doi:10.1111/codi.12665
Original article
Impact of chronic kidney disease on postoperative outcome following colorectal cancer surgery A. Currie*, G. Malietzis*, A. Askari*, S. Nachiappan*, P. Swift†, J. T. Jenkins*‡, O. D. Faiz*‡ and R. H. Kennedy*‡ *Surgical Epidemiology, Trials and Outcomes Centre, St Mark’s Hospital, London, UK, †South West Thames Kidney and Transplantation Unit, St Helier Hospital, Epsom and St Helier Hospital NHS Trust, Surrey, UK and ‡Department of Surgery and Cancer, Imperial College London, London, UK Received 28 January 2014; accepted 1 April 2014; Accepted Article online 8 October 2013
Abstract Aim Chronic kidney disease (CKD) is increasing in prevalence and is associated with cardiovascular events and mortality in asymptomatic and vascular surgery populations. This study aimed to determine the role of CKD in stratifying peri- and postoperative risk for colorectal cancer (CRC) patients with nonmetastatic disease undergoing elective curative resection. Method Patients diagnosed with nonmetastatic colorectal adenocarcinoma and undergoing surgical resection between 2006 and 2011 were identified from a prospectively collated database. Further information on survival and cause of death was gathered from a regional cancer registry. Estimated glomerular filtration rates were calculated using the Modification of Diet in Renal Disease (MDRD) equation. Kaplan–Meier survival curves were constructed for disease-free and overall survival. Multivariate Cox regression models were used to determine the role of CKD after stratification by several clinicopathological factors. Results Seven-hundred and eight colorectal resections were studied [median follow up: 45 (interquartile range, 21–65) months). Overall postoperative complications were similar, but patients with CKD were more likely to develop cardiovascular morbidity (P < 0.001) and 30-day mortality [4.8% (six of 124) in the CKD group vs 2.1% (12/580) in the non-CKD group]. Kaplan–Meier analysis revealed poorer overall survival for localized (Stage I–II; P = 0.019) and Stage III (P = 0.001) CRC in the CKD population. Multivariate Cox regression analysis identified CKD as an independent prognostic factor for noncancer death [hazard ratio (HR) = 1.82 (95% CI: 1.07–3.10); P = 0.027] but not for overall survival [HR = 1.21 (95% CI: 0.90–1.47); P = 0.116].
Correspondence to: Prof. Robin Kennedy, Consultant Colorectal Surgeon, St. Mark’s Hospital, Watford Road, Harrow, Middlesex HA1 3UJ, UK. E-mail: [email protected]
Conclusion Patients with CKD may be more likely to develop cardiovascular complications following CRC resection and have an increased risk of a noncancer death. Future research should explore the interaction of CKD in competing mortality risks following CRC surgery. Keywords Colorectal cancer, chronic kidney disease, complications, survival What does this paper add to the literature? This paper is the first to compare the postoperative outcomes, long-term survival and cause of death between chronic kidney disease and normal renal function patients following elective colorectal cancer surgery.
Introduction Chronic kidney disease (CKD) affects 5% of the population, including patients dependent on dialysis [1]. CKD, defined as an estimated glomerular filtration rate (eGFR) of < 60 ml/min/1.73 m2, is predictive of increased all-cause and cardiovascular mortality in individuals with vascular disease [2] and in asymptomatic middle-aged individuals [3]. There are several key pathophysiological aetiologies of CKD that may contribute to postoperative morbidity, including increased levels of inflammatory factors, excess arterial calcification and endothelial dysfunction [4]. Meta-analysis of elective, noncardiac (largely vascular) surgical studies shows that CKD is an independent risk factor for postoperative death and cardiovascular events [5]. The relationship of chronic, mild renal dysfunction with postoperative morbidity and mortality is unclear, particularly in patients undergoing moderate-risk surgery, such as that for colorectal cancer (CRC). CRC surgery is associated with considerable risk of perioperative complications, particularly in an increasingly elderly population with significant comorbidities [6], such as CKD.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
879
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Postoperative complications have been shown to have a negative impact on longer-term outcomes (including mortality) following CRC resection [7]. Therefore, assessing the relationship of CKD with postoperative morbidity and longer-term mortality may be particularly instructive. In the cardiac surgical setting, the eGFR is of value in predicting cardiovascular postoperative events and cardiovascular and all-cause mortality in patients undergoing surgery [8]. We hypothesized that CKD, when compared with a population with normal renal function, would be associated with an increased risk of postoperative complications and reduced survival following elective CRC surgery.
Method Patient population and study criteria
Consecutive patients with primary CRC without distant organ metastases and who underwent elective and potentially curative surgery (R0) at St Mark’s Hospital, London, between January 2006 and December 2011, were identified. Patients were excluded if they had a recurrent, familial or inherited cancer. Patients were also excluded if they did not have the necessary laboratory data described below. All patienttreatment decisions were discussed and agreed at weekly multidisciplinary team (MDT) meetings. Our institution follows a standardized surveillance protocol and follows up patients for 5 years. Postoperative follow-up data included outpatient documentation, laboratory and radiology results, as well as endoscopic surveillance findings. CT scanning of the chest, abdomen and pelvis is the radiological surveillance imaging modality of choice in our institution and is performed at least annually. Either colonoscopy or CT colonography are used for luminal surveillance. All-cause mortality data were obtained from the UK Thames Cancer Registry and were linked to our institutional CRC data set. Our database is registered with the UK National Research Ethics Committee (reference number: 12/LO/1556).
CKD classification Renal function was evaluated as eGFR, which was calculated using the latest version of the Modification of Diet in Renal Disease (MDRD) study [9] equation, as follows: eGFR (ml/min/1.73 m2) = 186 9 (serum creatinine (mg/dl)/88.4) 1.154 9 (Age) 0.203 9 (0.742 if female) 9 (1.210 if black). Serum creatinine levels were measured using the enzymatic Jaffe method. We used eGFR measurements taken from preoperative blood tests within 4 weeks of the primary operation.
880
Based on eGFR, a well-established CKD staging system [9] classifies patients into one of five categories (Stages 1–5). According to this classification, we divided patients into two subgroups: CKD Stages 0–2 (normal to mild CKD, with eGFR ≥ 60 ml/min/1.73 m2); and CKD Stages 3–4 (moderate to severe CKD, with eGFR between 15 and 60 ml/min/1.73 m2). There were no patients with Stage 5 CKD in this study.
Other data collected and definitions Clinical variables, such as sex, age, American Society of Anesthesiology (ASA) score, concomitant diabetes mellitus and cardiovascular diseases, histological grading, TNM stage [Union for International Cancer Control (UICC) version 5], presence of vascular invasion on histopathology, location of cancer (colon or rectum), surgical approach (laparoscopic or open on an intentionto-treat analysis), operative procedure performed, postoperative complications, and dates of recurrence, death and/or latest hospital contact, were collated from our prospectively recorded database. All dates of death, together with certified cause of death, were gathered from the local Thames Cancer Registry. Postoperative complications were defined according to the Clavien–Dindo classification [10]. Cardiovascular complications after surgery included cardiac disease (such as atrial fibrillation, angina, myocardial infarction or congestive heart failure), cerebrovascular disease (such as stroke or transient ischaemic attack), peripheral vessel disease and other associated disorders. Anastomotic leak was defined as a communication from the anastomosis seen radiologically, at surgical intervention or on postmortem. Anastomotic leaks were further categorized into conservatively managed (including the use of antibiotics), radiologically drained or managed by operative intervention. Overall survival was defined as the time, in months, from the date of primary operation to an individual’s death from any cause. The secondary end-point was disease-free survival, which was defined as the time, in months, from the date of primary operation to the date of identification of disease recurrence by either radiological or histological methods. Statistical analysis
Comparisons of clinicopathological parameters among different groups were made using the v2 test for all categorical variables, and the t-test and Kruskall–Wallis test were used to evaluate differences between continuous normally and non-normally distributed variables, respectively. Disease-free survival and overall survival were assessed using the Kaplan–Meier method and groups were compared using the log-rank test. Stepwise multi-
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
variate Cox proportion analysis was performed. Some variables were dichotomised for the purposes of the Cox regression. A P value of < 0.05 was considered significant. All analyses were performed using the statistical software, Statistical Package for the Social Sciences, version 20.0 (SPSS, Chicago, Illinois, USA).
Results Of 1177 consecutive patients with a diagnosis of CRC who underwent surgical resection, 197 had an emergency operation, 137 had metastatic disease at the time of surgery, 52 patients did not undergo an R0 resection, 50 had recurrent cancer and 48 had a familial, inherited or inflammatory bowel disease-associated cancer. Preoperative biochemical data were not available for 14 patients. Exclusion of these patients resulted in a sample size of 708 patients who had undergone curative elective surgery for CRC. Comparison of the baseline characteristics of the CKD and non-CKD populations showed that the former was older with higher ASA scores and more likely to be female and to have comorbid cardiovascular disease and diabetes at presentation (Table 1). There were more poorly differentiated cancers and also a higher tumour stage in the CKD population. Proportionately more colonic cancers were present in the CKD population and therefore they were more likely to undergo a segmental colectomy and less likely to undergo a rectal resection than the non-CKD population. The postoperative outcome is shown in Table 2. Clavien–Dindo-defined complications occurred in approximately 40% of patients in each group, with no difference in susceptibility between the two populations. Despite the different distribution of surgical procedures, anastomotic leakage occurred in similar frequencies in the CKD (eight of 122; 7%) and non-CKD (40/504; 8%) populations (P = 0.115). Patients with CKD were more likely to experience a cardiovascular complication [25/126 (20%) vs 40/582 (7%); P < 0.001]. In the total patient population, 21% (149/708) had died at the median 48-month follow-up. Mortality at 30-day [six of 124 (4.8%) CKD patients vs 12/580 (2.1%) non-CKD patients; P = 0.012] and 90-day [nine of 122 (7%) CKD patients vs 17/582 (3%) non-CKD patients, P = 0.022) follow-up was greater in the CKD population. Overall survival was greater for patients with localized (Stage I–II) CRC than for Stage III disease. Kaplan–Meier analysis revealed that patients with CKD had poorer overall survival for localized and Stage III cancer (log-rank test, P = 0.019 and P = 0.001, respectively; Figs. 1 and 2). Recurrent disease was seen in 17/126 (14%) of the CKD population and in 97/
CKD and colorectal cancer surgery outcomes
582 (17%) of the non-CKD group. The median time to recurrence was 14 [interquartile range (IQR): 6–24 months] and no differences were seen between the two groups. Disease-free survival was unaffected by CKD stage (log-rank test, P = 0.465; Fig. 3). Colorectal cancer was the primary cause of death in 55% (82/149 deaths; Fig. 4). Noncancer causes of death were predominantly cardiovascular (24/67; 37%) or respiratory (12/67; 18%), whilst the remainder were infectious or other causes. Noncancer deaths were more prevalent in the CKD population [26/126 (21%) in the CKD group vs 41/582 (7%) in the non-CKD group; P < 0.001). Cox regression was undertaken to determine the independent impact of CKD on noncancer death (Table 3). After controlling for age, cancer stage, ASA grade, cardiovascular and diabetic comorbidity, surgical approach and tumour location, CKD was an independent predictor of noncancer death following CRC resection. Applying Cox regression modelling to overall survival, however, revealed that CKD did not have an independent effect: UICC stage, surgical approach and age over 80 years were the only independent predictors of outcome.
Discussion This study has uniquely, to our knowledge, compared the postoperative survival and causes of death following elective primary CRC resection between CKD and nonCKD patient populations. CKD has been shown to predict independently noncancer death, but not overall survival. Additionally, CKD did not appear to affect disease-free survival. Increased 30- and 90-day postoperative mortality were seen in the CKD group. Similar frequencies of postoperative complications were seen in CKD and non-CKD groups, despite the lower proportion of rectal cases in CKD patients. There were more cardiovascular complications observed in the CKD group, but anastomotic leakage rates were similar in CKD and non-CKD groups. The impact of CKD, as measured by low eGFR, on survival in asymptomatic and cardiac postoperative patient groups is well explored [2,3,11], but the outcome following colorectal surgery is less well known. A systematic review of 31 cohort studies found a graded impact of CKD on postoperative cardiac events and allcause mortality following elective noncardiac surgery [5]. This comprised mostly vascular surgery studies, but the general surgery cohorts included were heterogeneous, containing colorectal and other types of surgery, as well as malignant and benign disease. Most studies examining colorectal surgery in renal disease have focused on the end-stage or dialysis-dependent popula-
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
881
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 1 Patient and operative characteristics of colorectal cancer resection subjects compared according to chronic kidney disease (CKD) stage. CKD stage Variable
Stage 0–2 eGFR
Stage 3–4 CKD
P
Male gender Age (years) ASA Stage 3–4 eGFR Cardiovascular comorbidity Diabetes Tumour stage (TNM) T1 T2 T3 T4 Nodal stage (TNM) 0 1 2 Grade G1 (well differentiated) G2 (moderately differentiated) G3 (poorly differentiated) Venous invasion 0 1 2 UICC stage Stage 1 Stage 2 Stage 3 Rectal Operation Colectomy Anterior resection APER Extended/Other Laparoscopic approach (ITT) Length of follow-up
356/582 (61) 66.5 12.8 71/582 (12) 82.8 14.9 234/582 (40) 56/582 (10)
57/126 (45) 76.7 9.1 42/126 (33) 48.1 10.2 72/126 (57) 20/126 (16)
0.001 < 0.001 < 0.001 < 0.001 0.001 0.040 0.017
81/582 106/582 308/582 87/582
10/126 15/126 71/126 30/126
(14) (18) (53) (15)
(8) (12) (56) (24) 0.685
369/582 (63) 129/582 (22) 84/582 (14)
85/126 (67) 23/126 (18) 18/126 (14)
104/582 (18) 425/582 (73) 53/582 (9)
17/126 (13) 88/126 (70) 21/126 (17)
398/582 (68) 173/582 (30) 11/582 (2)
79/126 (63) 45/126 (36) 2/126 (2)
148/582 224/582 210/582 199/582
23/126 62/126 41/126 9/126
0.031
0.416
0.063 (25) (38) (36) (34)
277/582 (48) 210/582 (36) 20/582 (3) 75/582 (13) 313/582 (54) 44 (31–57)
(18) (49) (33) (7)
88/126 (70) 33/126 (26) 3/126 (2) 2/126 (2) 83/126 (66) 45 (33–57)
< 0.001 < 0.001
0.102 0.610
Data are given as no. of positive/total no. of patients (%), mean SD or median (interquartile range). APER, abdominoperineal excision of rectum; ASA, American Society of Anesthesiology; eGFR, estimated glomerular filtration rate; ITT, Intention-to-treat; UICC, Union for International Cancer Control.
tion [12,13], rather than the less severe group, which was investigated in the present study. In a series from the Cleveland Clinic, Canedo et al. [14]. undertook a case–control study to examine the impact of nondialysis-dependent (Stage 3–4) CKD on postoperative complications following colorectal surgery. Their CKD population had an increased risk of postoperative complications, particularly infection. However, their case series was not restricted to cancer operations and had a high proportion of stoma reversals, which present a dif-
882
ferent perioperative risk profile compared with resectional procedures. Nozawa et al. [15]. compared the outcome of elective CRC resection in their institution between CKD and non-CKD groups and found an increased risk of postoperative cardiovascular morbidity in Stage 3–4 CKD, but no effect on overall postoperative mortality. However, there were very high rates of overall survival in this study (5-year overall survival Stage I–III CRC of 92%) and no reports on causes of
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 2 Postoperative outcome and survival following colorectal cancer resection: comparison between chronic kidney disease (CKD) and non-CKD populations.
CKD Normal renal function
1.0
CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
Complications Anastomotic leakage* Conservative Nonoperative intervention Operative management Cardiovascular complication Clavien–Dindo grade complication† 1 2 3 4 5 90-day mortality All-cause mortality Cause of death Cancer Cardiovascular Infection Respiratory Other Combined noncancer death
Stage 0–2 eGFR
Stage 3–4 CKD
236/582 (41) 40/504 (8)
46/126 (37) 8/122 (7)
7/40 (18) 3/40 (8)
1/8 (13) 2/8 (25)
P 0.401 0.115
0.6
0.4
0.2
0.0 0
30/40 (75)
12
24
36 48 60 Follow up (months)
72
84
96
5/8 (63)
40/582 (7)
25/126 (20)
< 0.001
226/582 (39)
56/126 (44)
0.239
74/582 71/582 48/582 21/582 12/582 17/582 104/582
(13) (12) (8) (4) (2) (3) (18)
20/126 17/126 9/126 4/126 6/126 9/126 45/126
(16) (13) (7) (3) (5) (7) (36)
63/104 12/104 9/104 14/104 6/104 41/582
(61) (12) (9) (13) (6) (7)
19/45 12/45 3/45 9/45 2/45 26/126
(42) (27) (7) (20) (4) (21)
0.022 < 0.001 0.045
< 0.001
Data are given as no. of positive/total no. of patients (%). eGFR, estimated glomerular filtration rate. *Analyses performed for restorative resections only. †Analysis undertaken at 30-day postoperative follow-up.
death, which may limit the application of these data to other populations. This study has shown CKD to be an independent predictor of noncancer death in our population, with cardiovascular mortality being the most common cause. Cardiovascular complications have also been shown to be more common in the CKD group. Cardiovascular mortality and CKD are linked by endothelial dysfunction, which is common to both CKD and the atherosclerotic process [16]. This endothelial damage is not only explained by classic cardiovascular risk factors common to both CKD and atherosclerotic disease, but also by the proinflammatory environment in patients with CKD as a result of uraemia, malnutrition, volume overload or calcium-phosphorus metabolism alterations
Figure 1 Overall survival for the chronic kidney disease (CKD) and non-CKD populations with Stage I–II [Union for International Cancer Control (UICC) classification] colorectal cancer. P = 0.019 (log-rank test) among groups. Log rank: P = 0.019.
[4]. Additionally, CKD is a state of relative immune suppression because of nutritional deficiencies, loss of serum immune-system components (such as opsonins) and lymphocyte suppression [11]. This may have contributed to an increasing risk of death from infectious and respiratory disease. A number of limitations of this study deserve consideration. It is a single-institution study and as such may have limited external validity. The sample size has prevented an accurate analysis of the CRC stage-specific impact of CKD, which would certainly be of interest in future work. However, the prospective nature of the data collection adds strength and it seems unlikely that the CKD population identified in this study is significantly different from that in other western populations. Certainly, the level of diabetes and cardiovascular disease are consistent with primary-care CKD data from the UK [17]. This study used the MDRD formula to
Cumlatative survival
Variable
Cumlatative survival
0.8
1.0
CKD Normal renal function
0.8
CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
0.6
0.4
0.2
0.0 0
12
24 36 48 Follow up (months)
60
72
84
Figure 2 Overall survival for the chronic kidney disease (CKD) and non-CKD populations with Stage III [Union for International Cancer Control (UICC) classification] colorectal cancer P = 0.001 (log-rank test) among groups. Log rank: P = 0.001.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
883
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 3 Cox proportional hazard: multivariate analysis for overall mortality and noncancer death. Overall survival
Noncancer death
Variable
Hazard ratio
95% CI
P
Hazard ratio
95% CI
P
CKD UICC stage CV disease ASA ≥ 3 Diabetes Male gender Rectal case Laparoscopic approach Postoperative complication Age ≥ 80 years
1.21 1.67 1.10 1.36 1.10 1.40 1.31 1.05 1.20 2.98
0.90–1.47 1.07–2.61 0.70–1.54 0.93–1.65 0.70–1.51 0.95–1.86 0.89–1.74 0.92–1.18 0.95–1.43 1.44–6.12
0.116 0.025 0.598 0.101 0.390 0.111 0.202 0.138 0.099 0.002
1.82 0.92 1.20 1.30 1.07 0.65 0.94 0.89 1.94 2.28
1.07–3.10 0.66–1.27 0.72–1.99 0.72–2.32 0.50–2.27 0.39–1.11 0.78–1.18 0.74–1.05 1.18–3.19 1.34–3.89
0.027 0.593 0.492 0.384 0.862 0.113 0.107 0.102 0.009 0.002
Renal funcrtion groups Normal renal function CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
1.0
Cumlatative survival
0.8
0.6
0.4
0.2
0.0 0
12
24
36 48 60 Follow up (months)
72
84
Percentage of all mortalities within renal function group
ASA, American Society of Anesthesiology; CKD, chronic kidney disease; CV, cardiovascular; UICC, Union for International Cancer Control.
CKD Normal renal function CKD Stage 3-4 60
40
20
0 Cancer Cardiovascular Infection Respiratory Other
Figure 3 Disease-free survival between chronic kidney disease (CKD) and non-CKD populations for combined colorectal cancer stages I–III [Union for International Cancer Control (UICC) classification]. P = 0.465 (log-rank test) among groups. Log rank: P = 0.465.
estimate GFR. Some reports suggest that another calculation, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), has slightly greater accuracy [18], but this relies on isotope dilution mass spectrometry measurements, which are resource-intensive and costly. The MDRD approach uses routinely collected perioperative biochemical values and therefore represents a more practical assessment of eGFR in resource-limited settings. Cystatin C, an endogenous protein biomarker filtered exclusively by the glomerulus, may be superior to both MDRD and CKD-EPI as a predictor of survival, particularly in elderly patients [19]. Future large-scale studies examining cystatin C-based eGFR measurements in estimating perioperative risk following CRC surgery would be of considerable interest.
884
Causeof death
Figure 4 Comparison of cause-specific mortality between chronic kidney disease (CKD) and non-CKD populations. P = 0.031 (v2 test). Chi square: P = 0.031. P value for difference in distribution of cancer mortality between normal renal function and CKD stage 3–4 groups.
Routine measurement of eGFR and identification of patients with CKD may allow more accurate assessment of competing mortality risks within an often elderly and infirm colorectal surgery population. This study has shown that a large proportion of deaths following CRC surgery are not secondary to cancer and that CKD is independently associated with those patients who die from noncancer death. While competing hazard analysis has been explored in prostate [20] and renal [21] cancers, it has not been thoroughly examined in CRC. With the evolution of novel, more minimally invasive treatments, such as transanal endoscopic microsurgery for rectal cancer and the endoscopic resection of malignant polyps [22], the need for more nuanced
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
estimation of competing risks from CRC and patient comorbidity is clear. Current perioperative scoring systems, such as the Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM), consistently overscore mortality [23,24] and do not address longer-term survival, which frequently governs the preoperative discussion with the cancer patient. Identification of CKD in routine preoperative assessment of CRC surgery may assist in more balanced decision-making, particularly in an increasingly older and comorbid patient group.
Author contributions Study conception and design: AC, ODF, RHK. Acquisition of data: AC, GM, AA, SN, JTJ, RHK. Analysis and interpretation of data: AC, GM, AA, SN. Writing manuscript: AC, PS, JTJ, ODF, RHK.
Conflicts of interest None.
Sources of funding None.
References 1 Levey AS, Coresh J. Chronic kidney disease. Lancet 2012; 379: 165–80. 2 Shadman R, Allison MA, Criqui MH. Glomerular filtration rate and N-terminal pro-brain natriuretic peptide as predictors of cardiovascular mortality in vascular patients. J Am Coll Cardiol 2007; 49: 2172–81. 3 Ryan TP, Fisher SG, Elder JL et al. Increased cardiovascular risk associated with reduced kidney function. Am J Nephrol 2009; 29: 620–5. 4 Gansevoort RT, Correa-Rotter R, Hemmelgarn BR et al. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. Lancet 2013; 382: 339– 52. 5 Mathew A, Devereaux PJ, O’Hare A et al. Chronic kidney disease and postoperative mortality: a systematic review and meta-analysis. Kidney Int 2008; 73: 1069–81. 6 Faiz O, Haji A, Bottle A et al. Elective colonic surgery for cancer in the elderly: an investigation into postoperative mortality in English NHS hospitals between 1996 and 2007. Colorectal Dis 2011; 13: 779–85. 7 Mirnezami A, Mirnezami R, Chandrakumaran K et al. Increased local recurrence and reduced survival from colorectal cancer following anastomotic leak: systematic review and meta-analysis. Ann Surg 2011; 253: 890–9.
CKD and colorectal cancer surgery outcomes
8 Foot CL, Chinthamuneedi M, Fraser JF et al. The association between preoperative eGFR and outcomes in cardiac surgical patients. Crit Care Resusc 2009; 11: 184–90. 9 Levey AS, de Jong PE, Coresh J et al. The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies Conference report. Kidney Int 2011; 80: 17–28. 10 Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240: 205–13. 11 Maizel J, Deransy R, Dehedin B et al. Impact of non-dialysis chronic kidney disease on survival in patients with septic shock. BMC Nephrol 2013; 14: 77. 12 Drolet S, Maclean AR, Myers RP et al. Morbidity and mortality following colorectal surgery in patients with end-stage renal failure: a population-based study. Dis Colon Rectum 2010; 53: 1508–16. 13 Stewart DB, Hollenbeak C, Boltz M. Renal transplant status in patients undergoing colorectal surgery: is immunosuppression safer than kidney disease? Surgery 2012; 152: 537–47. 14 Canedo J, Ricciardi K, DaSilva G et al. Are postoperative complications more common following colon and rectal surgery in patients with chronic kidney disease? Colorectal Dis 2013; 15: 85–90. 15 Nozawa H, Kitayama J, Sunami E et al. Impact of chronic kidney disease on outcomes of surgical resection for primary colorectal cancer: a retrospective cohort review. Dis Colon Rectum 2012; 55: 948–56. 16 Stafford-Smith M. Heart and kidneys: sharing more than just blood. Curr Opin Anaesthesiol 2007; 20: 65–9. 17 Kearns B, Gallagher H, de Lusignan S. Predicting the prevalence of chronic kidney disease in the English population: a cross-sectional study. BMC Nephrol 2013; 14: 49. 18 Schaeffner ES, Ebert N, Delanaye P et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann Intern Med 2012; 157: 471–81. 19 Inker LA, Schmid CH, Tighiouart H et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012; 367: 20–9. 20 Briganti A, Spahn M, Joniau S et al. Impact of age and comorbidities on long-term survival of patients with high-risk prostate cancer treated with radical prostatectomy: a multi-institutional competing-risks analysis. Eur Urol 2013; 63: 693–701. 21 Sun M, Trinh QD, Bianchi M et al. A non-cancer-related survival benefit is associated with partial nephrectomy. Eur Urol 2012; 61: 725–31. 22 Brigic A, Fraser C, Sibbons P et al. Individualization of surgical management for early-stage colonic cancer. Colorectal Dis 2011; 13(Suppl 7): 59–62. 23 Tan KY, Kawamura YJ, Tokomitsu A et al. Assessment for frailty is useful for predicting morbidity in elderly patients undergoing colorectal cancer resection whose comorbidities are already optimized. Am J Surg 2012; 204: 139–43. 24 Leung E, McArdle K, Wong LS. Risk-adjusted scoring systems in colorectal surgery. Int J Surg 2011; 9: 130–5.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
885
Copyright of Colorectal Disease is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Original article
Impact of chronic kidney disease on postoperative outcome following colorectal cancer surgery A. Currie*, G. Malietzis*, A. Askari*, S. Nachiappan*, P. Swift†, J. T. Jenkins*‡, O. D. Faiz*‡ and R. H. Kennedy*‡ *Surgical Epidemiology, Trials and Outcomes Centre, St Mark’s Hospital, London, UK, †South West Thames Kidney and Transplantation Unit, St Helier Hospital, Epsom and St Helier Hospital NHS Trust, Surrey, UK and ‡Department of Surgery and Cancer, Imperial College London, London, UK Received 28 January 2014; accepted 1 April 2014; Accepted Article online 8 October 2013
Abstract Aim Chronic kidney disease (CKD) is increasing in prevalence and is associated with cardiovascular events and mortality in asymptomatic and vascular surgery populations. This study aimed to determine the role of CKD in stratifying peri- and postoperative risk for colorectal cancer (CRC) patients with nonmetastatic disease undergoing elective curative resection. Method Patients diagnosed with nonmetastatic colorectal adenocarcinoma and undergoing surgical resection between 2006 and 2011 were identified from a prospectively collated database. Further information on survival and cause of death was gathered from a regional cancer registry. Estimated glomerular filtration rates were calculated using the Modification of Diet in Renal Disease (MDRD) equation. Kaplan–Meier survival curves were constructed for disease-free and overall survival. Multivariate Cox regression models were used to determine the role of CKD after stratification by several clinicopathological factors. Results Seven-hundred and eight colorectal resections were studied [median follow up: 45 (interquartile range, 21–65) months). Overall postoperative complications were similar, but patients with CKD were more likely to develop cardiovascular morbidity (P < 0.001) and 30-day mortality [4.8% (six of 124) in the CKD group vs 2.1% (12/580) in the non-CKD group]. Kaplan–Meier analysis revealed poorer overall survival for localized (Stage I–II; P = 0.019) and Stage III (P = 0.001) CRC in the CKD population. Multivariate Cox regression analysis identified CKD as an independent prognostic factor for noncancer death [hazard ratio (HR) = 1.82 (95% CI: 1.07–3.10); P = 0.027] but not for overall survival [HR = 1.21 (95% CI: 0.90–1.47); P = 0.116].
Correspondence to: Prof. Robin Kennedy, Consultant Colorectal Surgeon, St. Mark’s Hospital, Watford Road, Harrow, Middlesex HA1 3UJ, UK. E-mail: [email protected]
Conclusion Patients with CKD may be more likely to develop cardiovascular complications following CRC resection and have an increased risk of a noncancer death. Future research should explore the interaction of CKD in competing mortality risks following CRC surgery. Keywords Colorectal cancer, chronic kidney disease, complications, survival What does this paper add to the literature? This paper is the first to compare the postoperative outcomes, long-term survival and cause of death between chronic kidney disease and normal renal function patients following elective colorectal cancer surgery.
Introduction Chronic kidney disease (CKD) affects 5% of the population, including patients dependent on dialysis [1]. CKD, defined as an estimated glomerular filtration rate (eGFR) of < 60 ml/min/1.73 m2, is predictive of increased all-cause and cardiovascular mortality in individuals with vascular disease [2] and in asymptomatic middle-aged individuals [3]. There are several key pathophysiological aetiologies of CKD that may contribute to postoperative morbidity, including increased levels of inflammatory factors, excess arterial calcification and endothelial dysfunction [4]. Meta-analysis of elective, noncardiac (largely vascular) surgical studies shows that CKD is an independent risk factor for postoperative death and cardiovascular events [5]. The relationship of chronic, mild renal dysfunction with postoperative morbidity and mortality is unclear, particularly in patients undergoing moderate-risk surgery, such as that for colorectal cancer (CRC). CRC surgery is associated with considerable risk of perioperative complications, particularly in an increasingly elderly population with significant comorbidities [6], such as CKD.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
879
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Postoperative complications have been shown to have a negative impact on longer-term outcomes (including mortality) following CRC resection [7]. Therefore, assessing the relationship of CKD with postoperative morbidity and longer-term mortality may be particularly instructive. In the cardiac surgical setting, the eGFR is of value in predicting cardiovascular postoperative events and cardiovascular and all-cause mortality in patients undergoing surgery [8]. We hypothesized that CKD, when compared with a population with normal renal function, would be associated with an increased risk of postoperative complications and reduced survival following elective CRC surgery.
Method Patient population and study criteria
Consecutive patients with primary CRC without distant organ metastases and who underwent elective and potentially curative surgery (R0) at St Mark’s Hospital, London, between January 2006 and December 2011, were identified. Patients were excluded if they had a recurrent, familial or inherited cancer. Patients were also excluded if they did not have the necessary laboratory data described below. All patienttreatment decisions were discussed and agreed at weekly multidisciplinary team (MDT) meetings. Our institution follows a standardized surveillance protocol and follows up patients for 5 years. Postoperative follow-up data included outpatient documentation, laboratory and radiology results, as well as endoscopic surveillance findings. CT scanning of the chest, abdomen and pelvis is the radiological surveillance imaging modality of choice in our institution and is performed at least annually. Either colonoscopy or CT colonography are used for luminal surveillance. All-cause mortality data were obtained from the UK Thames Cancer Registry and were linked to our institutional CRC data set. Our database is registered with the UK National Research Ethics Committee (reference number: 12/LO/1556).
CKD classification Renal function was evaluated as eGFR, which was calculated using the latest version of the Modification of Diet in Renal Disease (MDRD) study [9] equation, as follows: eGFR (ml/min/1.73 m2) = 186 9 (serum creatinine (mg/dl)/88.4) 1.154 9 (Age) 0.203 9 (0.742 if female) 9 (1.210 if black). Serum creatinine levels were measured using the enzymatic Jaffe method. We used eGFR measurements taken from preoperative blood tests within 4 weeks of the primary operation.
880
Based on eGFR, a well-established CKD staging system [9] classifies patients into one of five categories (Stages 1–5). According to this classification, we divided patients into two subgroups: CKD Stages 0–2 (normal to mild CKD, with eGFR ≥ 60 ml/min/1.73 m2); and CKD Stages 3–4 (moderate to severe CKD, with eGFR between 15 and 60 ml/min/1.73 m2). There were no patients with Stage 5 CKD in this study.
Other data collected and definitions Clinical variables, such as sex, age, American Society of Anesthesiology (ASA) score, concomitant diabetes mellitus and cardiovascular diseases, histological grading, TNM stage [Union for International Cancer Control (UICC) version 5], presence of vascular invasion on histopathology, location of cancer (colon or rectum), surgical approach (laparoscopic or open on an intentionto-treat analysis), operative procedure performed, postoperative complications, and dates of recurrence, death and/or latest hospital contact, were collated from our prospectively recorded database. All dates of death, together with certified cause of death, were gathered from the local Thames Cancer Registry. Postoperative complications were defined according to the Clavien–Dindo classification [10]. Cardiovascular complications after surgery included cardiac disease (such as atrial fibrillation, angina, myocardial infarction or congestive heart failure), cerebrovascular disease (such as stroke or transient ischaemic attack), peripheral vessel disease and other associated disorders. Anastomotic leak was defined as a communication from the anastomosis seen radiologically, at surgical intervention or on postmortem. Anastomotic leaks were further categorized into conservatively managed (including the use of antibiotics), radiologically drained or managed by operative intervention. Overall survival was defined as the time, in months, from the date of primary operation to an individual’s death from any cause. The secondary end-point was disease-free survival, which was defined as the time, in months, from the date of primary operation to the date of identification of disease recurrence by either radiological or histological methods. Statistical analysis
Comparisons of clinicopathological parameters among different groups were made using the v2 test for all categorical variables, and the t-test and Kruskall–Wallis test were used to evaluate differences between continuous normally and non-normally distributed variables, respectively. Disease-free survival and overall survival were assessed using the Kaplan–Meier method and groups were compared using the log-rank test. Stepwise multi-
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
variate Cox proportion analysis was performed. Some variables were dichotomised for the purposes of the Cox regression. A P value of < 0.05 was considered significant. All analyses were performed using the statistical software, Statistical Package for the Social Sciences, version 20.0 (SPSS, Chicago, Illinois, USA).
Results Of 1177 consecutive patients with a diagnosis of CRC who underwent surgical resection, 197 had an emergency operation, 137 had metastatic disease at the time of surgery, 52 patients did not undergo an R0 resection, 50 had recurrent cancer and 48 had a familial, inherited or inflammatory bowel disease-associated cancer. Preoperative biochemical data were not available for 14 patients. Exclusion of these patients resulted in a sample size of 708 patients who had undergone curative elective surgery for CRC. Comparison of the baseline characteristics of the CKD and non-CKD populations showed that the former was older with higher ASA scores and more likely to be female and to have comorbid cardiovascular disease and diabetes at presentation (Table 1). There were more poorly differentiated cancers and also a higher tumour stage in the CKD population. Proportionately more colonic cancers were present in the CKD population and therefore they were more likely to undergo a segmental colectomy and less likely to undergo a rectal resection than the non-CKD population. The postoperative outcome is shown in Table 2. Clavien–Dindo-defined complications occurred in approximately 40% of patients in each group, with no difference in susceptibility between the two populations. Despite the different distribution of surgical procedures, anastomotic leakage occurred in similar frequencies in the CKD (eight of 122; 7%) and non-CKD (40/504; 8%) populations (P = 0.115). Patients with CKD were more likely to experience a cardiovascular complication [25/126 (20%) vs 40/582 (7%); P < 0.001]. In the total patient population, 21% (149/708) had died at the median 48-month follow-up. Mortality at 30-day [six of 124 (4.8%) CKD patients vs 12/580 (2.1%) non-CKD patients; P = 0.012] and 90-day [nine of 122 (7%) CKD patients vs 17/582 (3%) non-CKD patients, P = 0.022) follow-up was greater in the CKD population. Overall survival was greater for patients with localized (Stage I–II) CRC than for Stage III disease. Kaplan–Meier analysis revealed that patients with CKD had poorer overall survival for localized and Stage III cancer (log-rank test, P = 0.019 and P = 0.001, respectively; Figs. 1 and 2). Recurrent disease was seen in 17/126 (14%) of the CKD population and in 97/
CKD and colorectal cancer surgery outcomes
582 (17%) of the non-CKD group. The median time to recurrence was 14 [interquartile range (IQR): 6–24 months] and no differences were seen between the two groups. Disease-free survival was unaffected by CKD stage (log-rank test, P = 0.465; Fig. 3). Colorectal cancer was the primary cause of death in 55% (82/149 deaths; Fig. 4). Noncancer causes of death were predominantly cardiovascular (24/67; 37%) or respiratory (12/67; 18%), whilst the remainder were infectious or other causes. Noncancer deaths were more prevalent in the CKD population [26/126 (21%) in the CKD group vs 41/582 (7%) in the non-CKD group; P < 0.001). Cox regression was undertaken to determine the independent impact of CKD on noncancer death (Table 3). After controlling for age, cancer stage, ASA grade, cardiovascular and diabetic comorbidity, surgical approach and tumour location, CKD was an independent predictor of noncancer death following CRC resection. Applying Cox regression modelling to overall survival, however, revealed that CKD did not have an independent effect: UICC stage, surgical approach and age over 80 years were the only independent predictors of outcome.
Discussion This study has uniquely, to our knowledge, compared the postoperative survival and causes of death following elective primary CRC resection between CKD and nonCKD patient populations. CKD has been shown to predict independently noncancer death, but not overall survival. Additionally, CKD did not appear to affect disease-free survival. Increased 30- and 90-day postoperative mortality were seen in the CKD group. Similar frequencies of postoperative complications were seen in CKD and non-CKD groups, despite the lower proportion of rectal cases in CKD patients. There were more cardiovascular complications observed in the CKD group, but anastomotic leakage rates were similar in CKD and non-CKD groups. The impact of CKD, as measured by low eGFR, on survival in asymptomatic and cardiac postoperative patient groups is well explored [2,3,11], but the outcome following colorectal surgery is less well known. A systematic review of 31 cohort studies found a graded impact of CKD on postoperative cardiac events and allcause mortality following elective noncardiac surgery [5]. This comprised mostly vascular surgery studies, but the general surgery cohorts included were heterogeneous, containing colorectal and other types of surgery, as well as malignant and benign disease. Most studies examining colorectal surgery in renal disease have focused on the end-stage or dialysis-dependent popula-
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
881
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 1 Patient and operative characteristics of colorectal cancer resection subjects compared according to chronic kidney disease (CKD) stage. CKD stage Variable
Stage 0–2 eGFR
Stage 3–4 CKD
P
Male gender Age (years) ASA Stage 3–4 eGFR Cardiovascular comorbidity Diabetes Tumour stage (TNM) T1 T2 T3 T4 Nodal stage (TNM) 0 1 2 Grade G1 (well differentiated) G2 (moderately differentiated) G3 (poorly differentiated) Venous invasion 0 1 2 UICC stage Stage 1 Stage 2 Stage 3 Rectal Operation Colectomy Anterior resection APER Extended/Other Laparoscopic approach (ITT) Length of follow-up
356/582 (61) 66.5 12.8 71/582 (12) 82.8 14.9 234/582 (40) 56/582 (10)
57/126 (45) 76.7 9.1 42/126 (33) 48.1 10.2 72/126 (57) 20/126 (16)
0.001 < 0.001 < 0.001 < 0.001 0.001 0.040 0.017
81/582 106/582 308/582 87/582
10/126 15/126 71/126 30/126
(14) (18) (53) (15)
(8) (12) (56) (24) 0.685
369/582 (63) 129/582 (22) 84/582 (14)
85/126 (67) 23/126 (18) 18/126 (14)
104/582 (18) 425/582 (73) 53/582 (9)
17/126 (13) 88/126 (70) 21/126 (17)
398/582 (68) 173/582 (30) 11/582 (2)
79/126 (63) 45/126 (36) 2/126 (2)
148/582 224/582 210/582 199/582
23/126 62/126 41/126 9/126
0.031
0.416
0.063 (25) (38) (36) (34)
277/582 (48) 210/582 (36) 20/582 (3) 75/582 (13) 313/582 (54) 44 (31–57)
(18) (49) (33) (7)
88/126 (70) 33/126 (26) 3/126 (2) 2/126 (2) 83/126 (66) 45 (33–57)
< 0.001 < 0.001
0.102 0.610
Data are given as no. of positive/total no. of patients (%), mean SD or median (interquartile range). APER, abdominoperineal excision of rectum; ASA, American Society of Anesthesiology; eGFR, estimated glomerular filtration rate; ITT, Intention-to-treat; UICC, Union for International Cancer Control.
tion [12,13], rather than the less severe group, which was investigated in the present study. In a series from the Cleveland Clinic, Canedo et al. [14]. undertook a case–control study to examine the impact of nondialysis-dependent (Stage 3–4) CKD on postoperative complications following colorectal surgery. Their CKD population had an increased risk of postoperative complications, particularly infection. However, their case series was not restricted to cancer operations and had a high proportion of stoma reversals, which present a dif-
882
ferent perioperative risk profile compared with resectional procedures. Nozawa et al. [15]. compared the outcome of elective CRC resection in their institution between CKD and non-CKD groups and found an increased risk of postoperative cardiovascular morbidity in Stage 3–4 CKD, but no effect on overall postoperative mortality. However, there were very high rates of overall survival in this study (5-year overall survival Stage I–III CRC of 92%) and no reports on causes of
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 2 Postoperative outcome and survival following colorectal cancer resection: comparison between chronic kidney disease (CKD) and non-CKD populations.
CKD Normal renal function
1.0
CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
Complications Anastomotic leakage* Conservative Nonoperative intervention Operative management Cardiovascular complication Clavien–Dindo grade complication† 1 2 3 4 5 90-day mortality All-cause mortality Cause of death Cancer Cardiovascular Infection Respiratory Other Combined noncancer death
Stage 0–2 eGFR
Stage 3–4 CKD
236/582 (41) 40/504 (8)
46/126 (37) 8/122 (7)
7/40 (18) 3/40 (8)
1/8 (13) 2/8 (25)
P 0.401 0.115
0.6
0.4
0.2
0.0 0
30/40 (75)
12
24
36 48 60 Follow up (months)
72
84
96
5/8 (63)
40/582 (7)
25/126 (20)
< 0.001
226/582 (39)
56/126 (44)
0.239
74/582 71/582 48/582 21/582 12/582 17/582 104/582
(13) (12) (8) (4) (2) (3) (18)
20/126 17/126 9/126 4/126 6/126 9/126 45/126
(16) (13) (7) (3) (5) (7) (36)
63/104 12/104 9/104 14/104 6/104 41/582
(61) (12) (9) (13) (6) (7)
19/45 12/45 3/45 9/45 2/45 26/126
(42) (27) (7) (20) (4) (21)
0.022 < 0.001 0.045
< 0.001
Data are given as no. of positive/total no. of patients (%). eGFR, estimated glomerular filtration rate. *Analyses performed for restorative resections only. †Analysis undertaken at 30-day postoperative follow-up.
death, which may limit the application of these data to other populations. This study has shown CKD to be an independent predictor of noncancer death in our population, with cardiovascular mortality being the most common cause. Cardiovascular complications have also been shown to be more common in the CKD group. Cardiovascular mortality and CKD are linked by endothelial dysfunction, which is common to both CKD and the atherosclerotic process [16]. This endothelial damage is not only explained by classic cardiovascular risk factors common to both CKD and atherosclerotic disease, but also by the proinflammatory environment in patients with CKD as a result of uraemia, malnutrition, volume overload or calcium-phosphorus metabolism alterations
Figure 1 Overall survival for the chronic kidney disease (CKD) and non-CKD populations with Stage I–II [Union for International Cancer Control (UICC) classification] colorectal cancer. P = 0.019 (log-rank test) among groups. Log rank: P = 0.019.
[4]. Additionally, CKD is a state of relative immune suppression because of nutritional deficiencies, loss of serum immune-system components (such as opsonins) and lymphocyte suppression [11]. This may have contributed to an increasing risk of death from infectious and respiratory disease. A number of limitations of this study deserve consideration. It is a single-institution study and as such may have limited external validity. The sample size has prevented an accurate analysis of the CRC stage-specific impact of CKD, which would certainly be of interest in future work. However, the prospective nature of the data collection adds strength and it seems unlikely that the CKD population identified in this study is significantly different from that in other western populations. Certainly, the level of diabetes and cardiovascular disease are consistent with primary-care CKD data from the UK [17]. This study used the MDRD formula to
Cumlatative survival
Variable
Cumlatative survival
0.8
1.0
CKD Normal renal function
0.8
CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
0.6
0.4
0.2
0.0 0
12
24 36 48 Follow up (months)
60
72
84
Figure 2 Overall survival for the chronic kidney disease (CKD) and non-CKD populations with Stage III [Union for International Cancer Control (UICC) classification] colorectal cancer P = 0.001 (log-rank test) among groups. Log rank: P = 0.001.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
883
A. Currie et al.
CKD and colorectal cancer surgery outcomes
Table 3 Cox proportional hazard: multivariate analysis for overall mortality and noncancer death. Overall survival
Noncancer death
Variable
Hazard ratio
95% CI
P
Hazard ratio
95% CI
P
CKD UICC stage CV disease ASA ≥ 3 Diabetes Male gender Rectal case Laparoscopic approach Postoperative complication Age ≥ 80 years
1.21 1.67 1.10 1.36 1.10 1.40 1.31 1.05 1.20 2.98
0.90–1.47 1.07–2.61 0.70–1.54 0.93–1.65 0.70–1.51 0.95–1.86 0.89–1.74 0.92–1.18 0.95–1.43 1.44–6.12
0.116 0.025 0.598 0.101 0.390 0.111 0.202 0.138 0.099 0.002
1.82 0.92 1.20 1.30 1.07 0.65 0.94 0.89 1.94 2.28
1.07–3.10 0.66–1.27 0.72–1.99 0.72–2.32 0.50–2.27 0.39–1.11 0.78–1.18 0.74–1.05 1.18–3.19 1.34–3.89
0.027 0.593 0.492 0.384 0.862 0.113 0.107 0.102 0.009 0.002
Renal funcrtion groups Normal renal function CKD Stage 3-4 Normal renal function censored CKD Stage 3-4 censored
1.0
Cumlatative survival
0.8
0.6
0.4
0.2
0.0 0
12
24
36 48 60 Follow up (months)
72
84
Percentage of all mortalities within renal function group
ASA, American Society of Anesthesiology; CKD, chronic kidney disease; CV, cardiovascular; UICC, Union for International Cancer Control.
CKD Normal renal function CKD Stage 3-4 60
40
20
0 Cancer Cardiovascular Infection Respiratory Other
Figure 3 Disease-free survival between chronic kidney disease (CKD) and non-CKD populations for combined colorectal cancer stages I–III [Union for International Cancer Control (UICC) classification]. P = 0.465 (log-rank test) among groups. Log rank: P = 0.465.
estimate GFR. Some reports suggest that another calculation, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), has slightly greater accuracy [18], but this relies on isotope dilution mass spectrometry measurements, which are resource-intensive and costly. The MDRD approach uses routinely collected perioperative biochemical values and therefore represents a more practical assessment of eGFR in resource-limited settings. Cystatin C, an endogenous protein biomarker filtered exclusively by the glomerulus, may be superior to both MDRD and CKD-EPI as a predictor of survival, particularly in elderly patients [19]. Future large-scale studies examining cystatin C-based eGFR measurements in estimating perioperative risk following CRC surgery would be of considerable interest.
884
Causeof death
Figure 4 Comparison of cause-specific mortality between chronic kidney disease (CKD) and non-CKD populations. P = 0.031 (v2 test). Chi square: P = 0.031. P value for difference in distribution of cancer mortality between normal renal function and CKD stage 3–4 groups.
Routine measurement of eGFR and identification of patients with CKD may allow more accurate assessment of competing mortality risks within an often elderly and infirm colorectal surgery population. This study has shown that a large proportion of deaths following CRC surgery are not secondary to cancer and that CKD is independently associated with those patients who die from noncancer death. While competing hazard analysis has been explored in prostate [20] and renal [21] cancers, it has not been thoroughly examined in CRC. With the evolution of novel, more minimally invasive treatments, such as transanal endoscopic microsurgery for rectal cancer and the endoscopic resection of malignant polyps [22], the need for more nuanced
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
A. Currie et al.
estimation of competing risks from CRC and patient comorbidity is clear. Current perioperative scoring systems, such as the Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM), consistently overscore mortality [23,24] and do not address longer-term survival, which frequently governs the preoperative discussion with the cancer patient. Identification of CKD in routine preoperative assessment of CRC surgery may assist in more balanced decision-making, particularly in an increasingly older and comorbid patient group.
Author contributions Study conception and design: AC, ODF, RHK. Acquisition of data: AC, GM, AA, SN, JTJ, RHK. Analysis and interpretation of data: AC, GM, AA, SN. Writing manuscript: AC, PS, JTJ, ODF, RHK.
Conflicts of interest None.
Sources of funding None.
References 1 Levey AS, Coresh J. Chronic kidney disease. Lancet 2012; 379: 165–80. 2 Shadman R, Allison MA, Criqui MH. Glomerular filtration rate and N-terminal pro-brain natriuretic peptide as predictors of cardiovascular mortality in vascular patients. J Am Coll Cardiol 2007; 49: 2172–81. 3 Ryan TP, Fisher SG, Elder JL et al. Increased cardiovascular risk associated with reduced kidney function. Am J Nephrol 2009; 29: 620–5. 4 Gansevoort RT, Correa-Rotter R, Hemmelgarn BR et al. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. Lancet 2013; 382: 339– 52. 5 Mathew A, Devereaux PJ, O’Hare A et al. Chronic kidney disease and postoperative mortality: a systematic review and meta-analysis. Kidney Int 2008; 73: 1069–81. 6 Faiz O, Haji A, Bottle A et al. Elective colonic surgery for cancer in the elderly: an investigation into postoperative mortality in English NHS hospitals between 1996 and 2007. Colorectal Dis 2011; 13: 779–85. 7 Mirnezami A, Mirnezami R, Chandrakumaran K et al. Increased local recurrence and reduced survival from colorectal cancer following anastomotic leak: systematic review and meta-analysis. Ann Surg 2011; 253: 890–9.
CKD and colorectal cancer surgery outcomes
8 Foot CL, Chinthamuneedi M, Fraser JF et al. The association between preoperative eGFR and outcomes in cardiac surgical patients. Crit Care Resusc 2009; 11: 184–90. 9 Levey AS, de Jong PE, Coresh J et al. The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies Conference report. Kidney Int 2011; 80: 17–28. 10 Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240: 205–13. 11 Maizel J, Deransy R, Dehedin B et al. Impact of non-dialysis chronic kidney disease on survival in patients with septic shock. BMC Nephrol 2013; 14: 77. 12 Drolet S, Maclean AR, Myers RP et al. Morbidity and mortality following colorectal surgery in patients with end-stage renal failure: a population-based study. Dis Colon Rectum 2010; 53: 1508–16. 13 Stewart DB, Hollenbeak C, Boltz M. Renal transplant status in patients undergoing colorectal surgery: is immunosuppression safer than kidney disease? Surgery 2012; 152: 537–47. 14 Canedo J, Ricciardi K, DaSilva G et al. Are postoperative complications more common following colon and rectal surgery in patients with chronic kidney disease? Colorectal Dis 2013; 15: 85–90. 15 Nozawa H, Kitayama J, Sunami E et al. Impact of chronic kidney disease on outcomes of surgical resection for primary colorectal cancer: a retrospective cohort review. Dis Colon Rectum 2012; 55: 948–56. 16 Stafford-Smith M. Heart and kidneys: sharing more than just blood. Curr Opin Anaesthesiol 2007; 20: 65–9. 17 Kearns B, Gallagher H, de Lusignan S. Predicting the prevalence of chronic kidney disease in the English population: a cross-sectional study. BMC Nephrol 2013; 14: 49. 18 Schaeffner ES, Ebert N, Delanaye P et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann Intern Med 2012; 157: 471–81. 19 Inker LA, Schmid CH, Tighiouart H et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012; 367: 20–9. 20 Briganti A, Spahn M, Joniau S et al. Impact of age and comorbidities on long-term survival of patients with high-risk prostate cancer treated with radical prostatectomy: a multi-institutional competing-risks analysis. Eur Urol 2013; 63: 693–701. 21 Sun M, Trinh QD, Bianchi M et al. A non-cancer-related survival benefit is associated with partial nephrectomy. Eur Urol 2012; 61: 725–31. 22 Brigic A, Fraser C, Sibbons P et al. Individualization of surgical management for early-stage colonic cancer. Colorectal Dis 2011; 13(Suppl 7): 59–62. 23 Tan KY, Kawamura YJ, Tokomitsu A et al. Assessment for frailty is useful for predicting morbidity in elderly patients undergoing colorectal cancer resection whose comorbidities are already optimized. Am J Surg 2012; 204: 139–43. 24 Leung E, McArdle K, Wong LS. Risk-adjusted scoring systems in colorectal surgery. Int J Surg 2011; 9: 130–5.
Colorectal Disease ª 2014 The Association of Coloproctology of Great Britain and Ireland. 16, 879–887
885
Copyright of Colorectal Disease is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
