Albuminuria, Kidney Function, and Cancer Risk in the Community

Yejin Mok; Shoshana H. Ballew; Yingying Sang; Josef Coresh; Corinne E. Joshu; Elizabeth A. Platz; Kunihiro Matsushita


Am J Epidemiol. 2020;189(9):942-950. 

In This Article


In this large community-based cohort study, lower eGFR and higher ACR demonstrated dose-response associations with higher overall cancer risk in demographically adjusted models, with stronger associations for ACR than eGFR. These associations were attenuated after accounting for shared risk factors for CKD and cancer, although ACR remained significantly positively associated with nonprostate cancer incidence. Only ACR was associated with some site-specific cancer incidence—positive for lung and urinary tract cancer, and inverse for prostate cancer. The association of higher ACR with lung cancer appeared most robust in several sensitivity analyses.

Our study is, to our knowledge, the first to evaluate eGFR using serum creatinine and cystatin C, currently considered the best estimate of measured GFR, for cancer risk.[24] Despite this methodological advantage, the association of eGFR with overall cancer risk was weak, particularly once we accounted for shared risk factors for CKD and cancer. This result is consistent with several previous studies,[12,19–21] including a retrospective study of more than 1 million people.[21] Taken altogether, individuals with lower eGFR might have a slightly higher risk of cancer, but our study suggests that the etiological contributions of reduced kidney function beyond shared risk factors is probably limited.

Our observation of higher ACR showing more evident associations with cancer risk than lower eGFR is in line with several previous studies.[12–14,16,17] Although attenuated in some sensitivity analyses, ACR showed significant associations with overall nonprostate cancer incidence in our study. There are several plausible mechanisms linking albuminuria to cancer risk. Inflammation might play a role given that inflammation has been shown to contribute to the development of some cancers,[37,38] and it also induces glomerular permeability of albumin.[39,40] In addition, vascular endothelial growth factor might be another potential contributor[41] given that it causes albuminuria and is known to play a critical role in progression, invasion, and metastasis of lung cancer.[42]

Lung cancer is the only cancer consistently showing a significant association with albuminuria in several previous studies[12,13,16,17] and in ours in ARIC (Web Figure 11). The mechanism behind the albuminuria–lung cancer association is unclear, but aforementioned biological mechanisms of inflammation and vascular endothelial growth factor might contribute to their relationship as well. Because the association of ACR and lung cancer was seen mainly in current and former smokers, it is possible that ACR is an indicator of susceptibility to carcinogenesis of smoking or indicative of tissue damage due to smoking. Nonetheless, this robust association of ACR with lung cancer seems to deserve future targeted investigations.

We also observed an inverse association of higher ACR with prostate cancer. We are not aware of any other studies reporting this association, although Lin et al.[17] showed a positive association of ACR with prostate cancer mortality. However, interestingly, a few studies reported the inverse association between the duration of diabetes, a leading cause of albuminuria, and risk of prostate cancer.[43–45] Also, a prior study found that prostate-specific antigen levels were lower in men with proteinuria than in those without.[46] Moreover, prostate cancer screening is not recommended among men with life expectancy of <10 years,[47–49] so men with more advanced CKD might be less likely to receive prostate cancer screening than those without CKD (i.e., detection bias).

Our findings warrant discussion of any clinical implications. First, it is important to acknowledge the higher risk of cancer in persons with CKD relative to those without in demographically adjusted models because current cancer screening or prevention strategies do not necessarily highlight CKD as a high-risk group.[50] From another point of view, given that increased cancer risk due to CKD is largely explained by potential confounders such as hypertension and diabetes, cancer-prevention strategies can probably target persons with these conditions before they reach CKD. The robust association of albuminuria with lung cancer suggests that health-care providers should pay attention to albuminuria levels among current smokers aged 55–80 years who are eligible for lung cancer screening using computed tomography.[51] Importantly, the measurement of albuminuria is recommended in patients with hypertension and diabetes, and thus there are a number of clinical scenarios in which albuminuria levels are readily available.

The present study has several limitations. Our study has a small number of participants with severely reduced eGFR and elevated ACR, and thus specific investigations in severe CKD cases are needed. Also, we had limited number of cases for some site-specific cancers. The assessment of kidney disease measures at a single time point does not take into account changes in exposure levels, which usually bias the results toward null. We used a common set of cancer risk factors as covariates and did not tailor to individual site-specific cancers.

Kidney disease measures, particularly high ACR, were significantly associated with cancer risk in demographically adjusted models, with implications for cancer screening and preventive strategies. The attenuation of this association after accounting for potential confounders suggests relatively limited pathophysiological contributions of these kidney disease measures to cancer development. However, the association of ACR with lung cancer was uniquely robust, warranting additional studies to explore potential mechanisms.