Ассоциация по борьбе с раком и острыми почечными травмами среди населения, получающего медицинскую плату за услуги.

Acute kidney injury (AKI) is a heterogeneous syndrome characterized by an abrupt decline in kidney function with negative short-term and long-term impact on quality of life, mortality, and potential need for kidney replacement therapy. 1 , 2 The epidemiology of AKI varies across settings, eg, community-acquired or hospital-acquired, and underlying comorbid conditions. 3 In onco-nephrology, AKI 4 , 5 is increasingly recognized as a treatment-related complications and factors that can influence outcomes, escalating the number and length of hospitalizations and associated costs. 4 , 6 , 7 In cancer survivors, AKI may lead to chronic kidney disease (CKD) and subsequent kidney failure, 8 , 9 which has been associated with poor survival, particularly among elderly cancer patients. 4 , 10 , 11 Given the additional burden that AKI may impose on patients with cancer and their health care expenditure, AKI awareness, prevention, and early detection are paramount for better cancer survivorship. 12

Few population-based studies have examined the long-term association between cancer and AKI. 13 , 14 A Chinese cross-sectional survey of 44 hospitals showed 14%-20% of all AKI cases were related to malignancy. 15 A Danish study using data from 1999-2006 found that 5-year cumulative incidence of AKI in patients with cancer was 27%, 13 whereas a Canadian study using data from 2007-2014 reported a 5-year cumulative incidence of 7.8%. 14 However, no population-based studies have been conducted in the United States.

For this study, we hypothesized that US patients with incident cancer have a higher long-term risk of AKI than those without cancer. We investigated this hypothesis in a cohort of Medicare beneficiaries aged 66 years or older with incident cancer and a matched control group without cancer. Because risk of AKI may depend on survival prognosis of each cancer subtype, the risk of AKI was assessed overall and by type of cancer, which may provide useful information for a multidisciplinary clinical approach to prevent and detect AKI.

Methods

Data Source and Patients

Medicare is a US federal health insurance program that provides coverage for eligible beneficiaries aged 65 years or older and younger individuals with certain disabilities or end-stage kidney disease. For the present analysis, we selected fee-for-service beneficiaries aged 66 years or older, who were continuously enrolled in Medicare Parts A and B from January 2013 to December 2014 and did not have kidney failure, using data from the Chronic Conditions Warehouse at Centers for Medicare and Medicaid Services. The STROBE flowchart is shown in Fig S1 . Incident cancer was defined by the presence of 2 outpatient or one inpatient International Classification of Diseases 9th Revision Clinical Modification (ICD-9 CM) code ( Table S1 ) for cancer during 2014 among beneficiaries without cancer history during 2013. Incident cancers included in this study are breast, prostate, kidney, renal pelvis and ureteral, bladder, ovarian, uterine, stomach, liver, gallbladder or biliary, colon, rectal, pancreatic, lung and bronchial, lymphoma, and multiple myeloma. Beneficiaries without a cancer diagnosis during 2013-2014 were selected as controls after matching 1:1 for age, sex, and race and ethnicity. Table S2 shows the baseline characteristics of the cohort by cancer site. We used the term of sex in accordance with the variable available in the Centers for Medicaer and Medicaid Services administrative data.

Study Outcomes

The main outcomes were AKI or death among beneficiaries with cancer compared with those without cancer. AKI was defined by ICD-9 CM or ICD-10 codes ( Table S1 ). The Death date was ascertained from the linkage between Medicare data and the National Death Index.

Covariates

For the matched cohort analysis, the models included age, sex, and race and ethnicity (Hispanic, non-Hispanic [NH] White, NH Black, and Other). We also controlled for comorbid conditions identified by Centers for Medicare & Medicaid Services Chronic Condition Data Warehouse, 16 including cardiovascular disease (any of acute myocardial infarction, atrial fibrillation, ischemic heart disease, and stroke/transient ischemic attack), chronic obstructive pulmonary disease, depression, hypertension, and diabetes. Because CKD codes used by the Centers for Medicare & Medicaid Services Chronic Condition Data Warehouse include several malignancies, such as renal cell carcinoma that could cause bias due to misclassification, we instead used ICD-9 codes to identify CKD ( Table S1 ).

Statistical Analysis

The observational period started at the date of cancer diagnosis for cases—or December 31, 2014, for controls—and ended at the date of AKI diagnosis, death, incident cancer diagnosis in controls, or 5 years, whichever occurred first. Controls were matched 1:1 for age, sex, and race and ethnicity with the overall cancer cohort and with each site-specific cancer population.

We estimated the unadjusted cumulative incidence function (CIF) of AKI accounting for death as a competing risk to describe the 5-year absolute risk of AKI. 17 The cumulative incidence function was chosen instead of the Kaplan-Meier method to avoid the overestimation of absolute cumulative incidence. 18 The CIF was computed using the PROC LIFETEST procedure in SAS 9.4 (SAS Institute) with 95% confidence intervals (CIs). 19

Cause-specific hazard ratios (csHRs) were computed using Cox regression models to explore the risk of AKI and death related to cancer versus no cancer. To account for the competing risk of death, Fine-Gray regression models were used, which provide subdistribution hazard ratios (sdHRs) relative to the matched controls. 20 With both approaches, we conducted both unadjusted and adjusted analyses. The csHR is the instantaneous rate of occurrence of the primary event in subjects who are currently event-free while the sdHR is interpreted as having an effect on the CIF over time. Thus, the csHRs are interpreted as an etiological association of cancer with AKI while sdHRs are interpreted as an estimate of prognostic association of cancer on CIF of AKI. The sdHR for AKI can represent the opposite direction of csHR for AKI (eg, sdHR for AKI < 1 with csHR for AKI > 1) if the hazard for death as a competing event is high. 21 The adjusted models included age, sex, race and ethnicity, and comorbid conditions. We included the matched variables in multivariable models because they are related to the outcome. 22 , 23 The csHR and sdHRs with 95% CI were computed for all cancers and by site-specific cancers ( Table S1 ).

This study was conducted to be consistent with applicable federal law and Centers for Disease Control and Prevention policy and was not subject to institutional review board approval per 45 C.F.R. part 46.102[l][2], 21 C.F.R. part 56; 42 U.S.C §241[d]; 5 U.S.C.§ 552a; 44 U.S.C. §3501 et seq.

Results

The study included 482,016 beneficiaries with cancer and the same number of matched controls. At baseline, mean age was 77.5 (standard deviation 7.6) years, 241,090 (50%) were women, and 409,339 (84.9%) beneficiaries were NH White ( Table 1 ). The most frequent incident cancer sites were prostate (109,154, 22.6%), breast (94,506, 19.6%), lung and bronchial (72,274, 15.0%), and colon (46,133, 9.6%) ( Table S2 ). Beneficiaries with cancer were more likely to have associated comorbid conditions, including CKD, than controls. The characterization of the study population by site-specific cancer is shown in Table S2 .

The median follow-up time was 4.1 years (interquartile range [IQR], 1.0-5.0 years) in the cancer group and 4.8 years (IQR, 1.7-5.0 years) in the control group. Five-year unadjusted cumulative incidence of AKI was 19.8% (95% CI, 19.7%-19.9%) in the cancer group and 19.1% (95% CI, 18.9%-19.2%) in the control group. The 5-year cumulative incidence of death was 34.7% (95% CI, 34.5%-34.8%) in the cancer group and 14.8% (95% CI, 14.7%-14.9%) in the control group. Cancer sites with the highest 5-year cumulative incidence of AKI were the renal pelvis and ureter, multiple myeloma, and kidney ( Table 2 ). Cancer sites with the highest cumulative incidence of death included pancreatic, lung and bronchial, liver, and gallbladder and biliary tract ( Fig S2 ). Occurrence of AKI tended to be higher during the first year after the diagnosis of cancer. The unadjusted and adjusted csHR for each site-specific cancer is shown in Table S3 .

In unadjusted competing risk models, the overall risk of AKI was 6% higher in beneficiaries with than without cancer (sdHR = 1.06; 95% CI, 1.05-1.07, P < 0.001); however, this association was reversed in the multivariable-adjusted model (sdHR = 0.84; 95% CI, 0.84-0.85, P < 0.001). Cumulative incidence curves of site-specific cancer are provided in Fig S2 . In adjusted models, the risk of AKI was lower versus matched controls for cancer sites with the highest mortality rates, such as pancreatic and lung and bronchial, and higher than in controls for multiple myeloma ( Fig 1 ; Table S4 ).

The sdHR of AKI was 30% higher in men than women (sdHR = 1.30; 95% CI, 1.28-1.31, P < 0.001), and 34% higher in NH Black than NH White beneficiaries (sdHR = 1.34; 95% CI, 1.32-1.36, P < 0.001) based on the subdistribution hazard model, which incorporated all covariates simultaneously ( Fig 2 ). Those with CKD, hypertension, diabetes, or cardiovascular disease had a 69% to 39% higher risk of AKI than those without these conditions.

Discussion

Among Medicare fee-for-service beneficiaries diagnosed with cancer in 2014, about 1 in 5 patients (19.8%) developed AKI over the next 5 years. Overall adjusted risk of AKI was 16% lower among those with cancer than among those without cancer, largely due to the significantly higher risk of death among beneficiaries with cancer. For some specific cancers, the risk of AKI correlated inversely with mortality. The discrepancy in the directions of csHR (>1) and sdHR (<1) observed in some cancers, such as pancreatic cancer can be attributed to higher mortality because patients with cancer were more likely to die before developing AKI. Accordingly, the adjusted risk of AKI was significantly higher in patients with multiple myeloma and was lowest among patients with pancreatic or lung and bronchial cancers, who had the highest mortality.

The overall risk of AKI increased with age, was 30% higher in men than women, and was 34% higher in NH Black than NH White beneficiaries. Although these disparities were pronounced, they have been consistently described in previous reports based on noncancer patients. 24 , 25 For example, findings from the Atherosclerosis Risk in Communities study indicated that a higher risk of AKI among NH Black adults may be explained by inadequate insurance or low income. 24 In contrast, we found that AKI risk was significantly lower in Hispanic beneficiaries and those of other race and ethnicities, compared with NH White beneficiaries. These differences may be related in part to the different distribution of comorbid conditions because those with CKD, hypertension, diabetes, or cardiovascular disease had a 69% to 39% higher risk of AKI than those without these conditions—even after accounting for the risk of death. Future studies may examine social determinants of health on risk of AKI. 26

In a systematic review of 16 studies including nearly 300,000 adult patients with malignancy, the pooled incidence of AKI was 23%, ranging from 4% to 45%. 27 This higher risk of AKI may be due to the overall younger population than the Medicare cohort and to AKI being defined by laboratory-based criteria instead of ICD codes. However, the studies included in the review were highly heterogenous and they lacked controls without cancer. A Canadian study of about 163,000 adult patients initiating chemotherapy or targeted agents for cancer found a much lower 5-year cumulative incidence of AKI, at 7.8%, likely due to AKI being defined by hospital discharge codes, suggesting a distribution toward more severe cases. Similar to the systematic review, our analysis showed multiple myeloma was associated with the highest risk for AKI. Kidney disease is highly prevalent in multiple myeloma, and AKI is often the initial presentation of this malignancy. 13 , 14 , 28 The kidney injury is commonly caused by monoclonal free light chain casts and other paraproteins with direct toxic effects on the proximal tubular cells. The injury can also be precipitated by secondary hypercalcemia, through dehydration and vasoconstriction. Although recovery from AKI improves survival in patients with multiple myeloma, their life expectancy and kidney survival is lower than among those with normal kidney function. 29 , 30 As prognosis in patients with cancer has improved due to earlier detection, targeted treatment, or adjunctive therapy, 31 comorbid conditions related to AKI, including CKD, may become more relevant to the shifting of causes of death from cancer-related to non-cancer–related deaths. 32

One strength of our study is the use of population-based data from Medicare fee-for-service beneficiaries, which allows for a comprehensive overview of AKI across representative cancers. There are some limitations to our study. First, the use of ICD-9 or ICD-10 codes to capture cancer, 33 AKI, and other comorbid conditions may be susceptible to misclassification. For example, non-differential misclassification of AKI could cause the possible associations between site-specific cancers and AKI to be biased toward the null. Second, the observational design of the study allows for unaccounted residual confounding due to variables not available in the data —such as laboratory data and social determinants of health such as income—which may play an important role in AKI risk. 26 Third, because Medicare data does not contain information on the cancer stage and histology, we could not assess the heterogeneity of susceptibility to AKI within each cancer group that may be attributed to cancer severity. Fourth, we did not consider cancer treatment including medication. Our research compared the incidence of AKI among the Medicare population diagnosed with cancer, reflecting the current distribution of cancer stages and treatments, to that among those without cancer. Finally, findings may not be generalizable to the entire older population because the study cohort is restricted to Medicare fee-for-service beneficiaries aged ≥66 years, who represent two-thirds of the Medicare population. There may be differences in quality of care between Medicare fee-for-service and Medicare Advantage, 34 which may affect the incidence of AKI through preventive care for AKI. 35 Further studies may be warranted to ascertain whether the observed associations are consistent.

In conclusion, among Medicare fee-for-service beneficiaries with newly diagnosed cancer, about 1 in 5 patients developed AKI over the subsequent 5 years. By cancer site, the risk for AKI depended largely on the corresponding survival rate; multiple myeloma had a significant association with a higher cumulative risk of AKI.

Article Information

Authors’ Full Names and Academic Degrees

Yoshihisa Miyamoto, MD, PhD, Linda J. Andes, PhD, Alain K. Koyama, ScD, Fang Xu, PhD, and Meda E. Pavkov, MD, PhD

Authors’ Contributions

Research idea, study design, and drafting: YM; Data acquisition: MP; Data analysis: LA; Critical revision: AK, FX, and MP. Each author contributed important intellectual content during article drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.

Support

The study receives no financial support.

Financial Disclosure

The authors declare that they have no relevant financial interests. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Peer Review

Received July 8, 2024. Evaluated by 1 external peer reviewer, with direct editorial input from the Statistical Editor, an Associate Editor, and the Editor-in-Chief. Accepted in revised form April 23, 2025.

Footnotes

Complete author and article information provided before references.

Supplementary File (PDF)

Figure S1: STROBE flowchart of the study.

Figure S2: Unadjusted cumulative incidence of AKI and death by site-specific cancer.

Table S1: ICD-9CM and ICM-10 Codes.

Table S2: Baseline Characteristics of Beneficiaries by Site-Specific Cancers, Medicare fee-for-Service 2014.

Table S3: Unadjusted and Adjusted Cause-Specific Hazard Ratio for Acute Kidney Injury and Death in all Cancers and by Site-Specific Cancers.

Table S4: Unadjusted and Adjusted Subdistribution Hazard Ratio for Acute Kidney Injury in all Cancers and by Site-Specific Cancers.

Supplementary Materials

Supplementary File (PDF)

Figures S1, S2; Tables S1-S4.

References

1. Kellum J.A., Romagnani P., Ashuntantang G., Ronco C., Zarbock A., Anders H.J. Acute kidney injury. Nat Rev Dis Primers. 2021;7(1):52. doi: 10.1038/s41572-021-00284-z.

2. Ronco C., Bellomo R., Kellum J.A. Acute kidney injury. Lancet. 2019;394(10212):1949–1964. doi: 10.1016/S0140-6736(19)32563-2.

3. Hoste E.A.J., Kellum J.A., Selby N.M., et al. Global epidemiology and outcomes of acute kidney injury. Nat Rev Nephrol. 2018;14(10):607–625. doi: 10.1038/s41581-018-0052-0.

4. Campbell G.A., Hu D., Okusa M.D. Acute kidney injury in the cancer patient. Adv Chronic Kidney Dis. 2014;21(1):64–71. doi: 10.1053/j.ackd.2013.08.002.

5. Gallieni M., Cosmai L., Porta C. Acute kidney injury in cancer patients. Contrib Nephrol. 2018;193:137–148. doi: 10.1159/000484970.

6. Rosner M.H., Perazella M.A. Acute kidney injury in patients with cancer. N Engl J Med. 2017;376(18):1770–1781. doi: 10.1056/NEJMra1613984.

7. Braet P., Sarto G.V.R., Pirovano M., Sprangers B., Cosmai L. Treatment of acute kidney injury in cancer patients. Clin Kidney J. 2022;15(5):873–884. doi: 10.1093/ckj/sfab292.

8. Kim C.S., Kim B., Suh S.H., et al. Risk of kidney failure in patients with cancer: a South Korean population-based cohort study. Am J Kidney Dis. 2022;79(4):507–517.e1. doi: 10.1053/j.ajkd.2021.06.024.

9. Mansouri I., Alencar de Pinho N., Snanoudj R., et al. Trends and outcomes with kidney failure from antineoplastic treatments and urinary tract cancer in France. Clin J Am Soc Nephrol. 2020;15(4):484–492. doi: 10.2215/CJN.10230819.

10. Lee M., Wang Q., Wanchoo R., Eswarappa M., Deshpande P., Sise M.E. Chronic kidney disease in cancer survivors. Adv Chronic Kidney Dis. 2021;28(5):469–476.e1. doi: 10.1053/j.ackd.2021.10.007.

11. Silver S.A., Long J., Zheng Y., Chertow G.M. Cost of acute kidney injury in hospitalized patients. J Hosp Med. 2017;12(2):70–76. doi: 10.12788/jhm.2683.

12. The Kidney Disease Improving Global Outcomes (KDIGO) Summary of Recommendation Statements. Kidney Int Suppl. 2011;3(1):5–14. doi: 10.1038/kisup.2012.77.

13. Christiansen C.F., Johansen M.B., Langeberg W.J., Fryzek J.P., Sorensen H.T. Incidence of acute kidney injury in cancer patients: a Danish population-based cohort study. Eur J Intern Med. 2011;22(4):399–406. doi: 10.1016/j.ejim.2011.05.005.

14. Kitchlu A., McArthur E., Amir E., et al. Acute kidney injury in patients receiving systemic treatment for cancer: a population-based cohort study. JNCI. 2019;111(7):727–736. doi: 10.1093/jnci/djy167.

15. Jin J., Wang Y., Shen Q., Gong J., Zhao L., He Q. Acute kidney injury in cancer patients: A nationwide survey in China. Sci Rep. 2019;9(1):3540. doi: 10.1038/s41598-019-39735-9.

16. The Centers for Medicare & Medicaid Services. Chronic Conditions. Vol 2024.

17. Aalen O.O., Johansen S. An Empirical Transition Matrix for Non-Homogeneous Markov Chains Based on Censored Observations. Scandinavian Journal of Statistics. 1978;5(3):141–150.

18. Schuster N.A., Hoogendijk E.O., Kok A.A.L., Twisk J.W.R., Heymans M.W. Ignoring competing events in the analysis of survival data may lead to biased results: a nonmathematical illustration of competing risk analysis. J Clin Epidemiol. 2020;122:42–48. doi: 10.1016/j.jclinepi.2020.03.004.

19. SAS. SAS/STAT® 14.2 User’s Guide The LIFETEST Procedure. Vol 2024.

20. Fine J.P., Gray R.J. A Proportional Hazards Model for the Subdistribution of a Competing Risk. J Am Statistical Assoc. 1999;94:446–509.

21. Lau B., Cole S.R., Gange S.J. Competing risk regression models for epidemiologic data. Am J Epidemiol. 2009;170(2):244–256. doi: 10.1093/aje/kwp107.

22. Mansournia M.A., Hernan M.A., Greenland S. Matched designs and causal diagrams. Int J Epidemiol. 2013;42(3):860–869. doi: 10.1093/ije/dyt083.

23. Sjolander A., Greenland S. Ignoring the matching variables in cohort studies - when is it valid and why? Stat Med. 2013;32(27):4696–4708. doi: 10.1002/sim.5879.

24. Grams M.E., Matsushita K., Sang Y., et al. Explaining the racial difference in AKI incidence. J Am Soc Nephrol. 2014;25(8):1834–1841. doi: 10.1681/ASN.2013080867.

25. Collins A.J., Foley R.N., Herzog C., et al. US Renal Data System 2012 Annual Data Report. Am J Kidney Dis. 2013;61(1 Suppl 1):e1–e476. doi: 10.1053/j.ajkd.2012.11.031.

26. Addison D., Branch M., Baik A.H., et al. Equity in Cardio-Oncology Care and Research: A Scientific Statement From the American Heart Association. Circulation. 2023;148:297–308. doi: 10.1161/CIR.0000000000001158.

27. Can W., Rong L., Lixia L. Incidence and risk factors of acute kidney injury in patients with malignant tumors: a systematic review and meta-analysis. BMC Cancer. 2023;23(1):1123. doi: 10.1186/s12885-023-11561-3.

28. Gaut J.P., Liapis H. Acute kidney injury pathology and pathophysiology: a retrospective review. Clin Kidney J. 2021;14(2):526–536. doi: 10.1093/ckj/sfaa142.

29. Bridoux F., Leung N., Belmouaz M., et al. Management of acute kidney injury in symptomatic multiple myeloma. Kidney Int. 2021;99(3):570–580. doi: 10.1016/j.kint.2020.11.010.

30. Binder M., Nandakumar B., Rajkumar S.V., et al. Mortality trends in multiple myeloma after the introduction of novel therapies in the United States. Leukemia. 2022;36(3):801–808. doi: 10.1038/s41375-021-01453-5.

31. Jemal A., Ward E.M., Johnson C.J., et al. Annual report to the nation on the status of cancer, 1975-2014, featuring survival. J Natl Cancer Inst. 2017;109 doi: 10.1093/jnci/djx030.

32. Zaorsky N.G., Churilla T.M., Egleston B.L., et al. Causes of death among cancer patients. Ann Oncol. 2017;28(2):400–407. doi: 10.1093/annonc/mdw604.

33. Cooper G.S., Yuan Z., Stange K.C., Dennis L.K., Amini S.B., Rimm A.A. The sensitivity of Medicare claims data for case ascertainment of six common cancers. Med Care. 1999;37(5):436–444. doi: 10.1097/00005650-199905000-00003.

34. Cohen K., Ameli O., Chaisson C.E., et al. Comparison of care quality metrics in 2-sided risk medicare advantage vs fee-for-service medicare programs. JAMA Netw Open. 2022;5(12) doi: 10.1001/jamanetworkopen.2022.46064.

35. Silver S.A., Nadim M.K., O'Donoghue D.J., et al. Community health care quality standards to prevent acute kidney injury and its consequences. Am J Med. 2020;133(5):552–560.e553. doi: 10.1016/j.amjmed.2019.10.038.

Associated Data

Supplementary Materials

Supplementary File (PDF)

Figures S1, S2; Tables S1-S4.