DISCUSSION
For cancer patients, navigating the COVID19 pandemic has brought unique challenges. Balancing the risk of contracting COVID19 against the risk of delaying oncologic care has thus far been informed by the natural history of malignancies, while the natural history of COVID19 in cancer patients remains an area of active research. Prior studies have attempted to define clinical outcomes in cancer patients with COVID19 (7, 8, 10-12, 21). One of the first such studies evaluated 1590 patients from China with COVID19, of which 18 patients had a history of malignancy, and found that a history of malignancy was associated with poor clinical outcomes(8). Another population based study from Spain evaluated 4,035 consecutively hospitalized patients with COVID19 and found that presence of an active malignancy (359 patients) was an independent predictor of mortality(7). A study of 5,366 COVID19 patients in New York City early in the pandemic found that a history of cancer (334 patients) conferred a greater risk of intubation, though not a greater risk of death(9).
Studies from cancer centers have evaluated the natural history of COVID19 in larger cohorts of cancer patients. A prospective observational study of 800 cancer patients in the United Kingdom with COVID19 reported a 28% mortality rate. Interestingly, the risk of mortality was independent of cancer type(11). However, in a cohort of 423 patients treated at a cancer center New York City, the mortality rate was notably lower at only 12%(12). Indeed, a multi-institutional study of 928 patients with COVID19 and cancer throughout the US and Canada found a mortality rate of 13%(10). Malignancy type was not a significant predictor of mortality. Notably, none of these studies compared outcomes to those of patients without cancer. One meta-analysis which included 46,499 COVID19 patients, including 1776 of whom had cancer, found that malignancy is associated with an increased risk of death and intubation(21). However, in a subgroup analysis of patients > 65, there was no increased risk of mortality from COVID19 amongst cancer patients.
Though our understanding of COVID19 and cancer has continued to evolve, little is known about the impact of a history of cancer on the risk of developing AKI in COVID19 patients. Indeed, to our knowledge, no such studies have evaluated AKI as an outcome. Furthermore, many of the existing studies had relatively small samples of cancer patients or did not have a non-cancer patient control. To this end, in the present study, we have a conducted a retrospective analysis of a large cohort of patients hospitalized with COVID19 in the New York City area to evaluate the impact of malignancy on patient outcomes, including AKI, with a focus on patients with genitourinary malignancies.
Notably, our cohort represents approximately 4% of positive cases in New York City and 7% of deaths in New York City as of August 10th, 2020(22). Approximately 7% of all hospitalized COVID19 patients in our cohort had a diagnosis AKI, an incidence concordant with the 4.5%-8.9% range reported in the literature(13, 14). Notably, there was no increased risk of AKI amongst patients with a general history of malignancy, or amongst patients with a history of genitourinary malignancy. This key finding informs us that patients with a history of cancer are not more likely to develop AKI when hospitalized with COVID19. AKI is of particular concern for cancer patients as AKI is associated with the subsequent development of CKD, and CKD is associated with adverse post-operative outcomes in patients with prostate, kidney, and bladder cancer(17-19). Furthermore, CKD may reduce the oncologist’s clinical armamentarium as several chemotherapeutic agents are nephrotoxic and optimal surveillance imaging often requires iodinated contrast enhancement(16).
Regarding mortality, overall, approximately 26% of patients who were hospitalized died, and amongst hospitalized cancer patients, approximately 34% of patients died. History of malignancy conferred a statistically significant 33% increased relative risk of mortality among hospitalized patients. However, upon analysis of malignancy sub-types, we found that this increased mortality risk appears to be driven by patients with pulmonary neoplasms. Indeed, patients with lung tumors had a greater than two-fold risk of death. On the contrary, patients with prostate, bladder, kidney, or general genitourinary malignancies were not at greater risk of mortality compared to the patients without cancer. Thus, though patients with malignancy may be at higher risk for mortality, this heightened risk does not appear to apply to patients with genitourinary malignancies. Interestingly, though patients with lung cancer and cancer in general had higher mortality rates, they did not have higher rates of ICU admission. This may be related to the shortage of ICU beds during the local peaks of the pandemic(23).
In addition to evaluating outcomes amongst hospitalized patients, we evaluated risk of hospitalization amongst all SARS-CoV-2 positive patients. Overall, we found no increased risk of hospitalization amongst patients with a history of cancer compared to those without a cancer diagnosis. Interestingly, we found an overall lower risk of hospitalization amongst patients with a history of genitourinary malignancy compared to patients without cancers. Indeed, on subgroup analysis this lower risk of hospitalization appears to be driven by patients with a history of kidney cancer. The explanation for this trend is unclear. However, one possibility is that the renin-angiotensin-system (RAS) is altered in patients with renal cancer. Indeed, SARS-CoV-2 binds via the angiotensin-converting enzyme 2 receptor(24) and previous research has shown that patients with renal cancer have differential expressions of RAS enzymes(25). Ultimately, further research is required to understand this pattern.
Our study has several notable limitations. We conducted a retrospective analysis of an existing database rather than a prospective study. Furthermore, given the nature of the database, we did not have data regarding tumor stage and grade and accordingly could not control for these variables. For example, in defining a patient as having a history of “lung cancer,” we included all patients with a diagnosis code indicative of a lung neoplasm, including secondary and unspecified lung neoplasms. Given that the lung is a common site of metastatic spread, our lung cancer cohort may have included patients with more advanced malignancies which may in turn be contributing to the higher mortality rate observed. Indeed, patients also were not separated by those with active malignancy and those on surveillance. Additionally, amongst the individual genitourinary malignancies analyzed (prostate, bladder, kidney), our sample sizes were small. Despite these limitations, we believe our study provides an important adjunct to the existing literature.