Endometrial cancer is the most common gynecologic malignancy in the United States, accounting for an estimated 38,300 cases and 6600 deaths in 2001.1 Since 1988, endometrial cancer has been surgically staged, which in turn has altered the clinical management.2 According to current recommendations, all medically operable patients should undergo total hysterectomy and bilateral salpingo-oophorectomy, and selected patients should have lymph node sampling. Despite the wealth of data available, there is no clear consensus as to which patients should undergo lymph node sampling. Consequently, the treatment strategies for endometrial cancer vary widely in nearly every aspect.
Increasingly, patients with endometrial cancer undergo surgery at community hospitals with involvement of a gynecologic oncologist. For example, in 1997, the Division of Gynecologic Oncology at the State University of New York began providing services to five local community hospitals and a second regional university hospital. In a typical scenario, the patient is seen initially in consultation by the gynecologic oncologist. The surgery is performed at the community hospital with the gynecologic oncologist assisting the gynecologist with the hysterectomy and performing the lymph node sampling, if indicated. The postoperative treatment is supervised by the gynecologist. In contrast, the surgical and postoperative treatment of patients who had surgery at a university hospital is entirely under the supervision of the gynecologic oncologist.
In the current health care environment, there is an increasing demand for performance measures from providers. Such measures will likely become integral components of credentialing and quality assurance processes. This study was undertaken to compare the outcomes of patients with endometrial cancer undergoing primary surgery with gynecologic oncology involvement at university or community hospitals.
MATERIALS AND METHODS
This retrospective review was approved by the Institutional Review Board at the State University of New York at Stony Brook. The study population consisted of all patients who had primary surgery for endometrial cancer with involvement of the attending physicians of the Division of Gynecologic Oncology. Patients with non-adenocarcinoma histology (ie, carcinosarcoma, endometrial stromal sarcoma) and those who had surgery for recurrent disease or complications were excluded. The patients were divided into two groups based on whether their surgery was performed at a university or community hospital. Demographic and clinical data were abstracted from the medical records.
An a priori sample size calculation was performed for several factors, and the largest minimum sample size was chosen. According to discharge records, the average length of stay for primary surgery for endometrial cancer at the State University of New York at Stony Brook in 1997 was 4 days. We were interested in detecting a 1-day difference compared with the stay at a community hospital. With a 3:1 ratio between the number of university and community hospital cases, a two-tailed α value of .05, and a β value of .2, a minimum of 126 patients in the university and 42 patients in the community hospital groups were required.
Statistical analysis was performed using NCSS 2000 (Number Cruncher Statistical Systems, Kaysville, UT) on a Dell Dimension XPS T700r (Dell Computers, Round Rock, TX). Categorical variables (ie, complication rates, performance of vaginal hysterectomy) were analyzed by the χ2 test with Yates correction or Fisher exact test, as appropriate. Continuous variables (ie, age, hospital stay, office-to-surgery interval) were assessed initially for normality numerically by testing skewness, kurtosis, and omnibus normality, and visually with a probability plot. Normally distributed continuous variables were analyzed by the unpaired Student t test or one-way analysis of variance with Bonferroni comparison. Non-normally distributed continuous variables were analyzed by the Mann-Whitney U test. Tests were two tailed when appropriate, and P < .05 was considered significant.
Between September 19, 1997 and August 15, 2001, 366 patients with endometrial cancer met the eligibility criteria. Of these, 295 (75%) patients had surgery at a university hospital, and 71 (25%) patients had surgery at a community hospital.
Table 1 summarizes the demographic data. The mean age for the entire study population was 62 years, and most women (87.8%) were postmenopausal. The university hospital group was significantly older than the community hospital group. The mean Quetelet index for the entire study population was 32.7 kg/m2, and there was no significant difference between the two groups. Most women (82%) had at least one comorbid condition. The most common comorbidity was hypertension (45.9%), followed by diabetes (17.0%). The mean preoperative severity index was significantly higher for the university hospital group than for the community hospital group (14.4 versus 11.8, P = .01). This index weighted 18 comorbid conditions, such as congestive heart failure (2 points), asthma (1 point), and age over 85 years (2 points), to obtain a maximum score of 22 which was normalized to a scale of 100.3
Table 2 summarizes the time intervals. There were no significant differences in the intervals between biopsy and consultation, consultation and surgery, and biopsy and surgery when the two groups were compared.
Table 3 summarizes operative outcome. The two groups were comparable with regard to estimated blood loss and incidence of operative complications. Patients were significantly more likely to have had a vaginal hysterectomy at a university hospital than at a community hospital. Overall, the frequencies of pelvic and paraaortic lymph node sampling were significantly higher in the community hospital group than in the university hospital group. When analyzed by surgical approach (vaginal versus abdominal or laparoscopically assisted), the frequencies of pelvic and paraaortic lymph node sampling were comparable between the community and university hospital groups (79% versus 67.4% and 77.8% versus 65.5%, respectively).
Table 4 summarizes the staging data. The frequency of appropriate surgical staging according to International Federation of Gynecology and Obstetrics (FIGO) criteria was comparable between the two groups. Most patients (71.5–80.3%) had stage I disease. There were no significant differences between the groups when they were analyzed by stage, substage, or early versus advanced stage (stage I versus stages II–IV).
Table 5 summarizes the histology and grade. Most patients (84–91%) had endometrioid histology, followed by papillary serous and clear cell histology. Nearly 50% of the patients had grade 1 tumors, with the remainder equally divided between grade 2 and 3 tumors. There were no significant differences between the two groups with regard to histology or grade.
Table 6 summarizes the hospital data. The two groups were comparable with regard to incidence of hospital complications and length of hospital stay. Patients in the university hospital group were significantly more likely to participate in a research protocol than those in the community hospital group.
According to National Cancer Data Base statistics, approximately 30% of women with endometrial cancer are treated at a community hospital.4 Several studies have cited an association between higher hospital volume and better outcome for patients with cancer.5–7 However, there are few data available to assess the surgical outcome of patients with endometrial cancer treated in these different settings. As the practice of medicine becomes increasingly evidence-based, the demand for these data continues to rise.
Historically, patients with endometrial cancer who were treated in a community hospital were managed without the involvement of a gynecologic oncologist. In 1988, FIGO modified the staging system for endometrial cancer to reflect surgicopathologic information.2 This modification paved the way for surgical staging for all operable patients but added fuel to the controversy over who should treat women with endometrial cancer. Few gynecologists are trained to perform lymph node sampling, so in the community hospital setting a surgical consultant (often a urologist, vascular surgeon, or general surgeon) performs the sampling or it simply is not done.2,8,9
Recently, gynecologic oncologists have expanded their activities to provide services at many community hospitals. Although this expansion has allowed more women with endometrial cancer access to the specialized knowledge provided by gynecologic oncologists, it has also added fuel to the management controversy. The Society of Gynecologic Oncologists has initiated a series of outcome studies to ensure that the care delivered to women with gynecologic cancer in the current highly competitive environment is of the highest quality. The first of these studies successfully “developed a tool for assessing outcomes for the treatment of endometrial cancer.”3
We were interested in evaluating several specific issues for patients undergoing surgery for endometrial cancer at either a community or university hospital. Unfortunately, because of privacy restrictions under the Health Insurance Portability and Accountability Act regulations, we were unable to study patients without gynecologic oncology involvement, so this study was limited to patients with gynecologic oncology involvement.
First, was there a difference in the interval to treatment? We found that the mean interval between the initial consultation with the gynecologic oncologist and surgery (17 days) was identical for patients who had surgery at either a community or university hospital, as was the interval between biopsy and surgery (32 days). This finding was surprising, as we assumed that there was a delay in performing surgery at the community hospitals. This assumption was based on two scheduling difficulties—coordinating the schedules of two physicians from different practices and obtaining operating room time in the absence of block time at the community hospitals. These data suggest that scheduling difficulties were not a substantial impediment, perhaps because the physicians and hospitals recognized the need for expeditious surgical treatment. The 32-day interval between biopsy and surgery was comparable to that previously reported for patients with gynecologic malignancies (20–35 days).10
Second, was there a difference in the patient population? Overall, the patients in the current study were morbidly obese and medically complex. The mean Quetelet index was 32.7 kg/m2 for both groups, comparable to findings in previous reports.3,11 Over 80% of the patients in both groups had at least one concomitant medical condition, most commonly hypertension or diabetes. However, the average age (63 versus 58 years) and the preoperative severity index (14.4 versus 11.8) were significantly higher for the university hospital group. In conjunction with the community obstetrics and gynecology departments, we had developed a triage system (Table 7) to help determine the most appropriate facility and service to treat individual patients. These data reflect the implementation of the triage system and indicate that some patients were preferentially treated at a university hospital because of their medical conditions. As has been suggested, “severity indexing must be performed in considering outcomes data or the results will appear unfavorable in comparison to physicians dealing with a healthier population.”3
Third, was there a difference in the complication rate? Despite their morbid obesity and medical complexity, the patients in both groups had comparable outcomes. The estimated blood loss and complication rates (in the operating room and hospital) were comparable and within acceptable standards.9,11 Similarly, the average hospital stay was similar for patients who were treated at a community or university hospital (3.6 versus 3.9 days, respectively) and was comparable to findings in previous reports (2.3–5.6 days).3,10,11 In both groups, dietary tolerance was the primary determinant of hospital discharge. The patients received a clear liquid diet on the day after surgery and were advanced to solid food as tolerated without waiting for flatus, resulting in short hospital stays. It has been suggested that participation in a managed care program is associated with a significantly shorter hospital stay than with fee-for-service programs.12 This was impossible to determine in the current study, as nearly 70% of the patients in each group were participants in a managed care program, 20% were covered by Medicare, and fewer than 5% were in a fee-for-service program.
Fourth, was there a difference in the incidence of appropriate surgical staging? Despite an abundance of available information, there is no consensus regarding which patients should undergo lymph node sampling. In the current study, we followed the recommendations for lymph node sampling of ACOG, including 1) absence of gross intraperitoneal disease, 2) greater than minimal myometrial invasion, 3) grade 2 or 3 lesions, 4) clear cell or papillary serous histology, or 5) visibly or palpably enlarged lymph nodes.13 On the basis of these criteria, 60–75% of patients had lymph node sampling, and 90% of the patients in both groups were staged appropriately. Patients were considered appropriately staged with regard to lymph node sampling if it was performed or omitted according to these criteria. Although the frequency of lymph node sampling was significantly higher in the community hospital group, this difference disappeared when the surgical approach (vaginal versus abdominal or laparoscopically assisted) was taken into account. Both the severity index (17.6 ± 9.3 versus 14.1 ± 7.6, P = .033) and the Quetelet index (35.8 ± 13.2 versus 32.6 ± 9.5 kg/m2, P = .042) were significantly higher for patients who did not have lymph node sampling despite meeting these criteria. These findings suggest that gynecologic oncologists opted not to perform a lymph node sampling given the risks imposed by these specific patients' comorbid conditions or body habitus.
In previous population-based reviews using data from the National Cancer Data Base, the Surveillance, Epidemiology and End Results Program, and the Health Care Cost and Utilization Project, the incidence of lymph node sampling was remarkably consistent at 33–38%.4,8,9 In contrast, our results paralleled those of the Society of Gynecologic Oncologists Outcomes Task Force, in which a lymph node sampling was performed in 80% of their patients.3 Data from the National Cancer Data Base and the current study (Tables 4 and 5) demonstrated that the distribution of stage and grade was comparable in community and university settings.4 Taken together, these data suggest that gynecologic oncologists, with their specialized knowledge and surgical skills, are more likely to appropriately stage (including lymph node sampling) patients with endometrial cancer when it is technically feasible regardless of the type of facility.
Finally, was there a difference in the surgical approach? Nearly 20% of the patients in the university hospital group had a vaginal hysterectomy, twice that of the community hospital group. A variety of factors appeared to contribute to the higher incidence in the university hospital group. Both the severity index and the Quetelet index were significantly higher for patients who had vaginal hysterectomy than for those who had either laparoscopically assisted vaginal or total abdominal hysterectomy (Table 8). In addition to their specific expertise with treating gynecologic malignancies, the gynecologic oncologists are experienced vaginal surgeons who can utilize this approach to treat selected patients. In contrast, on several occasions, the referring gynecologist specifically rejected the recommendation for a vaginal approach. In fact, the only patients who had vaginal hysterectomy at a community hospital were participants in the Gynecologic Oncology Group protocol LAP-2. This protocol is a phase III randomized study of laparoscopically assisted vaginal hysterectomy-bilateral salpingo-oophorectomy versus total abdominal hysterectomy-bilateral salpingo-oophorectomy plus pelvic and paraaortic lymph node sampling in patients with stage I or II endometrial cancer. Significantly more patients in the university hospital group agreed to participate in LAP-2, accounting for 50% of the patients who had vaginal hysterectomy.
It has been suggested that the vaginal approach is a safe and effective alternative to laparotomy for morbidly obese or medically compromised patients with endometrial cancer. Several studies have shown that the perioperative complication and survival rates for selected patients were comparable to those of patients who had abdominal hysterectomy.14,15 Despite this, a major objection to the vaginal approach has been the inability to perform lymph node sampling and assess extrapelvic sites for metastatic disease. The combination of laparoscopically assisted vaginal hysterectomy and laparoscopic lymph node sampling has been gaining popularity because it appears to decrease morbidity and hospital stay.11 Until long-term data are available, this approach should be utilized only by experienced gynecologic oncologists, ideally within a research protocol such as the Gynecologic Oncology Group LAP-2 study.
The results of this study demonstrate that involvement of a gynecologic oncologist at the time of primary surgery for endometrial cancer was associated with comparable outcomes in both the university and community hospital settings. Further studies are needed to define the outcomes for women with other gynecologic malignancies.
1. American Cancer Society. Cancer facts and figures 2001. Atlanta, GA: American Cancer Society, 2001.
2. Creasman W. New gynecologic cancer staging. Obstet Gynecol 1990;75:287–8.
3. Kennedy AW, Austin Max, Look KY, Munger CB. The Society of Gynecologic Oncologists Outcomes Task Force: Study of endometrial cancer. Gynecol Oncol 2000; 79:379–98.
4. Partridge EE, Shingleton HM, Menck HR. The national cancer data base report on endometrial cancer. J Surg Oncol 1996;61:111–23.
5. Hillner BE, Smith TJ, Desch CE. Hospital and physician volume or specialization and outcomes in cancer treatment: Importance in quality of cancer care. J Clin Oncol 2000;18:2327–40.
6. Roohan PJ, Bickell NA, Baptiste MS, Therriault GD, Ferrara EP, Siu AL. Hospital volume differences and five-year survival from breast cancer. Am J Public Health 1998;88:454–7.
7. Harmon JW, Tang DG, Gordon TA, Bowman HM, Choti MA, Kaufman HS, et al. Hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg 1999;230:404–11.
8. Trimble EL, Kosary C, Park RC. Lymph node sampling and survival in endometrial cancer. Gynecol Oncol 1998; 71:340–3.
9. Brooks SE, Ahn J, Mullins CD, Baguet CR, D'Andrea A. Health care cost and utilization project analysis of comorbid illness and complications for patients undergoing hysterectomy for endometrial cancer. Cancer 2001;92:950–8.
10. Chin S, Harrigill KM. Delay in gynecologic surgical treatment: A comparison of patients in managed care and fee-for-service plans. Obstet Gynecol 1999;93:922–7.
11. Scribner DR, Walker JA, Johnson GA, McMeekin SD, Gold MA, Mannel RS. Laparoscopic pelvic and paraaortic lymph node dissection: Analysis of the first 100 cases. Gynecol Oncol 2001;82:498–503.
12. Miller RH, Luft HS. Managed care performance since 1980. JAMA 1994;271:1512–9.
13. American College of Obstetricians and Gynecologists. Carcinoma of the endometrium. ACOG technical bulletin no. 162. Washington, DC: American College of Obstetricians and Gynecologists, 1991.
14. Massi G, Savino L, Susini T. Vaginal hysterectomy versus abdominal hysterectomy for the treatment of stage I endometrial adenocarcinoma. Am J Obstet Gynecol 1996;174:1320–6.
15. Bloss JD, Berman ML, Bloss LP, Buller RE. Use of vaginal hysterectomy in the management of stage I endometrial cancer in the medically compromised patient. Gynecol Oncol 1991;40:74–7.