Better knowledge of independent risk factors might decrease mortality and morbidity rates following colorectal surgery.
Prospective multicenter study.
From June to September 2002, consecutive patients undergoing open or laparoscopic surgery (electively or on an emergent basis) for colorectal cancers or diverticular disease were prospectively included. Exclusion criteria were colectomy for other causes (eg, inflammatory bowel diseases, benign polyps). The structured sheet of data collection included more than 200 items on all perioperative data concerning the patient, the disease, and the operating surgeons. Postoperative mortality and morbidity were defined as in-hospital death and complications.
Among 1421 patients, the in-hospital death rate was 3.4% and the overall morbidity rate was 35%. Four independent preoperative risk factors of mortality were found: emergency surgery, loss of more than 10% of weight, neurological comorbidity, and age older than 70 years. Six independent risk factors of morbidity were found: age older than 70 years, neurologic comorbidity, hypoalbuminemia, cardiorespiratory comorbidity, long duration of operation, and peritoneal contamination.
Colorectal resection in France is associated with a 3.4% mortality rate and a 35% morbidity rate. Knowledge of the risk factors could help surgeons manage cases.
Knowledge of the true frequency of both mortality and morbidity is crucial in planning health care and research and identifying risk factors. Despite improvement in surgical technique, bowel preparation, and prophylactic antibiotics, colorectal surgery was associated with a 5% to 6% mortality rate and a 20% to 40% morbidity rate.1-3 However, most studies about outcomes after potential colorectal resection have been retrospective analyses from highly specialized centers and often lack detailed information on how preoperative variables affected morbidity and mortality risks. Furthermore, few data have been reported so far from population-based studies that would more accurately reflect the situation in nonspecialized centers over time. To our knowledge, only 5 prospective studies,4-8 including more than 1000 patients, have recently reported either mortality or morbidity rates for colorectal surgery. Except for 1 study,7 no multivariate logistic regression analyses for both mortality and morbidity were performed. Furthermore, no study reported inhospital stay.
In evidence-based medicine, knowledge of the estimated risk associated with an operation might be an important factor to consider when a surgeon and patient consider surgery. Thus, the aim of this national prospective multicenter study was 2-fold:
All members of the Association Française de Chirurgie (French Association for Surgery) were encouraged to participate in a prospective multicenter study exploring both mortality and morbidity after colorectal surgery during a 4-month period (from June to September 2002). All consecutive patients undergoing open or laparoscopic surgery (electively or on an emergent basis) for colorectal cancers or diverticular disease were included from all centers. Exclusion criteria were colectomy for other causes (eg, inflammatory bowel diseases, benign polyps). The structured sheet of data collection included more than 200 items on all perioperative data (Figure).
All patients were followed up during 3 months. Both postoperative mortality and morbidity were defined as in-hospital death and specific complications, respectively. Postoperative morbidity conditions are presented in the Figure.
Statistical analyses used a χ2 test, Mann-Whitney U test, t test, analysis of variance, Kruskal-Wallis H test, and a Pearson correlation or Spearman rank correlation when appropriate. Multivariate logistic regression analyses were also performed to find the main independent factors for postoperative mortality and morbidity. Relative risk (RR) was calculated for each significant variable. Values are expressed as mean ± SD. Statistical significance was stated for P<.05. Data collection and statistical analyses were done using SEM software (SEM Software, Clermont-Ferrand, France).9
Of the 1426 patients who had operations in 81 participating centers, 5 were excluded because data were not available. Thus, 1421 patients were included in the final analysis of both postoperative mortality and morbidity. The mean ± SD inclusion number per center was 17.6 ± 12.8 patients within 4 months (range, 1-58 patients). A total of 690 patients (49%) were treated in teaching (university) hospitals, 463 (33%) in general hospitals, and 268 (18%) in private clinics. A total of 997 patients (70%) had colorectal cancer, and 424 (30%) had diverticular disease. Elective surgery was performed for 1152 patients (81%), and emergency surgery was performed for 269 patients (19%). A total of 652 patients (46%) were older than 69.9 years. Surgical procedures included right colectomy (n = 300), left colectomy (n = 644), total colectomy (n = 74), proctectomy with total mesorectal excision (n = 200), abdominoperineal resection (n = 67), Hartmann procedure (n = 105), and colostomy alone (n = 31).
Forty-eight patients died during their hospitalization, resulting in an in-hospital postoperative mortality rate of 3.4%. Forty-two patients died by day 30, a 2.9% 30-day mortality rate.
Causes of in-hospital deaths were septic shock (n = 11), terminal cancer (n = 6), cardiac failure (n = 5), bronchopneumonia (n = 4), acute respiratory distress (n = 4), myocardial infarction (n = 3), multivisceral failure (n = 3), gastrointestinal hemorrhage (n = 2), stroke syndrome (n = 2), miscellaneous (n = 4), and unknown (n = 4).
Univariate analysis showed 28 variables associated with a higher risk of in-hospital mortality (Table 1). Multivariate analysis found 4 independent factors significantly associated with a higher mortality rate: urgent surgery (RR, 4.41), loss of body weight greater than 10% (RR, 3.85), neurologic comorbidity (RR, 3.43), and age older than 70 years (RR, 2.15). Table 2 presents the variations of mortality rates and RRs according to the lack of these factors or the presence of 1 or more factors. The risk of death was low when there was only 1 or no factor; it was 9% when 2 factors were present, nearly 20% when 3 factors were present, and as high as 50% when all 4 factors were present.
Postoperative morbidity was observed in 495 patients (35%). Complications rates were as follows: wound, 9.6% (including wound dehiscence, wound infection, wound hematoma); clinical anastomotic dehiscence, 4.4% (of which half underwent reoperation); postoperative hemorrhage, 4.5%; stoma complications, 8.6%; prolonged ileus, 5.3%; cardiorespiratory tract, 10.7%; urinary tract, 7.7%; venous thrombosis, 1.2%; neuropsychiatric tract, 2%; postoperative gastrointestinal tract, 1.3%; and miscellaneous, 1.3%. Reoperation was necessary in 27 patients (2%).
Univariate analysis showed 29 variables associated with a higher risk of postoperative morbidity (Table 3). Multivariate analysis found 6 independent factors significantly associated with a higher postoperative morbidity rate: age older than 70 years (RR, 1.02), neurologic comorbidity (RR, 1.58), hypoalbuminemia (RR, 0.55), cardiorespiratory comorbidity (RR, 1.50), mean operation time more than 120 minutes (RR, 1.00), and peritoneal contamination (RR, 1.59).
Mean ± SD hospital stay was 17 ± 13 days. It was significantly lower after elective surgery than after emergent surgery (16 ± 13 vs 20 ± 13 days; P<.001) and after colectomy for diverticular disease than after colectomy for cancer (15 ± 12 vs 18 ± 13 days; P<.001). Hospital stay was also significantly lower after colectomy performed in a private center than after colectomy performed in other centers (15 ± 10 vs 17 ± 14 vs 17 ± 13 days; P = .02).
To our knowledge, this is the first French prospective multicenter study of more than 1400 patients that evaluated postoperative mortality and morbidity after colorectal surgery. This large prospective database provided the opportunity to review the results of large numbers of events over a short period of time and to determine trends. After elective or emergency colorectal surgery for colorectal cancer or diverticular disease, postoperative mortality and morbidity rates were 3.4% and 35%, respectively. These results compare favorably with reported postoperative mortality ranging from 2% to 9% and postoperative morbidity ranging from 18% to 37% (Table 4).1-8,10-16 However, these data are difficult to interpret because of weaknesses in the study design. Among 14 recent studies,1-8,11-16 including more than 1000 consecutive patients who underwent colorectal resection, only 5 were prospective.4-8
Although our mortality rate appears low compared with the literature, our morbidity rate was higher than in other studies. Two reasons could explain this higher rate of postoperative morbidity. First, this multicenter study reflected more accurately the situation observed in our population-based study, in which patients had operations not only in teaching hospitals but also in nonspecialized hospitals. Second, our prospective study allowed us to completely register overall postoperative complications.
To our knowledge, only 9 recent studies in the literature1-4,6,7,15,16 have used multivariate logistic regression analyses to determine the independent risk factors of postoperative mortality. In our study, no relationship was observed between patient outcomes and the surgeon or hospital features (eg, training, certification, experience, and volume of activity). Results concerning the possible relation between volume and outcome in colorectal surgery are controversial. Prystowsky et al3 found that the patient’s outcome improved with the surgeon’s certification and experience. Harmon et al2 observed that higher surgeon volume was associated with significant improvement regardless of the hospital volume. However, medium-volume surgeons achieved results equivalent to those of high-volume surgeons when they operated in high- or medium-volume hospitals. We concluded that hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. As others did,2-4,10,15 we observed that ages equal to or older than 70 years (46% of our patients) were significantly associated with an increased mortality rate. Furthermore, in our study, the risk of death doubled by decade in patients older than 50 years. Except in 2 studies,2,3 we and others found that sex was not associated with a higher mortality rate. Comorbidities, including cardiorespiratory,3,10 neurologic,7 liver diseases,7 corticosteroid,7,10 diabetes mellitus,10 and prior radiotherapy,10 have been reported in several studies as independent risk factors for postoperative mortality. In our study, only neurologic comorbidity (including stroke syndrome and functional status) was an independent risk factor for immediate mortality rate. Although not observed in our study, an American Society of Anesthesiologists score higher than 2 represented another independent risk factor of mortality in 2 studies.7,15 Malnutrition is a risk factor of postoperative mortality. Malnutrition may be evaluated either by a weight loss of more than 10% within the past 6 months, as our study did, or by serum albumin level, as Longo et al did.7 Gibbs et al17 reported that serum albumin level was a strong predictor of mortality, and it is the best prognostic indicator of nutritional status. In our study, weight loss was 1 of the 4 independent risk factors for mortality and hypoalbuminemia 1 of the risk factors of morbidity. Our results emphasized the need for careful preoperative assessment of nutritional status before colectomy. A recent prospective randomized study reported that the administration of a supplemented diet before and after surgery was beneficial to outcomes for patients with cancer who were malnourished and undergoing major elective surgery. Perioperative treatment with immunonutrition seemed to be the best strategy to reduce complications and length of hospital stay.18
Beside the patient characteristics and among the features of surgical procedure, urgent surgery was associated in almost all of the studies with a significantly increased mortality rate. On the other side, no significant difference was observed according to the nature of the disease (colorectal cancer vs diverticular disease), the extent of colorectal resection, and the type of intestinal anastomoses. A recent study showed that laparoscopic colorectal surgery was not associated with a higher rate of postoperative mortality.6 Our univariate analysis showed that laparotomy was associated with a significantly increased rate of mortality. This increased mortality rate after laparotomy could be explained in part by the fact that most of the patients operated on through laparoscopy were operated on electively and for diverticulitis. However, most of the high-risk patients who had emergency operations for colorectal carcinoma were operated on through laparotomy. Furthermore, multivariate analysis failed to show any benefit effect of laparoscopy, probably because of the low number of patients. Thus, laparoscopy cannot be considered a factor of reduced mortality in colorectal surgery. Unlike 3 studies,1,7,10 our study did not find that surgical complications such as anastomotic dehiscence were associated with significantly increased mortality. Whereas anastomotic dehiscence following colorectal surgery represented the first etiology of deaths, accounting for 20% to 33% of all of the deaths,19 6 patients (12%) died with an anastomotic leakage in our study. Most of the postoperative deaths were also caused by postoperative medical complications, including cardiorespiratory and cerebrovascular complications.
Very few studies have used multivariate logistic regression analyses to determine the independent risk factors of postoperative morbidity.3,7 In our study, multivariate analysis showed 6 independent risk factors of postoperative morbidity. Four factors depend on the patient’s characteristics (age older than 70 years, neurologic comorbidity, cardiorespiratory comorbidity, hypoalbuminemia) and 2 factors on the surgical procedure (long duration of operation and fecal contamination). Elderly patients and associated comorbidities (eg, cardiorespiratory, neurologic disorders) were also reported by 2 other studies. Only Prystowsky et al3 observed that the surgeon’s training, certification, and experience were associated with a reduced complication rate. For several authors, serum albumin level is the best prognostic indicator of nutritional status because of its ability to detect protein-energy malnutrition, which is not necessarily accompanied by lower body weight and might not be clinically recognizable. Thus, Gibbs et al17 have observed among 54 125 major noncardiac surgery cases that a decrease in serum albumin from a concentration greater than 46 g/dL to less than 21 g/dL was associated with an increase in morbidity rates from 10% to 65%. Furthermore, albumin level was a better predictor of some types of morbidity, particularly sepsis and major infections, than other types.
Unlike Longo et al,7 we did not observe any other biologic parameters significantly associated with an increased morbidity or mortality rate after colectomy. Among surgical procedures, laparoscopy seems to reduce postoperative morbidity after colorectal surgery. Whereas laparoscopy significantly reduced postoperative morbidity on univariate analysis, multivariate analysis failed to confirm its beneficial effect.
In summary, our population-based prospective study showed that in France, colorectal resection was associated with a 3.4% mortality rate and a 35% morbidity rate. Four independent risk factors, including age older than 70 years, neurologic comorbidity, underweight, and emergency surgery, significantly increased postoperative mortality. The risk of death was low when there was 1 or no risk factor; it was 9% when 2 factors were present, nearly 20% when 3 factors were present, and as high as 50% when all 4 factors were present. Furthermore, our study found 6 independent factors of postoperative morbidity: age older than 70 years, neurologic comorbidity, cardiorespiratory comorbidity, hypoalbuminemia, peritoneal contamination, and operations lasting longer than 120 minutes. Finally, this prospective study suggested the consistent association between the poor status of the patient and a poor outcome. As suggested by other studies, attending to perioperative nutrition could reduce both mortality and morbidity rates.
Correspondence: Yves Panis, MD, PhD, Department of Digestive Surgery, Hôpital Lariboisière, 2 rue Ambroise Paré, 75475 Paris CEDEX 10, France (email@example.com).
Accepted for Publication: September 21, 2004.
EJ In-hospital mortality and associated complications after bowel surgery in Victorian public hospitals. Aust N Z J Surg
2000;706- 10PubMedGoogle ScholarCrossref
et al. Hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg
1999;230404- 413PubMedGoogle ScholarCrossref
JM Patient outcomes for segmental colon resection according to surgeon’s training, certification, and experience. Surgery
2002;132663- 672PubMedGoogle ScholarCrossref
AC Evaluation of POSSUM and P-POSSUM scoring systems in patients undergoing colorectal surgery. Br J Surg
2003;90340- 345PubMedGoogle ScholarCrossref
RJ Population-based audit of colorectal cancer management in two UK health regions. Br J Surg
1997;841731- 1736PubMedGoogle ScholarCrossref
et al. Laparoscopic colorectal anastomosis: risk of postoperative leakage. Results of a multicenter study. Laparoscopic Colorectal Surgery Study Group (LCSSG). Surg Endosc
1999;13639- 644PubMedGoogle ScholarCrossref
et al. Risk factors for morbidity and mortality after colectomy for colon cancer. Dis Colon Rectum
2000;4383- 91PubMedGoogle ScholarCrossref
et al. Relation of surgical volume to outcome in eight common operations: results from the VA National Surgical Quality Improvement Program. Ann Surg
1999;230414- 432PubMedGoogle ScholarCrossref
J SEM: a suitable statistical software adaptated for research in oncology. Bull Cancer
2000;87715- 721PubMedGoogle Scholar
H Multifactorial index of preoperative risk factors in colon resections. Dis Colon Rectum
1992;35117- 122PubMedGoogle ScholarCrossref
G Anastomotic integrity after operations for large-bowel cancer: a multicentre study. BMJ
1980;281411- 414PubMedGoogle ScholarCrossref
CD A population-based study of the incidence, mortality and outcomes in patients following surgery for colorectal cancer in Western Australia. Aust N Z J Surg
2000;7011- 18PubMedGoogle ScholarCrossref
et al. Factors associated with the occurrence of leaks in stapled rectal anastomoses: a review of 1014 patients. J Am Coll Surg
1997;185105- 113PubMedGoogle ScholarCrossref
et al. Postoperative morbidity and mortality following resection of the colon and rectum for cancer. Dis Colon Rectum
1995;38480- 487PubMedGoogle ScholarCrossref
PJ Use of the circular stapler in 1000 consecutive colorectal anastomoses: experience of one surgical team. Surgery
1995;117140- 145PubMedGoogle ScholarCrossref
G Are high-volume surgeons and hospitals the most important predictors of in-hospital outcome for colon cancer resection? Surgery
2002;132268- 273PubMedGoogle ScholarCrossref
SF Preoperative serum albumin level as a predictor of operative mortality and morbidity: results from the National VA Surgical Risk Study. Arch Surg
1999;13436- 42PubMedGoogle ScholarCrossref
V Nutritional approach in malnourished surgical patients: a prospective randomized study. Arch Surg
2002;137174- 180PubMedGoogle ScholarCrossref
P Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis in 707 patients. World J Surg
2002;26499- 502PubMedGoogle ScholarCrossref