Prognostic nomograms are used increasingly in clinical trials and to guide surveillance for patients with renal cell carcinoma (RCC). An elevated erythrocyte sedimentation rate (ESR) reportedly has been associated with a poor prognosis among patients with RCC, but the ESR is not incorporated into existing nomograms. Hence, the current study was conducted to expand on prior observations pertaining to the ESR as a prognostic indicator in patients with RCC.
The authors identified 3008 patients who underwent nephrectomy for RCC between 1970 and 2002. Disease-specific survival was estimated using the Kaplan–Meier method, and its association with the ESR and other clinical and pathologic features was evaluated using Cox proportional hazards regression analysis.
A preoperative ESR was available for 1075 patients (35.7%), 501 of whom (46.6%) exhibited an elevated ESR, including 437 of 881 patients (49.2%) with clear cell RCC, 41 of 134 patients (30.6%) with papillary RCC, and 20 of 48 patients (41.7%) with chromophobe RCC. An elevated ESR was associated with adverse clinical, laboratory, and pathologic profiles for all three histologic subtypes. The risk ratios (RRs) and 95% confidence intervals (95% CIs) for death because of clear cell RCC, papillary RCC, and chromophobe RCC for patients with an elevated ESR were 3.6 (95% CI, 1.1–1.9), 3.8 (95% CI, 1.4–10.6), and 10.3 (95% CI, 1.2–89.5), respectively. The association between an elevated ESR and death from clear cell RCC persisted even after multivariate analysis (RR of 1.5; 95% CI, 1.2–2.0).
Greater than 30% of patients with renal cell carcinoma (RCC) will develop metastatic progression after they undergo attempted curative surgery.1 Therefore, prognostic classification of RCC has been used in an effort to facilitate appropriate counseling of patients and to guide decisions pertaining to surveillance2 and adjunctive therapy.3 Currently accepted prognostic factors for RCC include tumor stage, size, and nuclear grade; patient performance status; and symptomatic presentation.4 These prognostic variables have been combined to construct various predictive models that frequently are used in the clinical setting.5–9 Nevertheless, further improvement of these models is warranted, preferably using readily available clinical indices that demonstrate predictive power.
The erythrocyte sedimentation rate (ESR) is a basic laboratory test that is available widely and is easy to use. Elevation of the ESR (i.e., a faster rate of red blood cell settling in anticoagulated blood) is largely regarded as a nonspecific finding and most frequently is ascribed to some form of inflammation, including infection, sepsis, or autoimmunity.10 Malignancy, especially metastatic disease, also may elevate the ESR.11 In particular, among patients with RCC, elevation of the ESR is frequent and, occasionally, dramatically high, even with tumors localized to the kidney. In addition, the prognostic utility of ESR in patients with RCC has been suggested in multiple previous studies.12–20 In fact, a work group of the International Union Against Cancer (UICC) and the American Joint Committee on Cancer (AJCC) has recognized that the ESR has potential use as a prognostic variable in the management of RCC.4
Despite these observations, a preoperative ESR measurement is not obtained routinely during the evaluation of renal masses at many institutions, including our own. Furthermore, existing prognostic algorithms for RCC do not incorporate the ESR.5–9 One reason for this exclusion may be that the nonspecific nature of the ESR obfuscates its utility as a predictive variable. In addition, much of the reported literature on ESR in patients with RCC18, 21–25 now appears dated and often is compromised by relatively small numbers of patients and lack of stratification according to RCC histologic subtype. Hence, in the current study, our objective was expand on the information pertaining to ESR as a prognostic variable in patients with RCC.
In this report, we demonstrate that elevations in ESR are observed commonly for the three most common RCC histologic subtypes (clear cell, papillary, and chromophobe). In addition, an elevated ESR is associated with adverse clinical and pathologic features that are predictive of RCC progression and recurrence, including symptomatic presentation, anemia, hypercalcemia, and advanced tumor stage and grade. Finally, we report an increased risk of death from RCC among patients with an elevated ESR, even after multivariate adjustment.
MATERIALS AND METHODS
After approval was obtained from the Institutional Review Board, we identified 3008 patients who underwent open radical nephrectomy or nephron-sparing surgery for unilateral, sporadic RCC between 1970 and 2002 from the Mayo Clinic Nephrectomy Registry. There were 1075 patients (35.7%) with a recorded preoperative ESR, and they formed the current study cohort. There was no significant difference in disease-specific survival (DSS) between patients with and without an ESR available for analysis (P = 0.815).
Clinical and Laboratory Features
The clinical features studied included age, gender, body mass index (BMI), presence of symptoms, Eastern Cooperative Oncology Group performance status, tumor thrombus level, and type of surgery. The BMI was categorized as normal (< 25 kg/m2), overweight (≥ 25 kg/m2 but < 30 kg/m2), and obese (≥ 30 kg/m2). Patients who had palpable flank or abdominal masses, discomfort, macroscopic hematuria, acute-onset varicocele, or constitutional symptoms (including rash, sweats, weight loss, fatigue, early satiety, and anorexia) were considered symptomatic at presentation. The level of tumor thrombus was classified as 0 (limited to the renal vein), I (extending ≤ 2 cm above the renal vein), II (extending > 2 cm above the renal vein but below the hepatic veins), III (at or above the hepatic veins but below the diaphragm), or IV (above the diaphragm).
The ESR was evaluated on heparinized samples of freshly drawn whole blood that were allowed to stand for 1 hour at room temperature. The technique of ESR measurement remained the same over the entire study period, and an abnormal ESR was defined as elevation > 22 mm in male patients or > 29 mm in female patients. Other laboratory tests studied included hemoglobin, hematocrit, leukocyte count, hematuria, pyuria, and serum levels of creatinine, blood urea nitrogen, calcium, glucose, bilirubin, aspartate aminotransferase, and alkaline phosphatase. Laboratory values were categorized as low, normal, or high, as appropriate, according to relevant cut-off values (available on request).
The pathologic features studied included histologic subtype; tumor size; the 2002 primary tumor classification; perinephric fat invasion; regional lymph node involvement; distant metastases; the 2002 tumor, lymph node, metastases (TNM) stage groupings; nuclear grade; histologic coagulative tumor necrosis; and sarcomatoid differentiation. The microscopic slides from all specimens were reviewed by a urologic pathologist (J.C.C.) who had no knowledge of patient outcomes. Histologic subtype was classified according to the UICC/AJCC and Heidelberg guidelines.26, 27 Tumor size was analyzed using indicator variables for tumors ≥ 5 cm but < 7 cm, tumors ≥ 7 cm but < 10 cm, and tumors ≥ 10 cm, which reportedly contain prognostic information according to previous studies from our institution.1, 8 Nuclear grade was assigned by using standardized criteria, as described previously.28 Tumor necrosis was defined histologically as the presence of coagulative necrosis characterized by homogenous clusters and sheets of degenerating and dead cells.29 Sarcomatoid differentiation was defined as a spindle cell malignancy that had the histologic appearance of a sarcoma.30
Comparisons among the clinical, laboratory, and pathologic features were evaluated using chi-square and Fisher exact tests. DSS was estimated using the Kaplan–Meier method. The duration of follow-up was calculated from the date of surgery to the date of death or last follow-up. Associations with DSS were assessed using Cox proportional hazards regression models and were summarized using risk ratios (RRs) and 95% confidence intervals (95% CIs). Statistical analyses were performed using the SAS software package (SAS Institute, Inc., Cary, NC), and P values < 0.05 were considered statistically significant.
There were 889 patients (82.7%) with clear cell RCC, 134 patients (12.5%) with papillary RCC, 48 patients (4.5%) with chromophobe RCC, and 4 patients (0.4%) with RCC, not otherwise specified. In total, an elevated ESR was present in 501 of 1075 the patients studied (46.6%), but the prevalence of this feature differed significantly by histologic subtype. An elevated ESR was observed in 437 patients (49.2%) who had clear cell RCC compared with 41 patients (30.6%) who had papillary RCC (P < 0.001). Patients with chromophobe RCC had an intermediate rate of ESR elevation (41.7%) that did not differ statistically from the other 2 subtypes.
Clear Cell RCC
The associations of an elevated ESR with the clinical, laboratory, and pathologic features for the 889 patients with clear cell RCC are summarized in Table 1. Patients who had an elevated ESR were more likely to have symptoms at presentation (81% vs. 58%), including constitutional symptoms (54% vs. 22%) and tumor thrombus (32% vs. 12%), compared with patients who had a normal ESR. A greater proportion of patients who had an elevated ESR also had anemia (66% vs. 12%), leukocytosis (16% vs. 9%), hematuria (63% vs. 42%), pyuria (56% vs. 44%), elevated serum creatinine (23% vs. 14%), hypercalcemia (14% vs. 9%), and hyperglycemia (43% vs. 35%) compared with patients who had a normal ESR. In addition, patients who had an elevated ESR were more likely to have tumors with adverse pathologic features, including TNM Stage III or IV (52.6% vs. 19.7%), nuclear Grade 3 or 4 (63% vs. 23.9%), coagulative tumor necrosis (46.7% vs. 11.5%), and sarcomatoid differentiation (10.1% vs. 1.1%).
Table 1. Comparison of Clinical, Laboratory, and Pathologic Features by Elevated Erythrocyte Sedimentation Rate for 889 Patients with Clear Cell Renal Cell Carcinoma
At last follow-up, 646 of 889 patients with clear cell RCC had died, including 312 patients who died from their disease at a median of 1.7 years after surgery (range, 0–26 yrs). Among the 243 patients who were alive at last follow-up, the median duration of follow-up was 11.6 years (range, 0–34 yrs). The estimated DSS rates were 71.2% at 5 years after surgery (standard error [SE], 1.6%; n = 485 patients still at risk) and 61.5% at 10 years after surgery (SE, 1.8%; n = 288 patients still at risk). The univariate association of elevated ESR with death from clear cell RCC is illustrated in Figure 1. Patients who had an elevated ESR were close to 4 times more likely to die from clear cell RCC compared with patients who had a normal ESR (RR of 3.60; 95% CI, 2.82–4.59 [P < 0.001]). This difference in outcome persisted even after patients were stratified according to various pathologic features (Table 2) and after adjusting for primary tumor classification, the presence of lymph node or distant metastases, nuclear grade, and tumor necrosis (RR of 1.52; 95% CI, 1.16–2.00 [P = 0.003]).
Table 2. Associations of Elevated Erythrocyte Sedimentation Rate with Death from Clear Cell Renal Cell Carcinoma Stratified by Pathologic Features
The risk ratio represents the association of elevated erythrocyte sedimentation rate with death from renal cell carcinoma stratified by each of the pathologic features.
2002 Primary tumor classification
pT1 and pT2
pT3 and pT4
Perinephric fat invasion
Primary tumor size
Regional lymph node involvement
pNx and pN0
pN1 and pN2
TNM stage groupings
Grades 1 and 2
Grades 3 and 4
Coagulative tumor necrosis
The associations of an elevated ESR with the clinical, laboratory, and pathologic features for the 134 patients with papillary RCC are summarized in Table 3. Patients who had an elevated ESR were more likely to have constitutional symptoms at presentation (39% vs. 20%), anemia (61% vs. 8%), elevated serum creatinine (37% vs. 15%), and TNM Stage of III or IV disease (29% vs. 8%) compared with patients who had a normal ESR.
Table 3. Comparison of Clinical, Laboratory, and Pathologic Features by Elevated Erythrocyte Sedimentation Rate for 134 Patients with Papillary Renal Cell Carcinoma
At last follow-up, 76 of 134 patients with papillary RCC had died, although there were only 15 deaths from papillary RCC at a median of 2.8 years after surgery (range, 0–16 yrs). The median follow-up for the 58 patients who remained alive was 10 years (range, 0–33 yrs). The estimated DSS rates were 90.7% at 5 years after surgery (SE, 2.7%; n = 90 patients still at risk) and 87.1% at 10 years after surgery (SE, 3.3%; n = 60 patients still at risk). The univariate association of elevated ESR with death from papillary RCC is illustrated in Figure 2. Similar to patients in the clear cell RCC group, patients who had an elevated ESR were nearly 4 times more likely to die from papillary RCC compared with patients who had a normal ESR (RR of 3.84; 95% CI, 1.38 –10.63 [P = 0.010]). No multivariate adjustment was performed because there were only 15 deaths from papillary RCC reported.
Given the small number of patients with chromophobe RCC, there was minimal power to detect associations of an elevated ESR with the clinical, laboratory, and pathologic features. Nonetheless, patients with tumors of this subtype who had an elevated ESR were more likely to have constitutional symptoms at the time of presentation (50% vs. 11%; P = 0.003), anemia (55% vs. 0%; P < 0.001), Grade 3 or 4 tumors (50% vs. 14%; P = 0.007), areas of necrosis (45% vs. 4%; P < 0.001) or contained areas of sarcomatoid differentiation (25% vs. 0%; P = 0.009) compared with patients who had a normal ESR.
At last follow-up, 29 of 48 patients with chromophobe RCC had died, although there were only 6 deaths from chromophobe RCC reported at a median of 1.3 years after surgery (range, 0–7 yrs). The median duration of follow-up for the 19 patients who remained alive was 10 years (range, 0–33 yrs). The estimated DSS rates were 88.6% at 5 years after surgery (SE, 4.8%; n = 30 patients still at risk) and 85.4% at 10 years after surgery (SE, 5.6%; n = 21 patients still at risk). The univariate association of elevated ESR with death from chromophobe RCC is illustrated in Figure 3. An elevated ESR was associated with a significant increase in the risk of death for this RCC subtype as well (RR of 10.33; 95% CI, 1.19–89.53 [P = 0.034]). However, the 95% CI for the association was wide, because only 1 of 28 patients who had a normal ESR and 5 of 20 patients who had an elevated ESR died from chromophobe RCC.
Initially suggested in the 1960s as a putative prognostic marker for RCC,11, 18, 21–24 the ESR has been associated with DSS in some recent studies,12, 14, 19, 20 although not in others.25, 31–33 Further confusion pertaining to the prognostic value of ESR has resulted from a lack of histologic stratification and the relatively small numbers of patients in each series. In the current report, we examine the ESR in a large cohort of patients with RCC, with particular emphasis given to histologic subtype and other clinical features associated with RCC.
In the current series of 1075 patients with RCC, elevations in ESR were observed in 47% of patients, which is consistent with previously reported rates ranging between 23% and 50%.16, 17, 19, 20, 24, 33, 34 The ESR was elevated in nearly 50% of the patients with clear cell RCC but in only 30% of patients with papillary RCC, further supporting the biologic differences between these subtypes.35 These differences go beyond the pathologic profile of the tumors, which generally is more adverse for clear cell RCC compared with papillary RCC. It is clear from Tables 1 and 3 that patients who had clear cell RCC had a greater probability of an elevated ESR than patients with papillary RCC, even after stratification by pathologic features.
In addition, we also believe the current study is the first to report that an elevated ESR is associated with a number of clinical and laboratory features that are predictive of aggressive disease among patients with clear cell RCC. Therefore, symptoms at the time of presentation, anemia, leukocytosis, renal impairment, hypercalcemia, hyperbilirubinemia, hematuria, and pyuria all are noted more frequently in the presence of an elevated ESR.
Furthermore, the results of the current study demonstrate that an elevation in ESR is associated with virtually every adverse pathologic feature of RCC, including larger tumor size, advanced stage, higher nuclear grade, the presence of tumor necrosis or sarcomatoid differentiation, and lymph node or distant metastases. Sene et al.20 previously reported that an elevated ESR was more frequent with increasing Robson stage in patients with RCC tumors. Similarly, Hannisdal et al.17 found that the ESR was elevated in 75% of patients with metastatic disease and in 47% of patients without metastatic disease. Metastatic RCC was associated with an elevated ESR in 76% of our patients and in 83% of the patients studied by Casamassima et al.31
However, it is worth noting that an elevated ESR is not directly indicative of adverse pathology. For example, among patients who had clear cell RCC, an elevated ESR also was observed in 30% of patients with tumors that measured < 5 cm, in 33% of patients with pathologic T1 tumors, in 47% of patients without lymph node metastases, and in 44% of patients without distant metastases. In addition, on stratified analysis, an elevated ESR was found to have the greatest impact on the risk of death from RCC among patients with low-risk or intermediate-risk pathologic features, as shown in Table 3. Therefore, it is among this group of patients that a preoperative ESR may be the most useful for identifying those who are at risk of recurrence after apparently curative surgical treatment. Because novel and potentially effective adjuvant therapies are continually being developed and tested for patients who have high-risk forms of RCC, it remains important to explore prognostic indicators that may lead to refinements of the algorithms that are used to identify patients as potential candidates for future clinical trials.
It is important to also recognize that an elevated ESR is not simply a surrogate marker for adverse pathologic findings. The association between elevated ESR and death from clear cell RCC persisted even after stratifying by various pathologic features or adjusting for tumor stage, grade, and necrosis. Similarly, Ljungberg et al.14 found that ESR, TNM stage, and tumor grade together were predictive of overall survival in a cohort of 196 patients with RCC. In other studies, it was found that ESR was a predictor of overall survival or DSS in patients with RCC, even after adjusting for necrosis,19 tumor size,16 disease stage, and histologic pattern.20 Therefore, the prognostic information provided by the preoperative ESR appears to be independent of other clinical and pathologic factors.
One limitation of the current retrospective study is that only 36% of patients treated at our institution for RCC had a preoperative ESR recorded. Although this may represent a selection bias, there was no significant difference in survival reported among patients with and those without an ESR. On the basis of our data and data from previous studies, we believe that a preoperative ESR in patients with RCC should be investigated further in a prospective fashion. Accordingly, we have modified our practice to obtain an ESR routinely in patients who present with RCC, and we hope to present more comprehensive data in the future. We currently use the ESR in addition to other clinical and laboratory features to help counsel patients preoperatively. The potential future utility of the preoperative ESR, as discussed earlier, may include the selection or stratification of patients for trials of systemic therapy after suitable prospective validation.
The mechanism of elevation of the ESR remains poorly understood. Previous investigations have suggested a possible correlation with the inflammatory cytokine interleukin 6,14, 36 perhaps mediated through hepatic synthesis of fibrinogen and α-globulins.36 In addition, a low hematocrit is found frequently in patients with RCC, and this alone may cause elevation of the ESR.10 However, 29% of the patients in our current cohort who had a normal or high hematocrit had an elevated ESR, suggesting that this cannot be the only factor implicated. Currently, further study is underway in our laboratory to elucidate precisely which pathways lead to elevation of the ESR, especially in localized RCC. This may yield useful insights into the immunologic alterations that are associated particularly with clear cell RCC.
In the current study, we examined the preoperative ESR in a large cohort of patients with RCC, providing to our knowledge the first comprehensive description of the variation in ESR by histologic subtype and its association with other clinical and pathologic features of interest. These findings reinforce previous reports relating to the prognostic significance of an elevated ESR in patients with RCC. On this basis, we recommend prospective investigation of the impact on patient management of the routine preoperative assessment of ESR.