Background
The number of people affected by chronic kidney disease (CKD) and end-stage kidney disease (ESKD) is substantial and increasing. The number of new patients with ESKD treated by renal replacement therapy has increased at an average of 8 % per year over the past 10 years globally [
1]. Currently, over one million patients are on dialysis worldwide, a number that is estimated to exceed two million over the next decade [
2].
CKD is a risk factor for disease affecting other organs. It is well established that people with CKD are at increased risk of developing and dying from cardiovascular disease compared to people without kidney disease [
3]. There is also evidence that cancer risk and cancer mortality may be increased in people with CKD, although the associations do appear to be site-specific. It has been reported that reduced renal function is associated with an increased risk of cancers of the kidney or urinary tract [
4‐
7], lip [
8], digestive tract [
9], lung [
4] and some soft tissue and haematological sites [
10]. Among dialysis patients, there have also been reports of an increased risk of cervical and possibly thyroid cancers and a reduced risk of prostate cancer [
8,
9]. A dose-dependent relationship between albuminuria and bladder or lung cancer risk was observed in a Scandinavian study [
11].
Previous observational studies have not examined the extent to which reduced kidney function is associated with increased risk of cancer and cancer death across the full spectrum of kidney disease and in different populations. We hypothesize that reduced kidney function is a risk factor for site-specific cancer and may be a prognostic indicator of poor cancer outcomes. The objective of this study was to determine the overall and site-specific risk for incident cancer and cancer deaths from a broader population of people with CKD, varying from mild to advanced stage disease requiring dialysis.
Discussion
We analysed individual patient data from six prospective studies of 32,057 participants with various levels of renal function, followed for an average of 5 years. Although in the pre-specified analyses there was no significant association between renal impairment and the overall risk of cancer or of cancer death, several notable findings emerged when these findings were examined in greater detail. First, as compared with people with eGFR ≥75 ml/min/1.73 m
2, patients on dialysis had a non-significant excess risk of any cancer (HR 1.24, 95 % CI 0.97–1.58) together with a statistically significant increase in the risk of cancer death (HR 1.58, 95 % CI 1.09–2.30). Second, closer inspection of data on cancer risk in particular sites indicated that the lack of any overall association masked clear associations for specific cancer types: in particular, with declining renal function there were trends towards increased risks of urinary tract and endocrine (mostly thyroid) cancers but also lower risk of prostate cancer. Taken together, our findings extend previous reports of associations between renal function and cancer risk to people with a wider degree of renal dysfunction, and our findings for particular cancer sites are consistent with these earlier reports [
4,
5,
7,
8].
It is well known that the lifespan of people on dialysis is reduced as a consequence of premature death from both cardiovascular and non-cardiovascular causes [
23‐
25]. Our study findings suggest that cancer is a contributing factor to the increased risk of non-vascular death among patients on dialysis. A 1.5-fold increase in the risk of cancer death is broadly consistent with previous observational studies that have reported an excess risk of cancer in the range of 1.2 and 1.4-fold among those on dialysis using registry analyses [
8,
10], but our findings make clear that the magnitude of any relative excess of cancer in a given dialysis population will be determined by the relative frequency of different cancers which, depending on the subtype, may be associated (positively or negatively) or unassociated with declining renal function. The distribution of cancer types will in turn depend on the gender, age, and ethnicity of the population, as well as other factors.
Some previous studies have reported an association between renal function and any cancer [
4,
8] whilst others did not [
5,
9]. For dialysis patients, other studies have reported an increased risk of cancer, especially of the kidney and urinary tract [
8,
9] but also of thyroid cancer [
8] and some digestive tract cancers [
8,
9]. Previous studies have also shown that the risk of prostate cancer is reduced among dialysis patients [
9]. In contrast to previous studies [
9] we did not observe an increased risk of oral cavity, respiratory or haematological cancers among those with reduced kidney function. Moreover, whilst a previous study suggested that women on dialysis were at increased risk of cervical cancer [
8,
9], we did not observe a significantly higher risk of female genital cancers for dialysis patients. This apparent heterogeneity of the available literature is consistent with the observation that associations between declining renal function and cancer risk are dependent on cancer subtype, which may vary between different study populations, and it implies that studies (or meta-analyses of studies) involving much larger numbers of cancers with detailed subtyping information are needed to gain a better understanding of these associations.
The present study adds to the current evidence that the excess cancer observed in people on dialysis may not be driven solely by viral carcinogenesis as previously suggested [
8], but could also be influenced by the uraemic milieu associated with severe renal dysfunction. Uraemia is often characterized as a state of immune dysfunction. The different types of uraemic toxins may exert antagonistic interactions of pro-inflammatory and immunosuppressive responses, leading to increased risks of infections and malignancy [
26]. In addition, people on dialysis retain solutes, which may impair the anti-tumour activity of certain immune cell types such as natural killer and dendritic cells, promote angiogenesis and enhance accelerated growth of aggressive tumours [
26]. Future studies that explore the relationship between impaired renal function and risk for particular cancer subtypes (rather than for cancer of all types) may be able to provide a better understanding of these processes.
Our study has several strengths. The present meta-analysis represents one of the largest cohorts of individuals with diverse patient characteristics to have examined the effects of reduced kidney function and risk of cancer and cancer death. The availability of individual data allowed for an assessment of the potential influence on estimates of competing risks and reverse causality bias. There are also some potential limitations. First, we may not have had sufficient follow-up time to reliably detect a small but significant effect among those with moderate stage CKD, particularly for cancers such as colorectal, breast and prostate cancer which have a long latency period relative to the period of observation in the included studies. Second, our study was not powered to detect a statistically significant interaction between gender and the effects of reduced kidney function on cancer incidence and death, or to reliably investigate the relevance of renal function to site specific cancer risk. Third, none of the included studies considered cancer as their primary outcome, so cancer reports may not have been confirmed, for example, by pathology reports. The reliability of the cancer outcomes may also have varied between the individual studies. In general, cancer incidence and mortality data were recorded by the treating physicians who confirmed the cancer diagnoses and/or deaths. It is likely that systematic coding errors may have occurred for the different studies and resulted in over or under-estimation of the causes and/or the potential missing causes of death. Fourth, only one study (SHARP) contributed data evaluating the link between dialysis and cancer, whereas all studies contributed data for earlier stage CKD. Finally, while adjustments were made for potential confounders, residual confounding from unmeasured factors may exist.
Conclusion
In summary, this study indicates that reduced renal function is associated with an increased risk of urinary tract, digestive tract and thyroid cancers, but also with a reduced risk of prostate cancer in men. The risk is most marked among dialysis patients, but our study did not have sufficient power to exclude an increase in risk of particular cancers among patients with less severe renal impairment. Much larger studies are needed to facilitate an understanding of the association between renal function and the risk of specific cancers, and to identify possible mechanisms through which renal impairment may modulate cancer risk.
Abbreviations
ADVANCE, action in diabetes and vascular disease: preterax and diamicron mr controlled evaluation study; BMES, blue mountains eye study; CAIFOS, calcium intake fracture outcome study; CI, confidence interval; CKD, chronic kidney disease; CKD-EPI, chronic kidney disease epidemiology collaboration; eGFR, estimated glomerular filtration rate; ESKD, end stage kidney disease; HR, hazard ratio; PROGRESS, perindopril-based blood-pressure-lowering regimen study; RCTs, randomised controlled trials; SHARP, study of heart and renal protection
Acknowledgements
The authors thank the study participants in each of the individual studies for their involvement.