End-Stage Kidney Disease following Surgical Management of Kidney Cancer

Abstract

Background and objectives

We investigated the incidence of ESKD after surgical management of kidney cancer in the Australian state of Queensland, and described patterns in the initiation of kidney replacement therapy resulting from kidney cancer across Australia.

Design, setting, participants, & measurements

All newly diagnosed cases of kidney cancer in the Australian state of Queensland between January of 2009 and December of 2014 were ascertained through the Queensland Cancer Registry. There were 2739 patients included in our analysis. Patients who developed ESKD were identified using international classification of disease–10–coded hospital administrative data. Incidence rate and 3-year cumulative incidence were calculated, and multivariable Cox proportional hazards models were used to identify factors associated with ESKD. Additional descriptive analysis was undertaken of Australian population data.

Results

The incidence rate of ESKD in all patients was 4.9 (95% confidence interval [95% CI], 3.9 to 6.2) per 1000 patient-years. The 3-year cumulative incidence was 1.7%, 1.9%, and 1.0% for all patients, and patients managed with radical or partial nephrectomy, respectively. Apart from preoperative kidney disease, exposures associated with increased ESKD risk were age≥65 years (adjusted hazard ratio [aHR], 2.0; 95% CI, 1.2 to 3.2), male sex (aHR, 2.3; 95% CI, 1.3 to 4.3), preoperative diabetes (aHR, 1.8; 95% CI, 1.0 to 3.3), American Society of Anesthesiologists classification ≥3 (aHR, 4.0; 95% CI, 2.2 to 7.4), socioeconomic disadvantage (aHR, 1.6; 95% CI, 0.9 to 2.7), and postoperative length of hospitalization ≥6 days (aHR, 2.1; 95% CI, 1.4 to 3.0). Australia-wide trends indicate that the rate of kidney replacement therapy after oncologic nephrectomy doubled between 1995 and 2015, from 0.3 to 0.6 per 100,000 per year.

Conclusions

In Queensland between 2009 and 2014, one in 53 patients managed with radical nephrectomy and one in 100 patients managed with partial nephrectomy developed ESKD within 3 years of surgery.

Podcast

This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2018_09_28_CJASNPodcast_18_1_.mp3

Visual Abstract

Introduction

In the developed world, the kidney is the sixth and 13th most common site for primary malignancy in men and women, respectively (1). Surgical management with either radical (complete) or partial (nephron-sparing) nephrectomy is the mainstay of treatment. Removal of functional kidney parenchyma can increase a patient’s risk of CKD (2). International guidelines recommend partial nephrectomy when feasible, due to associations with better postoperative kidney function and equivalent oncologic control compared with radical nephrectomy (3,4). This has led to increased contemporary utilization of partial nephrectomy (5).

For patients who have limited life expectancy, iatrogenic CKD is less likely to lead to a clinically significant event; however, for patients with a favorable prognosis, there is increased risk that CKD will lead to a clinically significant end point, such as ESKD or cardiovascular events. In Australia between 2011 and 2015, 0.6% of patients who commenced kidney replacement therapy (KRT) were recorded as having developed ESKD as a consequence of kidney cancer (6), which is similar to data from the USA, where kidney cancer patients constituted approximately 0.5% of the KRT population from 1983 to 2007 (7).

Globally, the incidence of kidney cancer is rising, particularly in Western countries (8), and in Australia >3000 patients are diagnosed with kidney cancer annually (9). Incidence rates have been steadily rising since the 1980s, but mortality rates have remained relatively static (9). It is unclear whether this is because patients are receiving more effective treatment or whether a diagnosis is being made earlier in the clinical course of disease; however, considering trends of increasing incidental diagnosis of kidney cancer, the latter seems more likely (10,11). Regardless of interpretation, these data demonstrate that the number of patients diagnosed with, and subsequently managed surgically for, kidney cancer is increasing and that, on average, patients diagnosed with kidney cancer are expected to live longer. These two factors contribute to an increased lifetime risk of developing clinically significant iatrogenic CKD.

Previous population-based studies from the USA and Canada have demonstrated that ESKD incidence after surgical resection of kidney cancer is more common after radical than partial nephrectomy, and in multimorbid/elderly patients (12–15). The burden of ESKD after oncological nephrectomy has not been previously evaluated in Australian populations.

The primary aim of this study was to characterize the burden of postoperative ESKD after radical and partial nephrectomy, and to identify associations between patient and health-service characteristics and this outcome, using a large, population-based cohort of patients managed for kidney cancer in the Australian state of Queensland. The secondary aim was to evaluate long-term nationwide trends in kidney cancer-associated mortality and KRT incidence.

Materials and Methods

Study Design and Population

All patients with any type of newly diagnosed kidney cancer across 43 hospitals in the Australian state of Queensland, diagnosed between January of 2009 and December of 2014, were ascertained through the Queensland Cancer Registry, a state-wide registry that records all newly diagnosed malignancies. Reporting cancer incidence to state-based registries is a legal requirement in Australia; therefore, ascertainment of all cases is likely close to complete. There were 3799 patients identified. We excluded those who: were aged <18 years old (n=47), did not receive surgery (n=988), underwent concurrent bilateral nephrectomy (n=8), or had ESKD before surgery (n=17). The final sample included 2739 patients. Patients younger than 18 years old were excluded because it is likely that there are physiologic differences in the response to nephrectomy between children and adults (16). Patients who had ESKD before surgery were identified using international classification of disease–10 (ICD-10)–coded hospital administrative data.

Ethical Considerations

Ethics approval was obtained from the University of Queensland Human Research Ethics Committee (Approval No. 2017001010). Access to patient information without consent was granted under the Queensland Public Health Act (Approval No. RD007218). All data were de-identified before being provided to investigators. This study was conducted in accordance with the Declaration of Helsinki.

Data Collection

Data were extracted from the Queensland Cancer Registry on the 14th of September 2017, and linked to electronic hospital admission and discharge data by the Queensland Cancer Control and Analysis Team. Comorbidities before and after surgery were determined from these administrative data using a validated algorithm that incorporates ICD-10 codes (17). The output is a list of comorbidities for each patient, which are determined using the same thresholds as the Charlson comorbidity index (18). From these comorbidities, moderate-to-severe CKD and diabetes mellitus were included in the primary analysis. Sensitivity analyses were conducted that included additional comorbidities. Determination of comorbidities was limited to hospitalizations within 12 preoperative months. American Society of Anesthesiologists (ASA) classification was recorded at the time of surgery, and linked to cancer registry data from hospital records, before being provided to investigators. ASA classification was grouped as 1–2 and ≥3. Urban/rural place of residence and hospital location were assigned on the basis of the Accessibility/Remoteness Index of Australia (19). Patients were evaluated by area-level socioeconomic status, grouped by tertiles (disadvantaged, middle, advantaged) in accordance with the Australian Socioeconomic Indexes for Areas (Index of Relative Socioeconomic Disadvantage) (20). Patient-level socioeconomic status and urban/rural status were determined from the postcode of the patient’s usual place of residence recorded in the cancer registry, and provided in the data extract. Hospital volume was estimated by dividing the total number of surgical resections during the study period at each center by six, as per our previous publication (21). Postoperative length of stay was calculated as the number of days between the date of surgery and the date of discharge.

Outcome

The primary outcome was ESKD. The date of ESKD was assigned using ICD-10–coded hospital admissions data for any of: ESKD (N18.6), dependence on dialysis (Z99.2), or kidney transplant (Z94.0). Patients undergoing a second radical nephrectomy were considered to have developed ESKD on the date of their second procedure.

Descriptive Analysis

Additional data were obtained to descriptively evaluate nationwide trends in kidney cancer–associated mortality and KRT. Incidence and mortality rates for kidney cancer in Australia were obtained from the Australian Institute of Health and Welfare (projected values used for 2016) (9,22), and KRT incidence data were provided by the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA). Incidence and mortality rates were standardized to a population of 100,000 using the Australian 2001 standard population, and trends were plotted over time.

Statistical Analyses

Patients were grouped by whether they developed ESKD by the date of data extraction; they were censored on the date of data extraction, or date of death. The 3-year cumulative incidence and incidence rate per 1000 person-years were also calculated for all patients. Cumulative incidence was calculated over a 3-year period because this was the minimum follow-up time available for all patients. Time-to-event outcomes were assessed using multivariable Cox proportional hazards regressions. Models were adjusted only for potential confounders, which were identified using directed acyclic graphs. The proportional hazards assumption was checked through formal hypothesis testing of Schoenfeld residuals, and visual tests of the interaction between groups and survival times. A sensitivity analysis was undertaken that excluded all patients who underwent more than one nephrectomy during the study period. As an additional sensitivity analysis, models were assessed using a competing risk regression, considering all-cause mortality as a competing risk, using the method of Fine and Gray (23). When adjusting for health-service characteristics, clustering by hospital code was accounted for using robust sandwich estimators. Analysis was performed using Stata 14.0 (StataCorp, College Station, TX) and two-sided α was set at 0.05.

Results

Incidence of ESKD

Of 2739 patients, 68 developed ESKD by the end of the follow-up period. Patient characteristics are outlined in Table 1, and ESKD by most-likely precipitating cause is presented in Figure 1. The median (interquartile range) follow-up time was 59.4 (42.9–79.2) months. The median time from the date of surgery until the development of ESKD was 27.9 (11.8–43.4) months. The incidence rate of ESKD in all patients, and patients managed with radical and partial nephrectomy, was 4.9 (95% confidence interval [95% CI], 3.9 to 6.2), 5.2 (95% CI, 4.0 to 6.8), and 3.7 (95% CI, 2.1 to 6.5) cases per 1000 person-years. The 3-year cumulative incidence of ESKD was 1.7% (1 in 59), 1.9% (1 in 53), and 1.0% (1 in 100) for all patients, and patients managed with radical and partial nephrectomy, respectively. A Kaplan–Meier graph comparing survival estimates for radical and partial nephrectomy is presented in Figure 2, and a cumulative incidence function curve is presented in Supplemental Figure 1.

Table 1.

Characteristics of 2739 patients managed surgically for kidney cancer

Characteristics	Postoperative ESKDa
	No (n=2671)	Yes (n=68)
Patient characteristics
Age at diagnosis, yr
Median (IQR)	62 (53–70)	66 (61–74)
<65	1606 (60)	30 (44)
≥65	1065 (40)	38 (56)
Sex
Female	933 (35)	13 (19)
Male	1738 (65)	55 (81)
Preoperative kidney disease
No	2637 (99)	53 (78)
Yes	34 (1)	15 (22)
Preoperative diabetes mellitus
No	2354 (88)	54 (79)
Yes	317 (12)	14 (21)
ASA classification
1–2	1419 (53)	15 (22)
≥3	883 (33)	42 (62)
Missing	369 (14)	11 (16)
Socioeconomic statusb
Disadvantaged	606 (23)	22 (32)
Middle	1696 (63)	39 (57)
Advantaged	369 (14)	7 (10)
Place of residencec
Major city	1618 (61)	41 (60)
Inner regional	648 (24)	17 (25)
Rural	404 (15)	10 (15)
Health-service characteristicsd
Nephrectomy type
Partial	625 (23)	12 (18)
Radical	2046 (77)	56 (82)
Surgical approach
Open	807 (30)	24 (35)
Laparoscopic	1864 (70)	44 (65)
Multiple surgeriese
No	2614 (98)	59 (87)
Yes	57 (2)	9 (13)
Hospital type
Private	1239 (46)	29 (43)
Public	1432 (54)	29 (57)
Hospital locationf
Major city	2162 (81)	53 (78)
Inner regional	353 (13)	11 (16)
Outer regional	156 (6)	4 (6)
Hospital volume, resections/yrf
Median (IQR)	25 (14–56)	25 (14–41)
<20	963 (36)	26 (38)
≥20	1708 (64)	42 (62)
Postoperative length of stay
Median (IQR)	4 (3–6)	6 (4–9)
1–5 d	1891	34
≥6 d	780	34

Data presented as: count (%) or median (interquartile range). IQR, interquartile range; ASA, American Society of Anesthesiologists.

^aESKD defined as admission for hemodialysis, kidney transplant, ESKD, or a second radical nephrectomy.

^bSocioeconomic indexes for areas index of relative socioeconomic advantage and disadvantage (tertiles).

^cAccessibility/remoteness index of Australia; rural encompasses outer regional and remote categories.

^dSurgery characteristics are for the first procedure only.

^eIncludes patients who had more than one procedure, in any combination of radical or partial nephrectomy. Of the nine patients who developed ESKD, seven of these underwent two radical nephrectomies at different time points.

^fNumber of surgeries for kidney cancer at each center/yr.

Figure 1.

The predisposing factor for ESKD was uncertain for most patients. Of the 68 patients who developed ESKD after surgical management of kidney cancer, 15 of these patients had preoperative moderate-to-severe CKD, seven patients went on to have a second nephrectomy for cancer management, and for 46 patients the cause is uncertain. Of these 46 patients, 28 had an ASA score ≥3 (black), 11 had an ASA score <3 (dark gray), and ASA score was not known for the remaining seven patients (light gray). There were eight patients in this dataset who were excluded because they received a bilateral radical nephrectomy (these patients are included in this figure because they also developed ESKD).

Figure 2.

Kaplan-Meier survival estimates curve showing ESKD occurred more frequently following radical nephrectomy, although estimates did not reach conventional levels of statistical significance. Graph comparing freedom from ESKD for patients managed with radical (broken line) compared with partial (solid line) nephrectomy. Differences in the incidence of ESKD between the two procedures did not reach conventional levels of statistical significance (P=0.26).

Associations with ESKD

Of the patient characteristics evaluated in this study, age≥65 years compared with younger patients (adjusted hazard ratio [aHR], 2.0; 95% CI, 1.2 to 3.2), male sex (aHR, 2.3; 95% CI, 1.3 to 4.3), preoperative kidney disease (aHR, 15.0; 95% CI, 8.3 to 27.0), preoperative diabetes mellitus (aHR, 1.8; 95% CI, 1.0 to 3.3), and an ASA score ≥3 (aHR, 4.0; 95% CI, 2.2 to 7.4) were significantly associated with ESKD (Table 2). Patients in the lowest tertile of socioeconomic status (disadvantaged) were also at increased risk of ESKD, although this association did not reach conventional levels of statistical significance (aHR, 1.6; 95% CI, 0.9 to 2.7). The only health-service characteristic related to ESKD was length of postoperative stay ≥6 days, compared with 1–5 days (aHR, 2.1; 95% CI, 1.4 to 3.0). There were no differences when analyses were performed with a competing risk regression, considering death as a competing event (Supplemental Table 1). No differences were observed when patients who underwent more than one nephrectomy were excluded (Supplemental Table 2). Refer to Supplemental Tables 3 and 4 for sensitivity analyses evaluating additional comorbidities.

Table 2.

Risk factors for developing ESKD after surgical management of kidney cancer (n=2739)

Characteristics	Person-Years	Number of Events	Rate	Crude	Adjusted
				HR (95% CI)	HR (95% CI)
Patient characteristics
Age at diagnosis, yr
<65	8543	30	3.5	1	1
≥65	5342	38	7.1	2.0 (1.2 to 3.2)	2.0 (1.2 to 3.2)
Per 5 yr	—	—	—	1.2 (1.1 to 1.3)	1.2 (1.1 to 1.3)
P value				0.005	0.005
Sex
Female	4990	13	2.6	1	1
Male	8894	55	6.2	2.3 (1.3 to 4.2)	2.3 (1.3 to 4.3)
P value				0.006	0.006
Preoperative kidney diseasea
No	13,670	53	3.9	1	1
Yes	214	15	70.0	17.0 (9.9 to 31.0)	15.0 (8.3 to 27.0)
P value				<0.001	<0.001
Preoperative diabetes mellitusb
No	12,350	54	4.4	1	1
Yes	1535	14	9.1	2.0 (1.1 to 3.6)	1.8 (1.0 to 3.3)
P value				0.02	0.05
ASA classificationb,c
1–2	7650	15	2.0	1	1
≥3	4396	42	9.6	4.7 (2.6 to 8.5)	4.0 (2.2 to 7.4)
P value				<0.001	<0.001
Socioeconomic status
Middle/advantaged	10,724	46	4.3	1	1
Disadvantaged	3160	22	7.0	1.6 (0.9 to 2.7)	1.6 (0.9 to 2.7)
P value				0.06	0.06
Place of residence
Major city	8412	41	4.9	1	1
Inner regional	3392	17	5.0	1.0 (0.6 to 1.8)	1.0 (0.6 to 1.8)
Rural	2080	10	4.8	1.0 (0.5 to 2.0)	1.0 (0.5 to 2.0)
P value				0.90	0.99
Health-service characteristics
Nephrectomy typed
Partial	3238	12	3.7	1	1
Radical	10,657	56	5.3	1.4 (0.9 to 2.2)	1.3 (0.8 to 2.0)
P value				0.08	0.26
Surgical approache
Open	4148	24	5.8	1	1
Laparoscopic	9736	44	4.5	0.8 (0.5 to 1.3)	0.7 (0.4 to 1.4)
P value				0.32	0.37
Hospital typef
Private	6508	29	4.5	1	1
Public	7378	39	5.3	1.2 (0.8 to 1.7)	1.3 (0.9 to 1.9)
P value				0.38	0.13
Hospital location
Major city	11,265	53	4.7	1	1
Inner regional	1849	11	5.9	1.3 (0.7 to 2.1)	1.2 (0.7 to 2.0)
Outer regional	771	4	5.2	1.1 (0.7 to 1.7)	1.0 (0.6 to 1.6)
P value				0.67	0.86
Hospital volume, resections/yrg
<20	4909	26	5.3	1	1
≥20	8974	42	4.7	0.9 (0.6 to 1.4)	1.0 (0.6 to 1.6)
P value				0.59	0.93
Postoperative length of stayh
1–5 d	9866	34	3.4	1	1
≥6 d	4018	34	8.6	2.5 (1.6 to 4.0)	2.1 (1.4 to 3.0)
Per day	—	—	—	1.07 (1.03 to 1.10)	1.04 (1.01 to 1.07)
P value				<0.001	0.002

Hazard ratio (HR) and 95% confidence interval (95% CI) estimated using Cox proportional hazards model. Reported rate is per 1000 patient-yr. All adjusted estimates were adjusted for age and sex; additional adjustment variables are included as footnotes. For all health-service characteristics, clustering by hospital code was accounted for using robust sandwich estimators. —, not applicable; ASA, American Society of Anesthesiologists.

^aAdditionally adjusted for diabetes mellitus and place of residence.

^bAdditionally adjusted place of residence.

^cThere were missing data for 369 observations.

^dAdditionally adjusted for hospital location and hospital volume.

^eAdditionally adjusted for nephrectomy type, hospital location, and hospital volume.

^fAdditionally adjusted for socioeconomic status.

^gAdditionally adjusted for hospital location.

^hAdditionally adjusted for preoperative diabetes and kidney disease, ASA classification, nephrectomy type, and surgical approach.

Nationwide Trends

The age-standardized incidence rate of kidney cancer in Australia has been increasing since 1982, whereas deaths attributable to kidney cancer have remained fairly static (Figure 3A). Incidence of KRT due to kidney cancer in Australia has been steadily increasing since the first recorded case of this in 1974 (Figure 3B). Similarly, the average proportion of annual incident KRT cases that are attributable to kidney cancer has risen from 0.2% in the 1970s to 0.6% in 2015 (Figure 3C). When assessing cancer-specific deaths as a ratio of incident kidney cancer cases, there is a marked negative trend over time; whereas, when considering KRT due to kidney cancer as a ratio of incident kidney cancer cases, this ratio steadily rises over time (Figure 3D).

Figure 3.

Australia-wide trends in the incidence of kidney replacement therapy, and incidence and mortality of kidney cancer show that mortality rates for kidney cancer are dropping, but incident kidney replacement therapy (KRT) following the management of kidney cancer is increasing. (A) Age-adjusted incidence (blue) and mortality (red) rates per 100,000 population for kidney cancer in Australia. (B) Age-adjusted incidence rate of KRT attributable to kidney cancer in Australia per 100,000 population. (C) Number of incident KRT cases in Australia (black) compared with the percentage of these cases attributable to kidney cancer (green). (D) The number of deaths as a ratio of incident kidney cancer cases (red) and the number of incident KRT cases due to kidney cancer as a ratio of incident kidney cancer cases (green). Data from the Australian Institute of Health and Welfare, and the Australia and New Zealand Dialysis and Transplant Registry.

Discussion

The aim of this study was to characterize the burden of ESKD in patients managed surgically for kidney cancer. Most significantly, we found that 1.7% of patients managed surgically for kidney cancer (1.9% and 1.0% of those managed with radical and partial nephrectomy, respectively) developed ESKD within 3 years of surgery (i.e., 1 in 53 patients managed with radical nephrectomy, and 1 in 100 patients managed with partial nephrectomy).

The incidence rates reported in this study were reasonably similar to the only other report of ESKD incidence after nephrectomy conducted in the Asian-Pacific region. A recent population-based study of 3636 Taiwanese nephrectomy patients conducted by Lin et al. (24) reported ESKD incidence rates of 6.9 and 5.5 cases per 1000 person-years in patients managed with radical and partial nephrectomy. Our corresponding point-estimates were 5.2 and 3.7 cases per 1000 person-years; although these were slightly lower, this is most likely explained by the fact that Taiwan has the world’s highest incidence of ESKD (25), indicating that Taiwanese kidney cancer patients may generally be at higher risk of ESKD than Australian kidney cancer patients. Lin et al. did not find a statistically significant difference in the incidence of ESKD between the two procedures (IRR, 1.3; 95% CI, 0.8 to 2.0), which was also the case in this study. This is most likely an issue of inadequate power due to the low event count. It is also possible that the incidence of ESKD after partial nephrectomy was biased due to unmeasured confounding by indication, where patients undergoing partial nephrectomy with an absolute indication (e.g., patients with a single functional kidney, preexisting severe CKD, or bilateral tumors) were at higher risk of ESKD than those undergoing elective partial nephrectomy (26).

In an analysis similar to this study conducted in Ontario, Canada, Yap et al. (13) found that, in a cohort of 7153 patients managed for kidney cancer between 2003 and 2010, partial nephrectomy was associated with a significantly lower risk of ESKD compared with radical nephrectomy (aHR, 0.4; 95% CI, 0.3 to 0.8). This supports our inference that failure to discern a difference between the two procedures in this study was a consequence of inadequate power. A recent systematic review of 21 studies comparing radical and partial nephrectomy (total n=11,204) also suggest benefits in terms of kidney function after partial nephrectomy (27). Mir et al. (27) demonstrated that patients managed with partial nephrectomy had a significantly lower risk of developing CKD (eGFR<60 ml/min per 1.73 m²) postoperatively compared with those managed with radical nephrectomy (relative risk, 0.36; P<0.001).

In clinical practice, it is assumed that only patients with clearly perturbed preoperative kidney function or anatomy are at significant risk of ESKD after nephrectomy (3). We were able to establish that potential risk factors for incident ESKD were older age, the presence of diabetes mellitus, and socioeconomic disadvantage, which are well known risk factors for CKD in the Australian population (28); and high ASA classification and male sex, both of which are associated with CKD after nephrectomy (29,30).

With these possible risk factors in mind, preoperative and ongoing screening for at-risk nephrectomy patients may improve patient outcomes, allowing for earlier appropriate referral to a nephrologist, with subsequent implementation of a multidisciplinary approach to ongoing management decisions (28,31). Foreseeable benefits of early involvement of a nephrologist could include slower disease progression, reductions in mortality and hospitalization rates, improvements in vascular access outcomes, and a higher likelihood of transplantation, all of which have been observed in the general population (32–35).

Another finding of interest was that a longer postoperative stay was associated with ESKD. Patients with a greater number of comorbidities, older patients, patients who experience perioperative surgical complications, and patients undergoing more extensive/invasive surgery are more likely to have a longer period of postoperative hospitalization (36). Although these variables were adjusted for (with the exception of perioperative complications), residual confounding may have played a role. We have previously observed that patients from our center who have a low eGFR before nephrectomy tend to have a longer postoperative length of stay (37), which may also partially explain the present finding.

Before this study, the only population-level data on ESKD after surgical management of kidney cancers in Australia were reported by the ANZDATA registry. Compared with Queensland kidney cancer patients in this study who developed ESKD, the proportion of patients recorded by ANZDATA to have developed ESKD as a consequence of kidney cancer is considerably smaller. ANZDATA presumably underestimates the true incidence of ESKD after surgical management of kidney cancers because it only reports patients who received KRT. It is not immediately clear how many patients who develop ESKD after nephrectomy do not begin KRT. As only one cause of kidney disease can be recorded, this may also affect the incidence reported in this registry, because the cause of ESKD recorded in the registry is that which the nephrologists considered to be the most important. The most salient observations from these nationwide data were that both kidney cancer incidence and the incidence of KRT as a consequence of kidney cancer are rising; whereas, the number of kidney cancer-associated deaths has remained stable. One explanation is that more patients are being diagnosed with incidental kidney masses and undergoing surgical management, hence potentially reducing the mortality rate due to overdiagnosis bias (38), while driving up the rate of KRT due to more frequent surgical management.

Another possible explanation is that the likelihood of KRT being offered to patients developing ESKD subsequent to nephrectomy has increased over time, due to systemic changes in the provision of KRT in Australian hospitals. Between the years of 1991 and 1998, there was no difference in the KRT commencement rate for patients aged ≤64 years old; however, for the age groups 65–74 and 75–84 years old, there was an increase of approximately 200 and 100 patients per million population, respectively, over this period (39). Considering that the average age of kidney cancer diagnosis is approximately 60 years old (40), and patients aged ≥65 years old were more likely to develop ESKD after nephrectomy compared with younger patients (Table 1), greater tendency to offer dialysis to older patients over time could also partially explain the trends seen in Figure 3. Notwithstanding, the observed trend most likely reflects, at least in part, a greater number of kidney cancer cases being diagnosed and managed surgically over time.

The strengths of this study were its large size, population-based sampling strategy, and minimum follow-up time of 3 years, which allowed for a comprehensive and complete evaluation of the incidence of ESKD after nephrectomy in Queensland, Australia. Limitations included the fact that there were minimal patient-level data available, which hindered our ability to investigate causes of ESKD in significant detail. In particular, individual measures of kidney function and albuminuria were unavailable, which may have provided greater insight into populations at higher risk of ESKD. The only indicator of preoperative CKD was obtained using ICD-10–derived comorbidity data, which is likely an underestimate (41). It is also likely that data for other specific comorbidities were underestimated, particularly because these were gathered only from hospital admissions within 1 year of cancer diagnosis. Although not specific, data for ASA classification can be considered reasonably accurate because they were recorded at point-of-care, and may provide a better understanding of baseline health status compared with the comorbidities ascertained using the algorithm. Cancer staging information was also unavailable, which may have been useful during the assessment of competing risks, but is unlikely to have directly affected the risk of ESKD, except with respect to the degree of functional tissue being removed. There is a high possibility of indication bias affecting the estimates for radical compared with partial nephrectomy; because this was not a randomized, controlled trial, caution should be applied in the interpretation of these results. Patients undergoing alternative management approaches (e.g., thermal ablation) were not included in this study; however, comparisons including these patients may be of interest for future similar studies. The most significant finding of this study was that 1.7% of patients managed surgically for kidney cancer within Queensland developed ESKD within 3 years of surgery, corresponding to 1 in 53 patients managed with radical nephrectomy and 1 in 100 patients managed with partial nephrectomy. Apart from having CKD before surgery and undergoing a second nephrectomy: age≥65 years-old, male sex, ASA classification ≥3, presence of diabetes mellitus, and socioeconomic disadvantage were all associated with the development of ESKD. Nationwide data indicated that KRT commencement after surgical management for kidney cancers increased proportionally to the number of patients diagnosed with kidney cancer, whereas the ratio of deaths to incident kidney cancer cases decreased. This study has highlighted that patients undergoing nephrectomy are subjected to potential harms, which need to be carefully balanced against benefits of surgery.

Disclosures

None.