential protective effects of KBs on cancer. Notably, most of these studies suggesting the antitumour effects of increasing KBs were conducted in the context of a ketogenic diet designed specifically for cancer patients.^{7, 8} One reason why the direction of the association between KBs and cancer is different between our findings and these studies may be that the beneficial effects of KBs on cancer are unique for patients with prevalent cancer (e.g., discriminately intensify the effects of anticancer therapy for tumour cells). An alternative explanation is that therapeutically induced KBs and chronically, spontaneously elevated KBs may have different effects on cancer development. Several authors have argued that therapeutically induced KBs can positively influence body composition and enhance the anticancer effects from radio- and chemotherapy,⁹ while chronically elevated KBs may be related to pro-tumour modulation as an onco-metabolite or act as a confounding factor, as described further below.⁸ Of note, marked elevations in circulating KB concentrations among the healthy general population are usually observed only in certain situations, including severe exercise, extended fasting or maintaining a ketogenic diet. Given that >99% of subjects of our study population had β-OHB concentrations within the normal range (.5 mM), and that the use of ketogenic diets was not widely introduced or recommended to the general population in the era when PREVEND subjects were included, the associations between KBs and cancer risk documented in our study are highly likely to reflect the associations for chronically, spontaneously elevated KBs. Importantly, there is accumulating evidence suggesting that chronically elevated KBs are involved in subclinical metabolic alterations.^{10, 11} Additionally, this also indicates that our study cannot rule out the potential benefits of KBs induced by ketogenic diet on slowing cancer progression.

To date, a few metabolomic profiling studies have also shown a positive association between KBs and cancer, which we did not find in the current study.^{5, 12} One of the possible explanations for the discrepant findings can be the lack of controlling for important confounders (e.g., type 2 diabetes and inflammation) in these metabolomics studies. This is crucial because it may not be the direct effects of increasing KBs but the underlying factor driving the increase in KBs that has pro-tumour effects, as suggested in our data by the attenuation in the association between KBs and cancer incidence after adjusting for confounders.

Several underlying stimuli of increasing KBs may confound the positive association between chronically elevated KBs and cancer. First, a higher level of KBs may capture subjects with older age in the general population, as KBs can re-balance the decreased mitochondrial pyruvate dehydrogenase complex activity related to aging.¹³ As indicated by the attenuation of the association between KBs and cancer incidence after adjusting for age in our data, it is likely that KBs per se do not predict higher cancer risk, but that the older subjects captured by higher KBs are intrinsically at a higher risk of developing cancer. Second, although there is no specific indication in our data, metabolic dysregulation itself may play a role in the observed association between KBs and cancer. In other words, the increase in KBs may indicate the process of maintaining energy homeostasis in response to metabolic dysregulation.¹⁴ Third, the increase in systemic inflammation that may occur in parallel with increasing KBs is another crucial consideration. Since inflammation is a well-established risk factor driving carcinogenesis,¹⁵ the higher cancer risk associated with increasing KBs may be at least partly explained by the concurrent activation of inflammation. Importantly, while our analyses cannot definitively establish whether inflammation, characterised by hs-CRP in this study, serves as a mediator or a confounder, our findings suggest that if hs-CRP acts as a mediator, it may account for approximately 21% of the association between ketone bodies and overall cancer incidence.

Our study had several strengths. To our knowledge, this is the first prospective cohort study to investigate the association of KBs with cancer incidence in the general population. Another strength is the well-phenotyped cohort, which enables us to explore the effects of several clinically important confounders on the association between KBs and cancer. Additionally, KBs concentration can be measured by analysis of plasma, urine and breath, but the plasma-based NMR method that we used to measure KBs concentration in this study possesses higher accuracy.¹⁶ Lastly, data on cancer incidence were verified via record linkage with Palga, the Dutch nationwide pathology databank, which has complete national coverage.¹⁷ Limitations include that the plasma samples used for measuring KBs were obtained after overnight fasting, meaning KBs concentrations could be slightly higher in the nonfasting state. However, such higher KBs concentrations are expected to consistently inflate the magnitude of the observed association among participants, whereas we found a null association, making our findings potentially more robust.

In conclusion, a higher level of KBs is associated with a higher risk of incident overall, urinary tract, lung and colorectal cancer in the general population, but these associations are nullified after adjustment for clinically important confounders.

All authors conceived and designed the study. L.M.K., B.v.d.V., E.G.G., S.J.L.B., R.A.d.B., M.A.C., R. P. F. D. and R.T.G. contributed to data acquisition. L.L. conducted data analysis. All authors contributed to the interpretation of the data. L.L., M.G.E.K., L.M.K. and R.T.G. drafted the manuscript. All authors revised the article. L.M.K. and R.T.G. supervised the work. All authors approved the final version of the manuscript.

The PREVEND study is supported by several grants from the Dutch Kidney Foundation (E.033) and the Dutch Heart Foundation (2001.005), the Dutch Government, the US National Institutes of Health and the University Medical Center Groningen, the Netherlands. L.L. is supported by a scholarship from the China Scholarship Council (CSC number: 202008440376). Dr. de Boer is supported by the European Research Council (ERC CoG 818715).

M.A.C. is an employee of and holds stock in Labcorp. Other authors declare not to have conflicts of interest for the present work.