Background
Older patients are at increased risk of perioperative morbidity and mortality. There are limited data on the safety of a robotic approach in the staging for endometrial cancer.
Objective
We compared outcomes in women undergoing laparotomy or robotic surgical staging for endometrial cancer.
Study Design
Using the Healthcare Cost and Utilization Project National Inpatient Sample database from 2008 through 2010, we abstracted records for patients who had surgery for endometrial cancer with either a robotic approach or laparotomy. Patients were categorized by age (<65 vs ≥65 years and 5-year increments). Medical comorbidity scores were calculated using the Charlson Comorbidity Index. Outcomes included intraoperative/perioperative/medical complications, death, length of stay (LOS), and discharge disposition. Student t and χ 2 tests were used to compare groups and approach. Multiple analysis of variance models were used to compare differences between robotics and laparotomy and age groups.
Results
We identified 16,980 patients who had surgery for endometrial cancer with either a robotic approach (age ≥65 years, n = 1228; age <65 years, n = 1574) or laparotomy (age ≥65 years, n = 5914; age <65 years, n = 8264). Older patients had a higher Charlson Comorbidity Index score at the time of surgery (2.6 vs 2.5, P < .001). In laparotomy cases, intraoperative complication rates were similar (4.1% vs 3.7%, P = .17). Older patients had higher rates of perioperative surgical (20.5% vs 15.4%, P < .001) and medical (23.3% vs 15.5%, P < .001) complications, longer LOS (5.1 vs 4.2 days, P < .001), and lower rates of discharge to home (71.2% vs 90.1%, P < .001). In robotic cases, rates of intraoperative complications were similar (5.9% vs 6.8%, P = .32). Older patients had higher rates of perioperative surgical (8.3% vs 5.2%, P = .001) and medical (12.3% vs 6.7%, P = .001) complications, longer LOS (2.00 vs 1.67 days, P < .001), and lower rates of discharge to home (88.8% vs 96.8%, P < .001). With both approaches, as age increased, perioperative surgical and medical complications also increased in a linear fashion. In a subanalysis of older patients (n = 7142), there were lower rates of perioperative surgical (8.3% vs 20.5%, P < .001) and medical (12.3% vs 23.3%, P < .001) complications, death (0.0% vs 0.8%, P < .001), shorter LOS (2.00 vs 5.13 days, P < .001) and higher rate of discharge to home (88.8% vs 71.2%, P < .001) in robotic compared to laparotomy cases.
Conclusion
Although the risks of surgery increase with age, in patients age ≥65 years, a robotic approach for endometrial cancer appears to be safe given current selection criteria utilized in the United States.
Introduction
In the United States, the number of persons age ≥65 years is projected to double from the current 35 million persons to 70 million in the next 25 years. Of newly diagnosed cancers, 56% occur in this demographic group and result in a disproportionate 71% of cancer deaths. Across many surgical subspecialties, those age >65 years have been shown to be at risk for adverse perioperative outcomes. In patients undergoing laparotomy for endometrial cancer, Wright et al demonstrated that as age increased, the risk of postoperative surgical complications, medical complications, length of stay (LOS), and mortality also increased.
The development of minimally invasive surgery (MIS) techniques such as laparoscopic and robotic surgery has revolutionized the treatment of gynecologic cancer and become the preferred approach for the treatment of apparent early-stage disease. A laparoscopic approach has been shown to decrease perioperative complications, especially in at-risk groups including the elderly and obese. In endometrial cancer, prospective studies and subsequent metaanalyses have been conducted demonstrating MIS is a feasible and safe alternative to laparotomy resulting in improved quality of life with no differences in disease free survival or the ability to detect occult disease. To date, no prospective trials have been published that compared robotics to traditional laparoscopy, but smaller retrospective studies have demonstrated reduced operating times, less blood loss, and fewer conversions to laparotomy, despite higher body mass index. Because of these advantages, the use of these MIS approaches in patients at risk for surgical complications, such as the elderly, are worthy of consideration. In an institutional review of 228 endometrial cancer cases, we explored the impact on morbidity and mortality in patients undergoing either robotic approach or laparotomy. We demonstrated that despite older patients carrying more medical comorbidities and having higher rates of lymph node dissection compared to younger patients, there were no differences in intraoperative complication rates. In those age >65 years, robotic surgery was associated with decreased blood loss, rate of ileus, and LOS. Because of a small sample size, we were unable to detect differences in rare events such as death.
In this study we continue to compare surgical outcomes between young and old patients undergoing surgical staging with either laparotomy or a robotic approach for endometrial cancer in a large nationally representative data set. In particular, we examine the morbidity and mortality associated with intraoperative, perioperative, surgical, and medical complications; LOS; and death.
Materials and Methods
Approval was obtained through the Colorado Multiple Institutional Review Board. The Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) is a database developed by the US Agency for Healthcare Research and Quality as part of a federal-state-industry partnership. HCUP-NIS contains all discharge data from >1000 short-term and nonfederal hospitals each year representing 90% of all hospitals in the United States. These data approximate a 20% stratified sample of US hospitals including academic and community facilities, totaling approximately 8 million visits per year. The HCUP-NIS is the largest all-payer database, containing charge information on all patients, including individuals covered by Medicare, Medicaid, and private insurance, as well as those who are uninsured. Robotic procedure codes were captured for the first time in 2008.
We retrieved hospital discharge data from the HCUP-NIS from Jan. 1, 2008, through Dec. 31, 2010. Relying on International Classification of Diseases, Ninth Revision, Clinical Modification procedure codes, we identified women who underwent a hysterectomy with a diagnosis of endometrial cancer. Traditional laparoscopic and vaginal approaches were excluded. Those patients undergoing a supracervical hysterectomy were excluded.
Variables extracted from the database included demographic information including race and age. Medical comorbidity was calculated using a modification of the Charlson Comorbidity Index (CCI), described and validated by Deyo et al, which represents the sum of a weighted index that takes into account the number and seriousness of preexisting comorbid conditions. Perioperative morbidity was based on International Classification of Diseases, Ninth Revision, Clinical Modification coding and was classified into the following categories as previously reported by Wright et al. Intraoperative complications included bladder, ureteral, intestinal, vascular, and other operative injuries. Perioperative surgical complications included reoperation, postoperative hemorrhage, wound complication, abscess, ileus, and venous thromboembolism (VTE). Postoperative medical complications included neurologic, cardiac, pulmonary, gastrointestinal, renal, and infectious complications as well as need for mechanical ventilation, shock, transfusion, and incontinence. Disposition locations included home without assistance, home with a home health agency, skilled nursing facility, or short-term hospital.
Primary outcomes were perioperative morbidity and mortality. Morbidity was defined as intraoperative complication, postoperative surgical complication, or postoperative medical complication. Secondary outcomes included LOS and return to home at the time discharge. Statistical analysis was performed with software (SPSS Statistics, Version 21; IBM Corp, Armonk, NY). Statistical significance was estimated with χ 2 tests for dichotomous and categorical variables and Student t tests were used to compare continuous variables. Multiple analysis of variance models were used to compare differences between robotics and laparotomy and age groups. Statistical significance was determined by P values < .01.
Materials and Methods
Approval was obtained through the Colorado Multiple Institutional Review Board. The Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) is a database developed by the US Agency for Healthcare Research and Quality as part of a federal-state-industry partnership. HCUP-NIS contains all discharge data from >1000 short-term and nonfederal hospitals each year representing 90% of all hospitals in the United States. These data approximate a 20% stratified sample of US hospitals including academic and community facilities, totaling approximately 8 million visits per year. The HCUP-NIS is the largest all-payer database, containing charge information on all patients, including individuals covered by Medicare, Medicaid, and private insurance, as well as those who are uninsured. Robotic procedure codes were captured for the first time in 2008.
We retrieved hospital discharge data from the HCUP-NIS from Jan. 1, 2008, through Dec. 31, 2010. Relying on International Classification of Diseases, Ninth Revision, Clinical Modification procedure codes, we identified women who underwent a hysterectomy with a diagnosis of endometrial cancer. Traditional laparoscopic and vaginal approaches were excluded. Those patients undergoing a supracervical hysterectomy were excluded.
Variables extracted from the database included demographic information including race and age. Medical comorbidity was calculated using a modification of the Charlson Comorbidity Index (CCI), described and validated by Deyo et al, which represents the sum of a weighted index that takes into account the number and seriousness of preexisting comorbid conditions. Perioperative morbidity was based on International Classification of Diseases, Ninth Revision, Clinical Modification coding and was classified into the following categories as previously reported by Wright et al. Intraoperative complications included bladder, ureteral, intestinal, vascular, and other operative injuries. Perioperative surgical complications included reoperation, postoperative hemorrhage, wound complication, abscess, ileus, and venous thromboembolism (VTE). Postoperative medical complications included neurologic, cardiac, pulmonary, gastrointestinal, renal, and infectious complications as well as need for mechanical ventilation, shock, transfusion, and incontinence. Disposition locations included home without assistance, home with a home health agency, skilled nursing facility, or short-term hospital.
Primary outcomes were perioperative morbidity and mortality. Morbidity was defined as intraoperative complication, postoperative surgical complication, or postoperative medical complication. Secondary outcomes included LOS and return to home at the time discharge. Statistical analysis was performed with software (SPSS Statistics, Version 21; IBM Corp, Armonk, NY). Statistical significance was estimated with χ 2 tests for dichotomous and categorical variables and Student t tests were used to compare continuous variables. Multiple analysis of variance models were used to compare differences between robotics and laparotomy and age groups. Statistical significance was determined by P values < .01.
Results
Baseline description
A total of 16,980 patients were identified in the database that met inclusion criteria. There were 9838 patients age <65 years; 8264 were staged with laparotomy and 1574 were staged robotically. There were a total of 7142 patients age ≥65 years; 5914 were staged with laparotomy and 1228 were staged robotically. Table 1 summarizes the study population. In general, the patients in the older cohort were more likely to be white and have more medical comorbidities. There were no differences in the rates of oophorectomy between the young and old (97.0% vs 97.5%, P = .07). The rates of pelvic or paraaortic lymph node dissection were lower in the older group (72.5% vs 75.6%, P < .001).
| Characteristic | Age <65 y, mean ± SD or % (n = 9838) | Age ≥65 y, mean ± SD or % (n = 7142) | P value, OR (95% CI) a | ||||
|---|---|---|---|---|---|---|---|
| Robotic | Laparotomy | Total | Robotic | Laparotomy | Total | ||
| Demographic | |||||||
| Median age, y | 55.1 ± 7.3 | 54.3 ± 7.8 | 54.4 ± 7.7 | 73.4 ± 6.7 | 73.6 ± 6.7 | 73.6 ± 6.7 | |
| Race | |||||||
| White | 81.4 | 72.1 | 73.6 | 86 | 77.3 | 78.8 | <.001 |
| Black | 5.3 | 11.1 | 10.2 | 4.6 | 11.1 | 10.0 | |
| Hispanic | 6.2 | 9.3 | 8.8 | 3.7 | 6.4 | 5.9 | |
| Asian | 2.4 | 4.1 | 3.9 | 1.1 | 1.7 | 1.6 | |
| Other | 4.7 | 3.4 | 3.6 | 4.7 | 3.4 | 3.6 | |
| CCI score | 2.38 ± 0.76 | 2.5 ± 1.0 | 2.5 ± 0.98 | 2.5 ± 0.83 | 2.6 ± 1.1 | 2.6 ± 1.03 | <.001 |
| Clinical | |||||||
| Obese | 34.6 | 34.6 | 34.6 | 20.1 | 22.8 | 22.3 | <.001 |
| Hysterectomy | 100 | 100 | 100 | 100 | 100 | 100 | |
| Oophorectomy | 97.4 | 96.9 | 97 | 98 | 97.4 | 97.5 | .07, 1.19 (0.99–1.44) |
| Lymph node dissection | 72.7 | 67.7 | 75.6 | 79.2 | 71.1 | 72.5 | <.001, 1.24 (1.14–1.30) |
a Compares total numbers for each characteristic between those ages <65 and ≥65 y.
Comparison between young and old undergoing laparotomy
The older group had higher CCI score at the time of surgery (2.63 ± 1.06 vs 2.51 ± 1.02, P < .001) and was less likely to be obese (22.8% vs 34.6%, P < .001). When the young and old who underwent laparotomy were compared, individual complications and rates of any intraoperative complication were similar (3.7% vs 4.1%, P = .17). Older patients had significantly higher rates of perioperative surgical complications including ileus, VTE, and postoperative hemorrhage, while no differences were noted in rates of reoperation, abscess, and wound complications. Significant increases in rates of perioperative medical complications were seen in all outcome measures except transfusion ( Table 2 ). Rates of death were also significantly higher for the older group (0.8% vs 0.3%, P < .001). In addition, LOS was significantly longer in the older cohort compared to the younger group (5.1 ± 4.9 vs 4.2 ± 4.0, P < .001). There were lower rates of discharge to home in the older group (71.2% vs 90.1%, P < .001) and 13% of patients were discharged to a skilled nursing facility or short-term hospital.
| Variable | Age <65 y, mean ± SD or % (n = 8264) | Age ≥65 y, mean ± SD or % (n = 5914) | P value | OR (95% CI) |
|---|---|---|---|---|
| Clinical characteristics | ||||
| Charlson Comorbidity Index score | 2.51 ± 1.02 | 2.63 ± 1.06 | <.001 | – |
| Obese | 34.6 | 22.8 | <.001 | 0.56 (0.52–0.60) |
| Intraoperative complication | ||||
| Bladder injury | 0.0 | 0.1 | .4 | 1.75 (0.47–6.51) |
| Intestinal injury | 0.2 | 0.3 | .92 | 1.75 (0.91–3.38) |
| Ureteral injury | 1.3 | 1.1 | .5 | .90 (0.66–1.22) |
| Vascular injury | 0.1 | 0.1 | .36 | 1.63 (0.55–4.90) |
| Other operative injury | 2.3 | 2.8 | .083 | 1.21 (0.98–1.50) |
| Any complication | 3.7 | 4.1 | .17 | 1.23 (0.95–1.34) |
| Perioperative surgical complication | ||||
| Reoperation | 0.7 | 0.7 | .83 | 0.96 (0.64–1.44) |
| Abscess | 1.5 | 1.4 | .59 | 0.93 (0.70–1.22) |
| GI | 9.2 | 13.1 | <.001 | 1.49 (1.34–1.66) |
| VTE | 4.1 | 5.6 | <.001 | 1.40 (1.20–1.64) |
| Wound | 1.1 | 1.0 | .555 | 0.90 (0.65–1.26) |
| Postoperative hemorrhage | 1.5 | 2.2 | .002 | 1.46 (1.15–1.89) |
| Any complication | 15.4 | 20.5 | <.001 | 1.42 (1.30–1.55) |
| Perioperative medical complication | ||||
| Neurological | 0.1 | 0.4 | .001 | 2.92 (1.47–5.82) |
| Cardiac | 1.0 | 2.3 | <.001 | 2.38 (1.8–3.14) |
| Pulmonary | 7.4 | 10.2 | <.001 | 1.41 (1.25–1.59) |
| Mechanical ventilation | ||||
| GI bleed | 0.2 | 0.4 | .01 | 2.37 (1.19–4.71) |
| Shock | 0.3 | 0.5 | .03 | 1.87 (1.06–3.29) |
| Transfusion | 11.5 | 12.5 | .76 | 1.10 (0.99–1.22) |
| Urinary | 3.8 | 6.4 | <.001 | 1.76 (1.51–2.06) |
| Incontinence | 0.9 | 2.2 | <.001 | 2.40 (1.81–3.20) |
| Infection | 5.8 | 9.6 | <.001 | 1.73 (1.53–1.97) |
| Any complication | 15.5 | 23.3 | <.001 | 1.66 (1.53–1.81) |
| Disposition | ||||
| Home | 90.1 | 71.2 | <.001 | – |
| Home health | 7.0 | 14.9 | ||
| Skilled nursing facility | 2.4 | 12.6 | ||
| Short term hospital | 0.2 | 0.4 | ||
| Death | 0.3 | 0.8 | <.001 | 2.92 (1.80–4.77) |
| Length of stay, d | 4.2 ± 4.01 | 5.1 ± 4.90 | <.001 | – |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
