- Email: [email protected]
Laparoscopic colorectal surgery in elderly patients: a case-control study of 15 years of experience
The American Journal of Surgery (2011) 201, 531–536
Clinical Science
Laparoscopic colorectal surgery in elderly patients: a case-control study of 15 years of experience Kok-Yang Tan, M.B.B.S.(Melb), M.Med.(Surg), F.R.C.S.a,b, Fumio Konishi, M.D., Ph.D.a,*, Yutaka J. Kawamura, M.D., Ph.D.a, Takafumi Maeda, M.D.a, Junichi Sasaki, M.D., Ph.D.a, Shingo Tsujinaka, M.D.a, Hisanaga Horie, M.D., Ph.D.c a
Department of Surgery, Saitama Medical Centre, Jichi Medical School, 1-847 Amanumacho, Omiyaku, Saitamashi, Saitama, Japan; bDepartment of Surgery, Alexandra Hospital, Singapore, the Republic of Singapore; cDepartment of Surgery, Jichi Medical School, Tochigi, Japan KEYWORDS: Laparoscopic colorectal surgery; Colorectal cancer; Elderly; Outcomes
Abstract INTRODUCTION: The aim of this study was to review the impact of age (ⱖ75 years) on the short-term outcomes of laparoscopic colorectal surgery. METHODS: Three hundred seventy-nine patients under 70 years of age and 91 patients 75 years and older were analyzed. Quantification of comorbidities was performed using the Charlson Weighted Comorbidity Index. Outcome measures were postoperative complications and 30-day mortality. RESULTS: There was no difference in the occurrence of postoperative complications between the younger and older patients. Bivariate analysis revealed that patient age was not a risk factor of major complications (odds ratio ⫽ 1.2; 95% confidence interval, .6–2.3). Although bivariate analysis revealed that older age had a statistically significant odds ratio for 30-day mortality (odds ratio ⫽ 12.8; 95% confidence interval, 1.3–125.4), multivariate analysis revealed that it was a weighted comorbidity index score of 5 or more (P ⫽ .02) and long operative time (P ⫽ .01) that were independent predictors of 30-day mortality and not age per se. CONCLUSIONS: Age is not an independent predictor of morbidity and mortality in laparoscopic colorectal cancer surgery. © 2011 Elsevier Inc. All rights reserved.
In 2007, more than 10% of the population of Japan was above 75 years of age. There are similar trends in the rest of the world. Surgeons are operating on more and more elderly patients.1,2 Although just more than 10 years ago patients over the age of 60 were taken as older patients in colorectal surgery,3 currently, patients 60 to 75 years old are not considered as old. In laparoscopic colorectal surgery, there have been reports of good outcomes in elderly patients.4,5 However, another recently published large study identified age over 75 as an independent predictor of complications.6 Thus, the impact
of old age (age ⱖ75 years) still requires elucidation. There have recently been some comparative studies comparing the perioperative outcomes of older patients with younger patients.7–9 However, these studies used a cutoff age of 70 years and are likely to have relatively few patients over the age of 75. The aim of this study was to review the impact of age (ⱖ75 years) on the shortterm outcomes of laparoscopic colorectal surgery.
Methods * Corresponding author. Tel.: ⫹048-647-2111; fax: ⫹048-648-5166. E-mail address: [email protected] Manuscript received May 17, 2009; revised manuscript August 9, 2009
0002-9610/$ - see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjsurg.2010.01.024
Patients who underwent laparoscopic resection for colorectal cancer between January 1993 and August 2008 at our
532
The American Journal of Surgery, Vol 201, No 4, April 2011
departments were reviewed. There were 595 cases of laparoscopic colorectal resections of which 545 were for colorectal cancer. In our present study, colorectal cancer patients were analyzed. Patients between the ages of 70 and 74 years were excluded in this study to provide a clear difference in age of the comparison groups. After exclusion, 379 patients under 70 years of age and 91 patients age 75 and above were analyzed. Data from our prospectively collected computer database were extracted, and further clinical information was extracted from review of clinical notes. Besides patient demographics and tumor characteristics, individual comorbidities were recorded. Quantification of comorbidities was performed using the Charlson Weighted Comorbidity Index.10 Outcome measures were postoperative complication rates and 30-day mortality rates. Analysis was performed on patient sex, body mass index (BMI), and tumor characteristics to ensure that there was no statistically significant difference of these parameters in patients of the comparative groups. Thus, the patients were matched for sex, BMI, and tumor size and stage. Analysis was then performed comparing comorbidities and outcome measures between patients younger than 70 years and patients 75 years of age and above. Further analysis was performed to elucidate independent predictors of major postoperative complications and 30-day mortality. Bivariate analysis was performed using the chi-square test using SPSS for Windows (version 15.0; SPSS, Inc, Chicago, IL). Results are expressed as odds ratios with 95% confidence intervals. Stepwise logistic regression analysis was used in multivariate analysis to identify parameters that
Table 1 Demographics and tumor characteristics of compared groups
N Male sex T stage T2 T3 T4 Mean tumor size in mm AJCC/UICC TNM stage I II III IV BMI Cardiac disease Diabetes mellitus Chronic lung disease Previous cerebrovascular accident Mean Charlson Weighted Comorbidity Index score
Age ⬍70
Age ⱖ75
379 60%
91 56%
18% 51% 2% 34 ⫾ 19
15% 59% 2% 38 ⫾ 19
43% 26% 29% 2% 23 ⫾ 3 4% 7% 2%
32% 25% 38% 5% 22 ⫾ 3 20% 11% 9%
P .456 .555
.098 .126
.230 ⬍.001 .264 .001
1%
4%
.056
2.2
2.6
⬍.001
Statistically significant variables are shown in italics.
Table 2
Operative details
N Colonic resection (%) Rectal resection (%) History of past surgery that impacts on laparoscopic surgery because of adhesions (%) D3 oncologic dissection (%) Mean blood loss (mL) Mean operative duration (min) Conversion (%)
Age ⬍70
Age ⱖ75
P
379 92 8
91 90 10
.660 .660
19
20%
.891
68 90 ⫾ 141
67 115 ⫾ 149
.327 .091
225 ⫾ 54 7
.901 .462
224 ⫾ 54 8
independently affect outcome. Only factors that were found on bivariate analysis to be statistically significant (P ⬍ .05) were used in the multivariate analysis.
Results Patient demographics and tumor characteristics are shown in Table 1. There was no statistically significant difference in sex, T stage of the lesions, American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC) TNM staging, and patients’ BMI. Patients’ comorbidities are also shown in Table 1. There were significantly more patients with cardiac disease and previous cerebrovascular accidents in the older patients. When the comorbidities were quantified using the Charlson Weighted Comorbidity Index, patients in the older age group had significantly higher index scores. The operative details are shown in Table 2. The sites of operation, the presence of previous surgery relevant to the laparoscopic resection because of adhesions, and the level of lymph node dissection according to the classification of the Japanese Society of Cancer of the Colon and Rectum11 were similar in both groups of patients. There was also no statistically significant difference in operative blood loss, operative duration, and conversion rates. There was no statistically significant difference in the occurrence of postoperative complications between the younger and older patients (Table 3). Encountered complications which were divided into major and minor are also listed in Table 3. However, there was a statistically significant higher 30-day mortality rate in the older patient group (Table 3). Bivariate analysis results for risk factors for major postoperative complications are shown in Table 4. Patient age was not found to be a risk factor of major complications. Although multiple factors were found on bivariate analysis to be risk factors for postoperative major complications,
K.-Y. Tan et al. Table 3
Laparoscopic colorectal surgery
533
Postoperative complications and mortality
N Overall complications (%) Major and minor (%) Major complications (%) 30-d mortality (%) Major complications (%) Anastomotic bleeding (%) Anastomotic leak (%) Intra-abdominal abscess (%) Intestinal obstruction (%) Prolonged ileus (%) Delirium (%) Cardiopulmonary (%) Bleeding gastrointestinal tract (%) Incisional hernia (%) Minor complications (%) Urinary infection (%) Wound infection (%)
Age ⬍70
Age ⱖ75
379
91
23 13 .3 .8 5 1 2 2 0 1 .5
30% 15 3
P
Factor .224 .537 .005
1 7 0 0 2 3 1
1
0 5
1 9
0 9
Male sex Rectal surgery Cardiac disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Blood loss Operative time Mean comorbidity index score
Odds ratio
95% CI
P
2.0 2.4 2.6
1.0–3.8 1.1–5.5 .9–7.0
.044 .032 .062
2.7 1.1 1.0
.5–14.0 .1–60.6 .4–2.9
.230 .604 .881 .044 .353 .137
Statistically significant variables are shown in italics.
Statistically significant variables are shown in italics.
multivariate analysis revealed that in this study only the 3 factors of male sex, rectal surgery, and operative blood loss were independent predictors of postoperative complications (Table 5). In this series of patients, there were only four 30-day mortalities, 1 under the age of 70 and 3 aged 75 and above. The younger patient died of pneumonia after a perioperative cerebrovascular accident. Of the 3 deaths in the older patients, 1 died suddenly secondary to a cardiac event,
Table 4
Table 5 Multivariate analysis for risk factors for major complications
whereas the other 2 died of septic complications after anastomotic leakage. Bivariate analysis results for risk factors for 30-day mortality are shown in Table 6. Older age was found to have a statistically significant odds ratio of nearly 13 for developing 30-day mortality. However, multivariate analysis (Table 7) revealed that it was a weighted comorbidity index score of 5 or more and long operative time that were independent predictors of 30-day mortality and not age per se.
Comments This study on laparoscopic colorectal cancer surgery in elderly patients (ⱖ75 years of age) is the largest compara-
Bivariate analysis for risk factors of major complications
Factor
Odds ratio
95% CI
P
Chi-square test Age ⱖ75 Male sex Rectal surgery Stage 3 and above Relevant previous surgery Cardiac disease Diabetes mellitus Lung disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Student t test
1.2 2.6 3.5 1.0 .9 5.0 2.1 1.0 8.6 13.2 2.4
.6–2.3 1.4–4.9 1.7–7.2 .6–1.7 .4–1.7 2.3–10.7 1.0–4.8 .2–4.5 2.3–33.1 1.2–148.3 1.1–5.1
.537 .002 ⬍.001 .984 .695 ⬍.001 .054 .988 ⬍.001 .007 .025
No complication
Major complication
35 ⫾ 18 82 ⫾ 122 220 ⫾ 53 23 ⫾ 3 2.2 ⫾ .5
36 ⫾ 19 150 ⫾ 229 240 ⫾ 60 23 ⫾ 3 2.6 ⫾ .9
Tumor size (mean in mm) Blood loss (mean in mL) Operative time (mean in min) BMI mean Mean comorbidity index score Statistically significant variables are shown in italics.
.570 ⬍.001 .012 .811 ⬍.001
534
The American Journal of Surgery, Vol 201, No 4, April 2011
Table 6
Bivariate analysis for risk factors of 30-day mortality
Factor
Odds ratio
95% CI
P
Chi-square test Age ⱖ75 Male sex Rectal surgery Stage 3 and above Relevant previous surgery Cardiac disease Diabetes mellitus Lung disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Student t test
12.8 1.0 1.0 2.6 1.4 13.9 1.0 1.0 65.3 463 1.0
1.3–125.4 1.0–1.0 1.0–1.0 .3–25.1 .1–13.6 1.9–102.5 1.0–1.0 1.0–1.0 8.0–532.7 28.9–7413.3 1.0–1.0
.005 .097 .535 .391 .771 .001 .550 .715 ⬍.001 ⬍.001 .540
No mortality
30-day mortality
35 ⫾ 19 90 ⫾ 144 224 ⫾ 54 23 ⫾ 3 2.3 ⫾ .6
38 ⫾ 17 75 ⫾ 119 283 ⫾ 56 24 ⫾ 5 3.5 ⫾ 1.7
Tumor size (mean in mm) Blood loss (mean in mL) Operative time (mean in minutes) BMI mean Mean comorbidity index score
.760 .818 .031 .708 ⬍.001
Statistically significant variables are shown in italics.
tive study to date. Patients between the age of 70 and 74 years were excluded to provide a clearer difference in age for comparison. We did not find laparoscopic surgery in elderly patients to be more difficult; there were no significant differences in the operative time, blood loss, and conversion rates compared with younger patients. Although the older patients had a significantly higher incidence of comorbidities, there was no significant difference in the postoperative complication rates. However, although the mortality rates were low in both groups, there was a statistically significant difference in the 30-day mortality rate compared with the younger group of patients. This is likely secondary to confounding by the incidence of comorbidities in elderly patients. Multivariate analysis for independent risk factors of postoperative major complications revealed that age was not an independent predictor in this study; the independent risk factors were male sex, rectal surgery, and operative blood loss. Although patients in the older age group had a signif-
Table 7 Multivariate analysis for risk factors of 30-day mortality Factor Age ⱖ75 Cardiac disease Previous cerebrovascular accident Comorbidity index ⱖ5 Operative time
Odds ratio 15.8 .2 3.8 1557.4
95% CI .4–669.6 .1–31.5 .1–489.1 4.2–584,085
Statistically significant variables are shown in italics.
P .149 .541 .587 .015 .012
icantly higher mortality rate, multivariate analysis on 30day mortality revealed that it is not age but comorbidities that predict mortality independently together with operative time. The confidence intervals of the analyses for major complications are relatively narrow, and these indicate the statistical reliability of our findings with regards to major complications. However, the confidence intervals for the analysis for 30-day mortality are very wide, and these findings may well be underpowered. This is secondary to our very low mortality rates (ie, too few events), and it is likely that much larger numbers are required for a more reliable statistical analysis for mortality. Ongoing data collection is being performed to validate our current findings in the future. Notwithstanding the limitations from our current data, we can still infer that age alone is not a predictor of major complications after laparoscopic colorectal surgery. Furthermore, the comorbidities are far more important than age in predicting mortality. Thus, an elderly patient with no comorbidities may do better than a younger patient with multiple comorbidities. It is interesting to note that, again in this cohort of patients, it is not the individual comorbidity but rather the quantification of overall comorbidities that is important in predicting perioperative mortality. As such, it was a comorbidity index of 5 and above and not individual comorbidity that predicted 30-day mortality. This finding echoes that of our previous studies on elderly patients.1,2 Although the Charlson Weighted Comorbidity Index score was an independent predictor of mortality in this study, it came short on multivariate analysis for postoperative complications. We believe this is secondary to the selection of patients with less comorbidities for laparoscopic surgery
K.-Y. Tan et al.
Laparoscopic colorectal surgery
during earlier experience, and, thus, there were not enough patients in this cohort with a high comorbidity index score to produce statistical significance. Although Charlson first described this index for the assessment of long-term mortality in elderly patients, this index has subsequently been used in the quantification of comorbidities by physicians and used to predict outcomes after acute medical events like acute stroke.12 Most of the time, it is a physician’s and geriatrician’s tool rather than a surgeon’s tool unless the surgeon has special interest in geriatric patients. Using this index is actually not new to surgery; however, it has been used in transplantation and risk prediction in acute mesenteric ischemia and perforated diverticular disease.13–15 From the findings of this study on mortality and from our findings from our earlier studies,1,2 we have found the Charlson Weighted Comorbidity Index scoring system to be an easy-to-use tool for quantifying comorbidities in elderly patients in the prediction of their outcomes of both morbidity and mortality after colorectal surgery. More complicated models for the prediction of perioperative mortality are available including the Physiology and Operative Severity Score for enUmeration of Mortality and Morbidity (POSSUM) model and the National Surgical Quality Improvement Program (NSQIP) model. These models use complex equations that take into account a number of preoperative and intraoperative parameters. These models may well provide more precise risk prediction using the equations; however, much more rigorous data collection is required and many centers require dedicated nurses just to perform this. These models no doubt allow valid comparisons of outcome. In practice, however, in the elective setting where patients are adequately optimized, there may be very little variation in many of the parameters considered in these 2 models, and, ultimately, the difference may then be just the patients’ comorbidities, which can be quantified by the simpler Charlson Index. There is ongoing prospective collection of these data of all the elderly surgical patients, and these issues will be further elucidated in the future. Another interesting finding from this study is that operative blood loss and operative time are independent predictors of postoperative complications and 30-day mortality, respectively. The findings suggest that prolonged laparoscopic surgery with prolonged pneumoperitoneum and reverse Trendelenburg positioning and excessive blood loss may have physiological impacts on the patient with adverse outcomes. Although having a low conversion rate may be something to be proud about, we advise caution in persisting with a difficult laparoscopic surgery. Early recognition of difficulty and decision to convert may well be the best policy to avoid excessive blood loss and prolonged surgery, and such conversions may not result in increase in morbidity and mortality.16 As seen from our results, conversion was not an independent predictor for postoperative complications and mortality; however, the amount of blood loss and operative time were independent factors.
535 The finding of male sex being a risk factor for postoperative major complications is not new and concurs with previous studies.6 It is also not surprising that laparoscopic rectal surgery was found to be a risk factor of postoperative complications because it is technically more difficult and rectal surgery is associated with higher complication rates even in open surgery.17 However, much more experience has been gained in recent years, and the results will likely improve. A multicenter prospective study is currently under way in Japan looking at complications after laparoscopic rectal surgery at a time when many surgeons have already gained adequate experience,18 and these results will likely be more conclusive with regards to laparoscopic rectal surgery.
Conclusions Age is not an independent predictor of morbidity and mortality in laparoscopic colorectal cancer surgery. Comorbidity index scores are useful in predicting poor outcomes. Excessive blood loss and operative time should be avoided and early conversion considered in laparoscopic colorectal surgery.
References 1. Tan KY, Kawamura Y, Mizokami K, et al. Colorectal surgery in octogenarian patients-outcomes and predictors of morbidity. Int J Colorectal Dis 2008;24:185–9. 2. Tan KY, Chen CM, Ng C, et al. Which octogenarians do poorly after major open abdominal surgery in our Asian population? World J Surg 2006;30:547–52. 3. Reissman P, Agachan F, Wexner SD. Outcome of laparoscopic colorectal surgery in older patients. Am Surg 1996;62:1060 –3. 4. Cheung HY, Chung CC, Fung JT, et al. Laparoscopic resection for colorectal cancer in octogenarians: results in a decade. Dis Colon Rectum 2007;50:1905–10. 5. De Santis L, Frigo F. Laparoscopic colorectal surgery in the elderly. Acta Biol Med 2005;76;Suppl 1:24 – 6. 6. Kirchhoff P, Dincler S, Buchmann P. A multivariate analysis of potential risk factors for intra- and postoperative complications in 1316 elective laparoscopic colorectal procedures. Ann Surg 2008;248: 259 – 65. 7. Chautard J, Alves A, Zalinski S, et al. Laparoscopic colorectal surgery in elderly patients: a matched case-control study in 178 patients. J Am Coll Surg 2008;206:255– 60. 8. Person B, Cera SM, Sands DR, et al. Do elderly patients benefit from laparoscopic colorectal surgery? Surg Endosc 2008;22:401–5. 9. Feng B, Zheng MH, Mao ZH, et al. Clinical advantages of laparoscopic colorectal cancer surgery in the elderly. Aging Clin Exp Res 2006;18:191–5. 10. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373– 83. 11. General rules for clinical and pathological studies on cancer of the colon, rectum and anus. part II. Histopathological classification. Japanese Research Society for Cancer of the Colon and rectum. Jpn J Surg 1983;13:574 –98. 12. Arnold M, Halpern M, Meier N, et al. Age-dependent differences in demographics, risk factors, co-morbidity, etiology, management, and clinical outcome of acute ischemic stroke. J Neurol 2008;255:1503–7.
536 13. Marchena-Gomez J, Acosta-Merida MA, Hemmersbach-Miller M, et al. The age-adjusted Charlson comorbidity index as an outcome predictor of patients with acute mesenteric ischemia. Ann Vasc Surg 2009;23:458 – 64. 14. Heldal K, Hartmann A, Leivestad T, et al. Clinical outcomes in elderly kidney transplant recipients are related to acute rejection episodes rather than pretransplant comorbidity. Transplantation 2009;87:1045–51. 15. Humes DJ, Solaymani-Dodaran M, Fleming KM, et al. A populationbased study of perforated diverticular disease incidence and associated mortality. Gastroenterology 2009;136:1198 –205.
The American Journal of Surgery, Vol 201, No 4, April 2011 16. Gonzalez R, Smith CD, Mason E, et al. Consequences of conversion in laparoscopic colorectal surgery. Dis Colon Rectum 2006;49:197– 204. 17. Guillou PJ, Quirke P, Thorpe H, et al. Short-term endpoints of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial. Lancet 2005;365:1718 –26. 18. Yamamoto S, Yoshimura K, Konishi F, et al. Phase II trial to evaluate laparoscopic surgery for stage 0/I rectal carcinoma. Jpn J Clin Oncol 2008;38:497–500.
Clinical Science
Laparoscopic colorectal surgery in elderly patients: a case-control study of 15 years of experience Kok-Yang Tan, M.B.B.S.(Melb), M.Med.(Surg), F.R.C.S.a,b, Fumio Konishi, M.D., Ph.D.a,*, Yutaka J. Kawamura, M.D., Ph.D.a, Takafumi Maeda, M.D.a, Junichi Sasaki, M.D., Ph.D.a, Shingo Tsujinaka, M.D.a, Hisanaga Horie, M.D., Ph.D.c a
Department of Surgery, Saitama Medical Centre, Jichi Medical School, 1-847 Amanumacho, Omiyaku, Saitamashi, Saitama, Japan; bDepartment of Surgery, Alexandra Hospital, Singapore, the Republic of Singapore; cDepartment of Surgery, Jichi Medical School, Tochigi, Japan KEYWORDS: Laparoscopic colorectal surgery; Colorectal cancer; Elderly; Outcomes
Abstract INTRODUCTION: The aim of this study was to review the impact of age (ⱖ75 years) on the short-term outcomes of laparoscopic colorectal surgery. METHODS: Three hundred seventy-nine patients under 70 years of age and 91 patients 75 years and older were analyzed. Quantification of comorbidities was performed using the Charlson Weighted Comorbidity Index. Outcome measures were postoperative complications and 30-day mortality. RESULTS: There was no difference in the occurrence of postoperative complications between the younger and older patients. Bivariate analysis revealed that patient age was not a risk factor of major complications (odds ratio ⫽ 1.2; 95% confidence interval, .6–2.3). Although bivariate analysis revealed that older age had a statistically significant odds ratio for 30-day mortality (odds ratio ⫽ 12.8; 95% confidence interval, 1.3–125.4), multivariate analysis revealed that it was a weighted comorbidity index score of 5 or more (P ⫽ .02) and long operative time (P ⫽ .01) that were independent predictors of 30-day mortality and not age per se. CONCLUSIONS: Age is not an independent predictor of morbidity and mortality in laparoscopic colorectal cancer surgery. © 2011 Elsevier Inc. All rights reserved.
In 2007, more than 10% of the population of Japan was above 75 years of age. There are similar trends in the rest of the world. Surgeons are operating on more and more elderly patients.1,2 Although just more than 10 years ago patients over the age of 60 were taken as older patients in colorectal surgery,3 currently, patients 60 to 75 years old are not considered as old. In laparoscopic colorectal surgery, there have been reports of good outcomes in elderly patients.4,5 However, another recently published large study identified age over 75 as an independent predictor of complications.6 Thus, the impact
of old age (age ⱖ75 years) still requires elucidation. There have recently been some comparative studies comparing the perioperative outcomes of older patients with younger patients.7–9 However, these studies used a cutoff age of 70 years and are likely to have relatively few patients over the age of 75. The aim of this study was to review the impact of age (ⱖ75 years) on the shortterm outcomes of laparoscopic colorectal surgery.
Methods * Corresponding author. Tel.: ⫹048-647-2111; fax: ⫹048-648-5166. E-mail address: [email protected] Manuscript received May 17, 2009; revised manuscript August 9, 2009
0002-9610/$ - see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjsurg.2010.01.024
Patients who underwent laparoscopic resection for colorectal cancer between January 1993 and August 2008 at our
532
The American Journal of Surgery, Vol 201, No 4, April 2011
departments were reviewed. There were 595 cases of laparoscopic colorectal resections of which 545 were for colorectal cancer. In our present study, colorectal cancer patients were analyzed. Patients between the ages of 70 and 74 years were excluded in this study to provide a clear difference in age of the comparison groups. After exclusion, 379 patients under 70 years of age and 91 patients age 75 and above were analyzed. Data from our prospectively collected computer database were extracted, and further clinical information was extracted from review of clinical notes. Besides patient demographics and tumor characteristics, individual comorbidities were recorded. Quantification of comorbidities was performed using the Charlson Weighted Comorbidity Index.10 Outcome measures were postoperative complication rates and 30-day mortality rates. Analysis was performed on patient sex, body mass index (BMI), and tumor characteristics to ensure that there was no statistically significant difference of these parameters in patients of the comparative groups. Thus, the patients were matched for sex, BMI, and tumor size and stage. Analysis was then performed comparing comorbidities and outcome measures between patients younger than 70 years and patients 75 years of age and above. Further analysis was performed to elucidate independent predictors of major postoperative complications and 30-day mortality. Bivariate analysis was performed using the chi-square test using SPSS for Windows (version 15.0; SPSS, Inc, Chicago, IL). Results are expressed as odds ratios with 95% confidence intervals. Stepwise logistic regression analysis was used in multivariate analysis to identify parameters that
Table 1 Demographics and tumor characteristics of compared groups
N Male sex T stage T2 T3 T4 Mean tumor size in mm AJCC/UICC TNM stage I II III IV BMI Cardiac disease Diabetes mellitus Chronic lung disease Previous cerebrovascular accident Mean Charlson Weighted Comorbidity Index score
Age ⬍70
Age ⱖ75
379 60%
91 56%
18% 51% 2% 34 ⫾ 19
15% 59% 2% 38 ⫾ 19
43% 26% 29% 2% 23 ⫾ 3 4% 7% 2%
32% 25% 38% 5% 22 ⫾ 3 20% 11% 9%
P .456 .555
.098 .126
.230 ⬍.001 .264 .001
1%
4%
.056
2.2
2.6
⬍.001
Statistically significant variables are shown in italics.
Table 2
Operative details
N Colonic resection (%) Rectal resection (%) History of past surgery that impacts on laparoscopic surgery because of adhesions (%) D3 oncologic dissection (%) Mean blood loss (mL) Mean operative duration (min) Conversion (%)
Age ⬍70
Age ⱖ75
P
379 92 8
91 90 10
.660 .660
19
20%
.891
68 90 ⫾ 141
67 115 ⫾ 149
.327 .091
225 ⫾ 54 7
.901 .462
224 ⫾ 54 8
independently affect outcome. Only factors that were found on bivariate analysis to be statistically significant (P ⬍ .05) were used in the multivariate analysis.
Results Patient demographics and tumor characteristics are shown in Table 1. There was no statistically significant difference in sex, T stage of the lesions, American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC) TNM staging, and patients’ BMI. Patients’ comorbidities are also shown in Table 1. There were significantly more patients with cardiac disease and previous cerebrovascular accidents in the older patients. When the comorbidities were quantified using the Charlson Weighted Comorbidity Index, patients in the older age group had significantly higher index scores. The operative details are shown in Table 2. The sites of operation, the presence of previous surgery relevant to the laparoscopic resection because of adhesions, and the level of lymph node dissection according to the classification of the Japanese Society of Cancer of the Colon and Rectum11 were similar in both groups of patients. There was also no statistically significant difference in operative blood loss, operative duration, and conversion rates. There was no statistically significant difference in the occurrence of postoperative complications between the younger and older patients (Table 3). Encountered complications which were divided into major and minor are also listed in Table 3. However, there was a statistically significant higher 30-day mortality rate in the older patient group (Table 3). Bivariate analysis results for risk factors for major postoperative complications are shown in Table 4. Patient age was not found to be a risk factor of major complications. Although multiple factors were found on bivariate analysis to be risk factors for postoperative major complications,
K.-Y. Tan et al. Table 3
Laparoscopic colorectal surgery
533
Postoperative complications and mortality
N Overall complications (%) Major and minor (%) Major complications (%) 30-d mortality (%) Major complications (%) Anastomotic bleeding (%) Anastomotic leak (%) Intra-abdominal abscess (%) Intestinal obstruction (%) Prolonged ileus (%) Delirium (%) Cardiopulmonary (%) Bleeding gastrointestinal tract (%) Incisional hernia (%) Minor complications (%) Urinary infection (%) Wound infection (%)
Age ⬍70
Age ⱖ75
379
91
23 13 .3 .8 5 1 2 2 0 1 .5
30% 15 3
P
Factor .224 .537 .005
1 7 0 0 2 3 1
1
0 5
1 9
0 9
Male sex Rectal surgery Cardiac disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Blood loss Operative time Mean comorbidity index score
Odds ratio
95% CI
P
2.0 2.4 2.6
1.0–3.8 1.1–5.5 .9–7.0
.044 .032 .062
2.7 1.1 1.0
.5–14.0 .1–60.6 .4–2.9
.230 .604 .881 .044 .353 .137
Statistically significant variables are shown in italics.
Statistically significant variables are shown in italics.
multivariate analysis revealed that in this study only the 3 factors of male sex, rectal surgery, and operative blood loss were independent predictors of postoperative complications (Table 5). In this series of patients, there were only four 30-day mortalities, 1 under the age of 70 and 3 aged 75 and above. The younger patient died of pneumonia after a perioperative cerebrovascular accident. Of the 3 deaths in the older patients, 1 died suddenly secondary to a cardiac event,
Table 4
Table 5 Multivariate analysis for risk factors for major complications
whereas the other 2 died of septic complications after anastomotic leakage. Bivariate analysis results for risk factors for 30-day mortality are shown in Table 6. Older age was found to have a statistically significant odds ratio of nearly 13 for developing 30-day mortality. However, multivariate analysis (Table 7) revealed that it was a weighted comorbidity index score of 5 or more and long operative time that were independent predictors of 30-day mortality and not age per se.
Comments This study on laparoscopic colorectal cancer surgery in elderly patients (ⱖ75 years of age) is the largest compara-
Bivariate analysis for risk factors of major complications
Factor
Odds ratio
95% CI
P
Chi-square test Age ⱖ75 Male sex Rectal surgery Stage 3 and above Relevant previous surgery Cardiac disease Diabetes mellitus Lung disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Student t test
1.2 2.6 3.5 1.0 .9 5.0 2.1 1.0 8.6 13.2 2.4
.6–2.3 1.4–4.9 1.7–7.2 .6–1.7 .4–1.7 2.3–10.7 1.0–4.8 .2–4.5 2.3–33.1 1.2–148.3 1.1–5.1
.537 .002 ⬍.001 .984 .695 ⬍.001 .054 .988 ⬍.001 .007 .025
No complication
Major complication
35 ⫾ 18 82 ⫾ 122 220 ⫾ 53 23 ⫾ 3 2.2 ⫾ .5
36 ⫾ 19 150 ⫾ 229 240 ⫾ 60 23 ⫾ 3 2.6 ⫾ .9
Tumor size (mean in mm) Blood loss (mean in mL) Operative time (mean in min) BMI mean Mean comorbidity index score Statistically significant variables are shown in italics.
.570 ⬍.001 .012 .811 ⬍.001
534
The American Journal of Surgery, Vol 201, No 4, April 2011
Table 6
Bivariate analysis for risk factors of 30-day mortality
Factor
Odds ratio
95% CI
P
Chi-square test Age ⱖ75 Male sex Rectal surgery Stage 3 and above Relevant previous surgery Cardiac disease Diabetes mellitus Lung disease Previous cerebrovascular accident Comorbidity index ⱖ5 Conversion Student t test
12.8 1.0 1.0 2.6 1.4 13.9 1.0 1.0 65.3 463 1.0
1.3–125.4 1.0–1.0 1.0–1.0 .3–25.1 .1–13.6 1.9–102.5 1.0–1.0 1.0–1.0 8.0–532.7 28.9–7413.3 1.0–1.0
.005 .097 .535 .391 .771 .001 .550 .715 ⬍.001 ⬍.001 .540
No mortality
30-day mortality
35 ⫾ 19 90 ⫾ 144 224 ⫾ 54 23 ⫾ 3 2.3 ⫾ .6
38 ⫾ 17 75 ⫾ 119 283 ⫾ 56 24 ⫾ 5 3.5 ⫾ 1.7
Tumor size (mean in mm) Blood loss (mean in mL) Operative time (mean in minutes) BMI mean Mean comorbidity index score
.760 .818 .031 .708 ⬍.001
Statistically significant variables are shown in italics.
tive study to date. Patients between the age of 70 and 74 years were excluded to provide a clearer difference in age for comparison. We did not find laparoscopic surgery in elderly patients to be more difficult; there were no significant differences in the operative time, blood loss, and conversion rates compared with younger patients. Although the older patients had a significantly higher incidence of comorbidities, there was no significant difference in the postoperative complication rates. However, although the mortality rates were low in both groups, there was a statistically significant difference in the 30-day mortality rate compared with the younger group of patients. This is likely secondary to confounding by the incidence of comorbidities in elderly patients. Multivariate analysis for independent risk factors of postoperative major complications revealed that age was not an independent predictor in this study; the independent risk factors were male sex, rectal surgery, and operative blood loss. Although patients in the older age group had a signif-
Table 7 Multivariate analysis for risk factors of 30-day mortality Factor Age ⱖ75 Cardiac disease Previous cerebrovascular accident Comorbidity index ⱖ5 Operative time
Odds ratio 15.8 .2 3.8 1557.4
95% CI .4–669.6 .1–31.5 .1–489.1 4.2–584,085
Statistically significant variables are shown in italics.
P .149 .541 .587 .015 .012
icantly higher mortality rate, multivariate analysis on 30day mortality revealed that it is not age but comorbidities that predict mortality independently together with operative time. The confidence intervals of the analyses for major complications are relatively narrow, and these indicate the statistical reliability of our findings with regards to major complications. However, the confidence intervals for the analysis for 30-day mortality are very wide, and these findings may well be underpowered. This is secondary to our very low mortality rates (ie, too few events), and it is likely that much larger numbers are required for a more reliable statistical analysis for mortality. Ongoing data collection is being performed to validate our current findings in the future. Notwithstanding the limitations from our current data, we can still infer that age alone is not a predictor of major complications after laparoscopic colorectal surgery. Furthermore, the comorbidities are far more important than age in predicting mortality. Thus, an elderly patient with no comorbidities may do better than a younger patient with multiple comorbidities. It is interesting to note that, again in this cohort of patients, it is not the individual comorbidity but rather the quantification of overall comorbidities that is important in predicting perioperative mortality. As such, it was a comorbidity index of 5 and above and not individual comorbidity that predicted 30-day mortality. This finding echoes that of our previous studies on elderly patients.1,2 Although the Charlson Weighted Comorbidity Index score was an independent predictor of mortality in this study, it came short on multivariate analysis for postoperative complications. We believe this is secondary to the selection of patients with less comorbidities for laparoscopic surgery
K.-Y. Tan et al.
Laparoscopic colorectal surgery
during earlier experience, and, thus, there were not enough patients in this cohort with a high comorbidity index score to produce statistical significance. Although Charlson first described this index for the assessment of long-term mortality in elderly patients, this index has subsequently been used in the quantification of comorbidities by physicians and used to predict outcomes after acute medical events like acute stroke.12 Most of the time, it is a physician’s and geriatrician’s tool rather than a surgeon’s tool unless the surgeon has special interest in geriatric patients. Using this index is actually not new to surgery; however, it has been used in transplantation and risk prediction in acute mesenteric ischemia and perforated diverticular disease.13–15 From the findings of this study on mortality and from our findings from our earlier studies,1,2 we have found the Charlson Weighted Comorbidity Index scoring system to be an easy-to-use tool for quantifying comorbidities in elderly patients in the prediction of their outcomes of both morbidity and mortality after colorectal surgery. More complicated models for the prediction of perioperative mortality are available including the Physiology and Operative Severity Score for enUmeration of Mortality and Morbidity (POSSUM) model and the National Surgical Quality Improvement Program (NSQIP) model. These models use complex equations that take into account a number of preoperative and intraoperative parameters. These models may well provide more precise risk prediction using the equations; however, much more rigorous data collection is required and many centers require dedicated nurses just to perform this. These models no doubt allow valid comparisons of outcome. In practice, however, in the elective setting where patients are adequately optimized, there may be very little variation in many of the parameters considered in these 2 models, and, ultimately, the difference may then be just the patients’ comorbidities, which can be quantified by the simpler Charlson Index. There is ongoing prospective collection of these data of all the elderly surgical patients, and these issues will be further elucidated in the future. Another interesting finding from this study is that operative blood loss and operative time are independent predictors of postoperative complications and 30-day mortality, respectively. The findings suggest that prolonged laparoscopic surgery with prolonged pneumoperitoneum and reverse Trendelenburg positioning and excessive blood loss may have physiological impacts on the patient with adverse outcomes. Although having a low conversion rate may be something to be proud about, we advise caution in persisting with a difficult laparoscopic surgery. Early recognition of difficulty and decision to convert may well be the best policy to avoid excessive blood loss and prolonged surgery, and such conversions may not result in increase in morbidity and mortality.16 As seen from our results, conversion was not an independent predictor for postoperative complications and mortality; however, the amount of blood loss and operative time were independent factors.
535 The finding of male sex being a risk factor for postoperative major complications is not new and concurs with previous studies.6 It is also not surprising that laparoscopic rectal surgery was found to be a risk factor of postoperative complications because it is technically more difficult and rectal surgery is associated with higher complication rates even in open surgery.17 However, much more experience has been gained in recent years, and the results will likely improve. A multicenter prospective study is currently under way in Japan looking at complications after laparoscopic rectal surgery at a time when many surgeons have already gained adequate experience,18 and these results will likely be more conclusive with regards to laparoscopic rectal surgery.
Conclusions Age is not an independent predictor of morbidity and mortality in laparoscopic colorectal cancer surgery. Comorbidity index scores are useful in predicting poor outcomes. Excessive blood loss and operative time should be avoided and early conversion considered in laparoscopic colorectal surgery.
References 1. Tan KY, Kawamura Y, Mizokami K, et al. Colorectal surgery in octogenarian patients-outcomes and predictors of morbidity. Int J Colorectal Dis 2008;24:185–9. 2. Tan KY, Chen CM, Ng C, et al. Which octogenarians do poorly after major open abdominal surgery in our Asian population? World J Surg 2006;30:547–52. 3. Reissman P, Agachan F, Wexner SD. Outcome of laparoscopic colorectal surgery in older patients. Am Surg 1996;62:1060 –3. 4. Cheung HY, Chung CC, Fung JT, et al. Laparoscopic resection for colorectal cancer in octogenarians: results in a decade. Dis Colon Rectum 2007;50:1905–10. 5. De Santis L, Frigo F. Laparoscopic colorectal surgery in the elderly. Acta Biol Med 2005;76;Suppl 1:24 – 6. 6. Kirchhoff P, Dincler S, Buchmann P. A multivariate analysis of potential risk factors for intra- and postoperative complications in 1316 elective laparoscopic colorectal procedures. Ann Surg 2008;248: 259 – 65. 7. Chautard J, Alves A, Zalinski S, et al. Laparoscopic colorectal surgery in elderly patients: a matched case-control study in 178 patients. J Am Coll Surg 2008;206:255– 60. 8. Person B, Cera SM, Sands DR, et al. Do elderly patients benefit from laparoscopic colorectal surgery? Surg Endosc 2008;22:401–5. 9. Feng B, Zheng MH, Mao ZH, et al. Clinical advantages of laparoscopic colorectal cancer surgery in the elderly. Aging Clin Exp Res 2006;18:191–5. 10. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373– 83. 11. General rules for clinical and pathological studies on cancer of the colon, rectum and anus. part II. Histopathological classification. Japanese Research Society for Cancer of the Colon and rectum. Jpn J Surg 1983;13:574 –98. 12. Arnold M, Halpern M, Meier N, et al. Age-dependent differences in demographics, risk factors, co-morbidity, etiology, management, and clinical outcome of acute ischemic stroke. J Neurol 2008;255:1503–7.
536 13. Marchena-Gomez J, Acosta-Merida MA, Hemmersbach-Miller M, et al. The age-adjusted Charlson comorbidity index as an outcome predictor of patients with acute mesenteric ischemia. Ann Vasc Surg 2009;23:458 – 64. 14. Heldal K, Hartmann A, Leivestad T, et al. Clinical outcomes in elderly kidney transplant recipients are related to acute rejection episodes rather than pretransplant comorbidity. Transplantation 2009;87:1045–51. 15. Humes DJ, Solaymani-Dodaran M, Fleming KM, et al. A populationbased study of perforated diverticular disease incidence and associated mortality. Gastroenterology 2009;136:1198 –205.
The American Journal of Surgery, Vol 201, No 4, April 2011 16. Gonzalez R, Smith CD, Mason E, et al. Consequences of conversion in laparoscopic colorectal surgery. Dis Colon Rectum 2006;49:197– 204. 17. Guillou PJ, Quirke P, Thorpe H, et al. Short-term endpoints of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial. Lancet 2005;365:1718 –26. 18. Yamamoto S, Yoshimura K, Konishi F, et al. Phase II trial to evaluate laparoscopic surgery for stage 0/I rectal carcinoma. Jpn J Clin Oncol 2008;38:497–500.