Radical versus partial nephrectomy

Effect on overall and noncancer mortality

Authors


Abstract

BACKGROUND:

Relative to radical nephrectomy (RN), partial nephrectomy (PN) performed for renal cell carcinoma (RCC) may protect from non-cancer-related deaths. The authors tested this hypothesis in a cohort of PN and RN patients.

METHODS:

The Surveillance, Epidemiology, and End Results-9 database allowed identification of 2198 PN (22.4%) and 7611 RN (77.6%) patients treated for T1aN0M0 RCC between 1988 and 2004. Analyses matched for age, year of surgery, tumor size, and Fuhrman grade addressed the effect of nephrectomy type (RN vs PN) on overall mortality (Cox regression models) and on non-cancer-related mortality (competing-risks regression models).

RESULTS:

Relative to PN, RN was associated with 1.23-fold (P = .001) increased overall mortality rate, which translated into a 4.9% and 3.1% absolute increase in mortality at 5 and 10 years after surgery, respectively. Similarly, non-cancer-related death rate was significantly higher after RN in competing-risks regression models (P < .001), which translated into a 4.6% and 4.5% absolute increase in non-cancer-related mortality at 5 and 10 years after surgery, respectively.

CONCLUSIONS:

Relative to PN, RN predisposes to an increase in overall mortality and non-cancer-related death rate in patients with T1a RCC. In consequence, PN should be attempted whenever technically feasible. Selective referrals should be considered if PN expertise is unavailable Cancer 2009. © 2009 American Cancer Society.

The standard of care for patients with T1a renal cell carcinoma (RCC) consists of either partial (PN) or radical nephrectomy (RN).1-3 For 2 decades, Novick et al and several other investigators have proven equivalent cancer control outcomes when either PN or RN is used to treat T1a lesions.2, 4-6 Relative to RN, PN preserves renal parenchyma and protects from renal function deterioration that may occur because of cancer-unrelated causes.7, 8 Lau et al (n = 1681) and McKiernan et al (n = 290) have confirmed the renal function preservation benefit of PN versus RN in 2 independent patient cohorts.9, 10

The benefits of PN may ascend well beyond renal function preservation. Relative to RN, PN may protect from a variety of systemic morbidities, such as cardiovascular diseases, anemia, malnutrition, and neuropathy. All of these may reduce quality of life and predispose to higher mortality.8 Recently, Thompson et al confirmed the beneficial effect of PN versus RN on overall mortality. The authors demonstrated an increase in overall survival when PN was performed instead of RN in a single-institution sample of 327 T1a RCC patients aged younger than 65 years.11

In the current article, we attempted to corroborate this finding in a large population-based cohort of patients with T1a RCC of all ages, who were treated with either PN or RN in 1 of 9 Surveillance, Epidemiology, and End Results (SEER) registries across the United States.

MATERIALS AND METHODS

Patients diagnosed with primary invasive kidney cancer between 1988 and 2004 were identified within 9 SEER cancer registries.12 The registries include the Atlanta, Detroit, San Francisco-Oakland, and Seattle-Puget Sound metropolitan areas, as well as the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah. Characteristics of the SEER population are comparable to the general US population.12 Two kidney cancer diagnostic codes (International Classification of Diseases for Oncology [ICD-O], second edition C64.9 code and ninth revision 189.0 code) were used as inclusion criteria. Presence of both diagnostic codes resulted in the identification of 43,143 RCC patients. This did not include upper tract transitional cell carcinoma or ureteral, noncortical renal tumors (ie, melanomas, sarcomas, and lymphomas). Of those, 9809 represented RN (n = 7611, 77.6%) or PN (n = 2198, 22.4%) for localized renal masses up to 4.0 cm in size. Of the 32,316 excluded patients, 19,571 had either locally advanced (T3-4N0-2M0, 7794 patients), metastatic (TanyNanyM1, 8808), or unknown stage RCC (n = 2969). In addition 12,745 had localized RCC that exceeded the 4-cm size threshold. Moreover, 148 excluded patients were treated with local tumor destruction, and 832 were treated nonsurgically. Finally, 38 patients aged younger than 18 years were excluded.

For all 9809 patients treated with either PN or RN, malignant histology was confirmed with the ICD-O-3 SEER histologic codes. The cause of death was defined according to the SEER specific cause of death (code 29020). Patients who did not die of RCC were considered to have died of other causes.

Statistical Analyses

We tested the effect of nephrectomy type (RN vs PN) on 2 mortality endpoints: overall mortality and mortality from non-cancer-related causes (non-cancer-related mortality). In the first part of the analyses, Cox regression models targeted overall mortality. Because the selection of PN or RN may largely depend on tumor size and patient age, we matched PN and RN cases for these 2 characteristics. Each patient treated with PN was matched with up to 4 RN patients. Assignment of up to 4 controls maximizes the power of matched comparisons.13 Caliper matching by decade was used for age and year of surgery. Matching was performed for year of surgery, because better treatment outcomes may be expected in more contemporary patients. Caliper matching, within 1 cm, was also used for tumor size. Finally, a subset analysis was performed in patients with available Fuhrman grade. Here the matching criteria were complemented with Fuhrman grade. Kaplan-Meier plots were used to graphically explore the observed overall mortality rate differences according to nephrectomy type.

In the second part of the analyses, competing-risks regression models targeted non-cancer-related mortality, after adjusting for RCC-specific mortality. Similarly to Cox regression models, the matching in competing-risks regression models was performed for age, tumor size, and year of surgery. A matched subset analysis was repeated in patients with available Fuhrman grade. Cumulative incidence plots were used to graphically explore the differences in non-cancer-related mortality rates according to nephrectomy type.

All statistical tests were performed using S-PLUS Professional, version 1 (Mathsoft, Seattle, Wash) or Statistical Package for Social Science, version 15.0 (SPSS, Chicago, Ill). Moreover, all tests were 2-sided with a significance level set at .05.

RESULTS

Of 9809 patients with T1aN0M0 RCC treated with either PN (n = 2198, 22.4%) or RN (n = 7611, 77.6%), 2153 PN and 5616 RN patients were respectively matched for age, tumor size, and year of surgery. Clinical and pathological characteristics are listed in Table 1. The median follow-up was 35 and 46 months for PN and RN patients, respectively. Fuhrman grade was available in 1342 PN (62.3%) and 3150 RN (56.0%) patients. Overall, 1283 PN and 3166 PN and RN patients were matched for Fuhrman grade in addition to age, tumor size, and year of surgery.

Table 1. Clinical Characteristics of the Population Matched for Age, Tumor Size, and Year of Surgery (n = 7769) and of the Population Matched for Age, Tumor Size, Year of Surgery, and Fuhrman Grade (n = 4449)
 Population Matched for Age, Tumor Size, and Year of Surgery (Unmatched for Fuhrman Grade, n = 7769)Population Matched for Age, Tumor Size, Year of Surgery, and Fuhrman Grade (n = 4449)
Partial Nephrectomy No. (%)Radical Nephrectomy No. (%)Partial Nephrectomy No. (%)Radical Nephrectomy No. (%)
  1. Data are stratified according to nephrectomy type (partial nephrectomy vs radical nephrectomy).

Total2153 (27.7)5616 (72.3)1283 (28.8)3166 (71.2)
Age, y, mean59.861.159.661.3
Male sex1331 (61.8)3335 (59.4)798 (62.2)1844 (58.2)
Tumor size, cm, mean2.42.72.52.8
Pathologic subtype    
 Clear cell1719 (79.8)4749 (84.6)1047 (81.6)2699 (85.2)
 Papillary199 (9.2)295 (5.3)104 (8.1)152 (4.8)
 Other235 (10.9)572 (10.2)132 (10.3)315 (9.9)
Fuhrman grade    
 1385 (17.9)929 (16.5)352 (27.4)917 (29.0)
 2752 (34.9)1733 (30.9)735 (57.3)1805 (57.0)
 3185 (8.6)441 (7.9)180 (14.0)412 (13.0)
 420 (0.9)47 (0.8)16 (1.2)32 (1.0)
 Unknown811 (37.7)2466 (43.9)00
Year of surgery    
 1988-199088 (4.1)312 (5.6)32 (2.5)105 (3.3)
 1991-2000858 (39.9)2935 (52.3)521 (40.6)1643 (51.9)
 2001-20041207 (56.1)2369 (42.2)730 (56.9)1418 (44.8)

Effect of PN Versus RN on Overall Survival

In Cox regression models performed on 7769 patients matched for age, tumor size, and year of surgery, RN was associated with a 1.23-fold (P = .001) higher mortality rate than PN. For example, at 5 and 10 years of follow-up, the survival rate of PN versus RN patients was, respectively, 89.3% versus 84.4% and 71.3% versus 68.2%, for an absolute difference of 4.9% and 3.1% in favor of PN (Fig. 1).

Figure 1.

Kaplan-Meier survival curves depicting the overall survival of the cohort matched for age, tumor size, and year of surgery (n = 7769), stratified according to treatment type (partial vs radical nephrectomy).

In Cox regression models performed on 4449 patients matched for age, tumor size, year of surgery, and Fuhrman grade, RN was associated with a 1.19-fold (P = .048) higher mortality rate than PN. For example, at 5 and 10 years of follow-up, the survival rate of PN versus RN patients was, respectively, 88.9% versus 85.5% and 70.9% versus 68.8%, for an absolute difference of 3.4% and 2.1% in favor of PN (Fig. 2).

Figure 2.

Kaplan-Meier survival curves depicting the overall survival of the cohort matched for age, tumor size, year of surgery, and Fuhrman grade (n = 4449), stratified according to treatment type (partial vs radical nephrectomy).

Effect of PN Versus RN on Noncancer-related Mortality

Competing-risks regression models were applied to 7769 patients matched for age, tumor size, and year of surgery. RN was associated with a statistically significantly higher rate of non-cancer-related mortality than PN, after adjustment for RCC-specific mortality (P < .001). For example, at 5 and 10 years of follow-up, the non-cancer-related mortality rate of PN versus RN patients was, respectively, 11.7 versus 16.3% and 27.1% and 31.6%, for an absolute difference of 4.6% and 4.5% in favor of PN (Fig. 3).

Figure 3.

Cumulative incidence plots depicting the noncancer-related mortality and renal cell carcinoma-specific mortality in the cohort matched for age, tumor size, and year of surgery (n = 7769), stratified according to treatment type (partial vs radical nephrectomy).

In competing-risks regression models performed on 4449 patients matched for age, tumor size, year of surgery, and Fuhrman grade, RN was also associated with a statistically significantly higher rate of non-cancer-related mortality than PN, after adjustment for RCC-specific mortality (P = .02). For example, at 5 and 10 years of follow-up, the non-cancer-related mortality rate of PN versus RN patients was 12.4% versus 15.6% and 27.1% versus 30.6%, for an absolute difference of 3.2% and 3.5% in favor of PN (Fig. 4).

Figure 4.

Cumulative incidence plots depicting the noncancer-related mortality and renal cell carcinoma-specific mortality in the cohort matched for age, tumor size, year of surgery, and Fuhrman grade (n = 4449), stratified according to treatment type (partial vs radical nephrectomy).

DISCUSSION

Current literature confirms the equivalence of cancer control outcomes of PN and RN for T1a RCC.14-18 Moreover, contemporary data indicate that PN may protect from renal function loss associated with the removal of the full renal unit, as is performed at RN.7, 8 Finally, PN appears to protect from the deleterious effects of renal insufficiency associated with RN, which may translate into increased mortality from non-cancer-related causes.7, 8 However this provocative finding has only been reported from a single institution and has not been corroborated with other data.

The preliminary nature of the data addressing the protective effect of PN on overall survival prompted us to revisit this topic in a large population-based cohort of PN and RN patients, from 9 SEER registries. Our analyses relied on 7769 patients matched for age, tumor size, and year of surgery and showed a 23% increase (hazard ration, 1.23; P = .001) in overall mortality rate when RN was performed instead of PN. A similar elevation in overall mortality rate (hazard ratio, 1.19; P = .048) was recorded when PN and RN patients were matched for Fuhrman grade. At 5 and 10 years after surgery, the absolute survival benefit of PN versus RN was 4.9% and 3.1%, respectively.

Because RCC may contribute to all-cause mortality, we repeated all analyses using competing-risks regression models, where non-cancer-related mortality represented the endpoint of interest and RCC-specific mortality was controlled for. Again matching was done for age, tumor size, and year of surgery. As in Cox regression models, RN was associated with statistically significantly higher non-cancer-related mortality rate than PN, with (P = .02) or without (P < .001) adjustment (matching) for Fuhrman grade. Again, as in overall mortality analyses, at 5 and 10 years after surgery, the absolute survival benefit of PN versus RN was 4.6% and 4.5%, respectively.

Taken together, these data indicate that the use of RN RCC may predispose to an excess in overall mortality and non-cancer-related mortality. In consequence, PN should be used instead of RN, as long as this procedure is technically feasible. RN should only be performed for centrally located lesions where a PN is truly not applicable.19, 20 Selective referral to surgeons with PN expertise should be considered for challenging cases.19

The magnitude of the survival disadvantage of RN is comparable to the results of Thompson et al, where at 5 and 10 years a 3% and 5% overall survival difference was recorded, respectively.11 However, Thompson et al only demonstrated the survival benefit in patients aged <65 years. Our findings are not restricted to younger individuals. Instead, our results apply to all examined age strata. Besides Thompson et al's report, no previous investigators examined the effect of PN on overall mortality and/or non-cancer-related mortality. In consequence, the current data cannot be compared with any other reported series. Nonetheless, the similarity of our findings to those of Thompson et al gives credence to both reports and highlights the robustness and the generalizability of our and Thompson et al's findings.

Despite the attractiveness and the novelty of the current data, their validity requires confirmation from other large-scale databases or ideally from randomized control trials assessing the effect of PN versus RN on mortality. Additional limitations stem from the data type used in the current analysis. First, the SEER registries are restricted to patients from the United States. It is possible that PN is used more frequently in Europe, which may dissipate the overall and non-cancer-related mortality rate differences. Second, comorbidity data were not available in this version of the SEER database. It is possible that RN patients had more comorbidities than their PN counterparts, which may have predisposed them to higher rates of non-cancer-related mortality. Comorbidity is proportional to age. Each decade increase in age has the same effect on mortality as 1 unit increase in the Charlson comorbidity score.21 Therefore, age is closely and directly related to comorbidities, and adjusting or matching for age is similar to matching for comorbidities. Nonetheless, a confirmation of these matched analyses would be ideal in a population where comorbidities are available. Third, we could not adjust for open versus laparoscopic PN and for baseline glomerular filtration rate. Hopefully, future analyses will address these limitations. Despite these potential concerns, our findings are unique in confirming a survival benefit of PN across all ages and Fuhrman grades of T1a RCC patients. We are confident that our findings will promote broader use of PN.

CONCLUSIONS

Relative to PN, RN predisposes to an increase in overall mortality and non-cancer-related death rate in patients with T1a RCC. In consequence, PN should be attempted whenever technically feasible. Selective referrals should be considered if PN expertise is unavailable.

Conflict of Interest Disclosures

Pierre I. Karakiewicz is partly supported by the University of Montreal Health Center Urology Associates, Fonds de la Recherche en Santé du Quebec, the University of Montreal Department Of Surgery, and the University of Montreal Health Center (CHUM) Foundation.

Laurent Zini is partly supported by the Association Française de Recherche sur le Cancer, the Fondation de France-Fédération Nationale des Centres de Lutte Contre le Cancer, the Association Française d'Urologie, and the Ministère Français des Affaires Etrangères et Européennes (Bourse Lavoisier).

Ancillary