Secondary Logo

Journal Logo

The Effect of Centralization of Primary Surgery on Survival in Ovarian Cancer Patients

Tingulstad, Solveig MD; Skjeldestad, Finn Egil MD, PhD; Hagen, Bjørn MD, PhD

ORIGINAL RESEARCH
Free

OBJECTIVE To examine the effect of centralized surgery on overall survival in patients with ovarian cancer and, in particular, patients with advanced disease (stage III/IV).

METHODS In a historical prospective study design, patients referred from community hospitals to a teaching hospital for primary surgery during the 2-year period, 1995–1997, were included as cases. For each referred case, two controls, matched for International Federation of Gynecology and Obstetrics (FIGO) stage and age, were selected among patients who had had primary surgery at the referral hospitals (nonteaching) in the years, 1992–1995. Kaplan–Meier survival curves were computed and tested statistically by the log rank test. Cox proportional hazard model was applied for estimation of prognostic factors of survival.

RESULTS There was no difference in postoperative mortality for stage I/II patients by level of care (community hospitals versus teaching hospital). However, for advanced stage disease (III + IV), the controls had significantly shorter crude survival than patients who had been operated on at the teaching hospital (5-year survival: 4% versus 26%; median survival: 12 months versus 21 months) (P = .01). Multivariable analyses showed that completed chemotherapy and size of residual tumor after primary surgery were independent prognostic factors of survival. Patients optimally operated on at the teaching hospital had significantly lower risk of death compared with all other groups, independently of chemotherapy. This indicates that the extent of cytoreductive surgery and the overall management undertaken in the teaching hospital are significant predictors of improved survival.

CONCLUSION Centralization of primary ovarian cancer surgery in one health region in Norway has improved survival for patients with advanced disease. Patients with apparent advanced ovarian cancer should be referred to a subspecialty unit for primary surgery, and every effort should be made to attain as complete cytoreduction as possible.

Centralized surgery of ovarian cancer improves survival in atients with advanced disease.

Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, Trondheim University Hospital; and Section of Epidemiologic Research, SINTEF, Unimed, Trondheim, Norway.

Address reprint requests to: Solveig Tingulstad, MD, Department of Obstetrics and Gynecology, Trondheim University Hospital, 7006 Trondheim, Norway; E-mail: solveig.tingulstad@medisin.ntnu.no.

Received December 26, 2002. Received in revised form March 28, 2003. Accepted April 17, 2003.

Surgery is crucial to the care of advanced ovarian cancer patients. Several studies have shown that survival in women with ovarian cancer improves when operated by gynecologists rather than general surgeons.1–3 Shorter survival time has been related to the failure of general surgeons' capability to debulk tumor mass.1 Furthermore, gynecologic–oncologists prove better outcome by removing a greater extent of the total tumor volume in comparison with general gynecologists.4 A population-based study from the southwest of England showed that gynecologic–oncologists were significantly more likely to attain optimal cytoreduction among patients with advanced disease when compared with general gynecologists.5 In a recently published meta-analysis of 81 studies on the effect of maximal cytoreductive surgery on survival in patients with advanced ovarian carcinoma, a statistically significant positive correlation between the percentage of maximal cytoreduction obtained in each cohort and median survival time was reported. The authors concluded that consistent referral of patients with apparent advanced ovarian cancer to expert centers for primary surgery may be the most efficient effort currently available for improving overall survival.6 Studies have shown that inappropriate surgery was more likely to be performed in nonteaching hospitals versus teaching hospitals, and that the survival rate was better for patients treated in teaching hospitals compared with nonteaching hospitals.7,8

The aim of the present study was to examine the effect of organized centralization of primary surgery on overall survival in patients with ovarian cancer, with particular emphasis on patients with advanced disease (stage III/IV).

Back to Top | Article Outline

MATERIALS AND METHODS

The Health Region IV in Norway comprises three counties with a total population of 633,000 according to the 2000 census. The region is served by one teaching hospital (regional gynecologic–oncology unit), with three gynecologic–oncologists responsible for all surgery, and seven nonteaching hospitals (community hospitals) with general gynecologists. From January 1995 to January 1997, the Risk of Malignancy Index algorithm9 was evaluated in a prospective study design. Consecutively admitted patients with a pelvic mass and a high Risk of Malignancy Index score (greater than 200) were referred from all seven community hospitals to the regional gynecologic–oncology unit for primary surgery.10

During the 2-year study period, a total of 115 women were diagnosed with primary ovarian cancer in the health region. Eighty-two (72%) patients underwent primary surgery at the teaching hospital, among whom 38 patients had been referred from community hospitals (nonteaching) according to the Risk of Malignancy Index study protocol10 and comprise the cases in the present study. The remaining 44 patients were living in the primary catchment area of the teaching hospital and were not eligible for participation because they were not referred to the teaching hospital. For each referred case, two historical patients at the referring hospital were picked as controls. The controls were matched for International Federation of Gynecology and Obstetrics (FIGO) stage and age (365 years) and had undergone primary surgery for ovarian cancer at the referral hospital during the period 1992–199411 (Table 1). The two controls for each case were selected in a backward consecutive order. Five controls, three with stage II and two with stage IV disease, had to be replaced by controls at a neighboring referral hospital because no match was found at the index hospital.

Table 1

Table 1

One of the investigators (ST) traced and reviewed all medical records at each participating hospital and systematically extracted medical and histopathologic data from the records. The data were transferred to a standardized case report form. Detailed information on type of surgery, size of residual disease at the end of surgery, histologic type and grade, type and number of courses of first-line chemotherapy, hospital stay, and postoperative mortality was abstracted from the medical records at each hospital. Uniform determination of FIGO stage12 was based on the outcome of preoperative diagnostic procedures, the operation note, and the pathology report. All histologic slides were reviewed by a single pathologist13 and classified according to the World Health Organization guidelines14 by histologic type and differentiation. Volume of residual disease after primary surgery, as documented in the operative note, was dichotomized into largest diameter of residual tumor less than 1 cm or 1 cm or greater. First-line chemotherapy was categorized into four groups: platinum alone, platinum in combination with a second drug, alkylating agent alone, or no chemotherapy. Taxanes were not used in first-line chemotherapy during the study period. Completed chemotherapy treatment was defined as five or more courses. Operative mortality was defined as death within 28 days of surgery. The number of days spent in the hospital along with the total number of survival days after primary surgery were recorded for each patient. The patients were followed until death or December 2001, resulting in a median follow-up of 82 months (range 60–147 months). No patients were lost to follow-up.

All statistical analyses were done in Statistical Package for Social Sciences 11.01 (SPSS Inc., Chicago, IL). Pearson χ2 test was used in univariate analyses, whereas Mann–Whitney U test was used for continuous variables (mean, median). Kaplan–Meier survival curves were computed and tested statistically by the log rank test. P values < .05 were considered statistically significant. Survival is presented as overall median survival and 5-year survival rates. Finally, the Cox proportional hazard models were applied for estimation of prognostic factors of survival. The study was approved by The Regional Medical Research Ethics Committee, Central Norway, Trondheim.

Back to Top | Article Outline

RESULTS

The distribution of FIGO stages and age among cases and controls is shown in Table 1. Figures 1–3 show Kaplan–Meier survival curves cases operated on at the teaching hospital and control subjects operated on at nonteaching hospitals by stage I/II, III/IV, and for all stages combined. There were no significant differences in survival between cases and controls for all stages combined or in the subgroup with early-stage disease. However, for advanced stage disease (III + IV), control subjects survived significantly shorter than patients operated on at the teaching hospital (5-year survival: 4% versus 26%, P = .01; median survival: 12 months versus 21 months, P = .01) (Table 2).

Table 2

Table 2

Tingulstad

Tingulstad

Tingulstad

Tingulstad

Tingulstad

Tingulstad

Patient and treatment characteristics in advanced disease (stage III + IV) by treatment setting are shown in Table 3. There were no significant differences in the distribution of substages within FIGO stage III, in histologic type, or differentiation grade between cases and controls. All cases and 54% of the control subjects (25 of 46) underwent the standard minimum required surgical procedure: total abdominal hysterectomy, bilateral salpingo-oophorectomy, and omentectomy. None of the controls underwent pelvic lymphadenectomy, whereas four (17%) cases had this procedure performed (P = .003). Also, intestinal resection, to increase the extent of cytoreduction, was done more frequently at the teaching hospital (among cases) (30%) than at nonteaching hospitals (among control subjects) (9%) (P = .02). Nine of the control subjects underwent relaparotomy at the teaching hospital because of inadequacy of the first operation. Finally, control subjects were less likely to have been optimally cytoreduced during primary surgery as only 24% obtained less than 1-cm residual disease compared with 48% of cases (P = .04). A few more cases received platinum-based chemotherapy compared with controls (P = .2). There was no difference in proportion of patients who completed chemotherapy (received five or more courses) and no difference in the mean number of courses given between cases (mean 5.7, standard deviation [SD] 1.1, range 0–9) and controls (mean 5.2, SD 2.4, range 0–15). Furthermore, for patients with advanced disease, the mean duration of the postoperative hospital stay was 11.2 days for cases and 12.8 days for controls (P = .86). Among those receiving chemotherapy, 55% of the patients operated at the teaching hospital received first course before discharge, compared with only 10% of the patients operated at the nonteaching hospitals (P < .001). The remaining patients received chemotherapy after a period of convalescence. The mean ratio of survival days spent in the hospital was 15% (SD 10.5, median 12, range 1–33) for cases and 28% (SD 16.8, median 17, range 1–75) for controls (P = .04).

Table 3

Table 3

Prognostic factors of survival were assessed in a forward stepwise Cox regression model by entering treating hospital (case/control) as an obligate exposure variable. In a model including histology (epithelial, nonepithelial), differentiation grade (well, moderate, poor, not graded), residual disease at end of surgery (less than 1 cm, 1 cm or greater), platinum therapy (non = 0, received = 1), and completeness of chemotherapy (less than five, five or more courses), only completeness of chemotherapy and residual disease were significant predictors of survival, whereas treating hospital became borderline significant (P = .11). There was no interaction between completeness of chemotherapy and residual disease. However, the analyses revealed a significant interaction between residual disease and treatment hospital (cases/controls) (P = .01), implying that survival varied among cases and controls by size of residual disease. Therefore, we explored a model by entering a combined variable of treating hospital and residual disease: teaching hospital/residual disease less than 1 cm, teaching hospital/residual disease 1 cm or greater, non-teaching hospital/residual disease less than 1 cm, and nonteaching hospital/residual disease 1 cm or greater. The results are shown in Table 4. With reference to those patients who were optimally operated on at the teaching hospital (residual disease less than 1 cm), patients operated on at the teaching hospital who did not reach this optimal surgical goal (left with residual disease 1 cm or greater) had a nine-fold increased risk of death within 5 years. These patients had a similar risk of mortality as patients undergoing primary surgery at a nonteaching hospital; there was no difference in mortality between patients treated in a nonteaching hospital setting by residual disease.

Table 4

Table 4

Back to Top | Article Outline

DISCUSSION

This report evaluates a policy of centralization of primary surgery for ovarian cancer in a population-based setting. The hospitals in the present study were the only ones in the health region where gynecologic surgery was performed (no private hospitals); thus, the study population was unselected. Despite the small sample size, the results show an improved survival among women with advanced disease who were operated on at a gynecologic–oncology subspecialty unit by gynecologic–oncologists compared with matched historical controls who were operated on at nonteaching hospitals by general gynecologists.

The size of residual disease after cytoreductive surgery has been reported by many authors to be the most consistent variable to predict prognosis.6,11,15,16 By combining the variables hospital setting (cases/controls) and residual disease (Table 4), patients at the teaching hospital not reaching treatment goal (residual disease less than 1 cm) had an equivalent mortality as patients treated at a nonteaching hospital regardless of residual tumor size. These data suggest that the overall management undertaken in the teaching hospital, aggressiveness in surgery by complete omentectomy, lymphadenectomy, and, if necessary, intestinal resection are important procedures to obtain a more correct estimation of residual disease. This phenomenon is shown by the difference in survival for the subgroup of patients who attained optimal cytoreduction (less than 1-cm residual tumor) operated on at nonteaching hospitals compared with those operated on at the teaching hospital.

The impact on survival of platinum-containing chemotherapy compared with surgical cytoreduction has been an issue of discussion.1,6,17 In our study, the hazard ratio of dying was three times higher for stage III/IV patients receiving less than five courses of chemotherapy in comparison with patients receiving five courses or more, independent of size of residual disease. However, there was no significant difference between cases and controls regarding type of chemotherapy or the number of courses of chemotherapy received. This indicates that the difference in survival among cases and controls in advanced stage disease could not be explained by the chemotherapy effect.

Patients who have undergone suboptimal primary surgery at a nonteaching hospital may be offered a second laparotomy at a more specialized oncology unit. In our study, nine of the patients in the control group (20%) were reoperated on such an indication at the teaching hospital, and seven of them obtained optimal cytoreduction (residual disease less than 1 cm). None of these patients received chemotherapy before relaparotomy. The median survival for these nine control patients, however, did not differ from the median survival rate for the whole control group (14 versus 12 months, respectively). This observation could indicate a superior effect of first versus second attempt to optimize cytoreductive surgery.

The need to maintain a critical volume of work to sustain surgical expertise has been a controversial issue.1,2,18,19 Olaiten et al reported that patients treated in hospitals managing fewer than ten ovarian cancer cases per year (P = .02) were less likely to obtain optimal cytoreduction compared with those treated in hospitals with ten or more cases per year.5 Trimbos et al have demonstrated a long learning curve associated with infrequent and complicated radical surgery for gynecologic cancer.20 In our setting, general gynecologists at the nonteaching hospitals operated on average five ovarian cancer patients per year (range 1–9), compared with a volume of 29 operations at the teaching hospital (range 20–41).11

The overall postoperative mortality was low, as no cases and two control subjects died within the first 28 days of surgery. It is noteworthy that the immediate postoperative mortality was not different between the two groups, despite the more extensive surgery performed in the cases. In addition, the fact that the duration of the hospital stay was similar for cases and controls reflects that more extensive surgery in the cases did not result in increased postoperative morbidity. In correspondence with our results, others have reported that radical surgery, including intestinal anastomosis, urologic procedures, and retroperitoneal lymphadenectomy to increase cytoreduction in advanced ovarian cancer surgery, did not increase operative mortality when patients were operated on by trained gynecologic–oncologists.21–23

Cases with advanced disease spent significantly less time of their overall survival time in hospitals compared with control subjects. Both the length of survival and the quality of life are important for cancer patients. Provided that the amount of survival time spent at home is a valuable experience for patients, our results suggest that primary surgery at a teaching hospital may add quality to life in addition to prolongation of survival. Although few studies have directly addressed the effect of surgical cytoreduction on the quality of the patient's life, Blythe and Wahl showed that ovarian cancer patients after extensive debulking surgery (to less than 2-cm residual tumor) enjoyed a better quality of life than patients rendered with tumor size greater than 2 cm.24 Centralization of primary surgery was introduced in the whole Health Region IV, Norway, from 1995 and onwards. To change the current referral practice of ovarian cancer patients in the health region for the benefit of a prospective randomized trial would ethically be questionable. Therefore, a pragmatic approach of applying a historical prospective design is the best scientific achievement in our setting. During the 10-year period 1987–1996, the health service in the health region has been reasonably stable, including surgical staff.11 Additionally, chemotherapy regimens have not been changed over the study period.

We conclude that centralization of ovarian cancer surgery to gynecologic–oncology subspecialty unit has improved survival, and may also have positively influenced the quality of life for patients with advanced disease. Therefore, our data suggest that patients with apparent advanced ovarian cancer should be referred to a subspecialty unit, and every effort should be made to obtain as complete cytoreduction as possible during primary surgery.4–6,11

Back to Top | Article Outline

REFERENCES

1. Junor EJ, Hole DJ, Gillis CR. Management of ovarian cancer: Referral to a multidisciplinary team matters. Br J Cancer 1994;70:363–70.
2. Kehoe S, Powell J, Wilson S, Woodman C. The influence of the operating surgeon's specialisation on patients survival in ovarian carcinoma. Br J Cancer 1994;70:1014–7.
3. Woodman C, Baghdady A, Collins S, Clyma JA. What changes in the organization of cancer services will improve the outcome for women with ovarian cancer? Br J Obstet Gynaecol 1997;104:135–9.
4. Junor EJ, Hole DJ, McNulty L, Mason M, Young J. Specialist gynecologists and survival outcome in ovarian cancer: A Scottish National Study of 1966 patients. Br J Obstet Gynaecol 1999;106:1130–6.
5. Olaitan A, Weeks J, Mocroft A, Smith J, Howe K, Murdoch J. The surgical management of women with ovarian cancer in the south west of England. Br J Cancer 2001;85:1824–30.
6. Bristow RE, Tomacruz RS, Armstrong DK, Trimble EL, Montz FJ. Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: A meta-analysis. J Clin Oncol 2002;20:1248–59.
7. Gillis CR, Hole DJ, Still RM, Davis J, Kaye SB. Medical audit, cancer registration, and survival in ovarian cancer. Lancet 1991;337:611–2.
8. Wolfe CD, Tilling K, Raju KS. Management and survival of ovarian cancer patients in south east England. Eur J Cancer 1997;33:1835–40.
9. Jacobs I, Oram D, Fairbanks J, Turner J, Frost C, Gruszinskas JG. A risk of malignancy index incorporating CA125, ultrasound and menopausal status for the accurate pre-operative diagnosis of ovarian cancer. Br J Obstet Gynaecol 1990;97:922–9.
10. Tingulstad S, Skjeldestad FE, Halvorsen T, Nustad K, Onsrud M. The risk-of-malignancy index to evaluate potential ovarian cancers in local hospitals. Obstet Gynecol 1999;93:448–52.
11. Tingulstad S, Skjeldestad FE, Halvorsen TB, Hagen B. Survival and prognostic factors in patients with ovarian cancer. Obstet Gynecol 2003;101:885–91.
12. Pecorelli S, Benedet JL, Creasman WT, Shephard JH, Pettersson F. Annual report on the results of treatment in gynecological cancer. Vol 23. London: International Federation of Gynecology and Obstetrics, 1998.
13. Tingulstad S, Halvorsen T, Norstein J, Hagen B, Skjeldestad FE. Completeness and accuracy of registration of ovarian cancer in the Cancer Registry of Norway. IJC 2002;98:907–11.
14. Scully RE, in collaboration with Sobin LH and pathologists in five countries. Histological typing of ovarian tumors. 2nd ed. Geneva: World Health Organization, 1999.
15. Griffiths TC, Parker LM, Fuller AF. Rule of cytoreductive surgery in the management of advanced ovarian cancer. Cancer Treat Rep 1979;63:235–40.
16. Hacker NF, Berek JS, Lagassa LD, Nieberg RK, Elashoff RM. Primary cytoreductive surgery for epithelial ovarian cancer. Obstet Gynecol 1983;61:413–20.
17. Clarc TG, Stewart ME, Altman DG, Gabra H, Smyth JF. A prognostic model for ovarian cancer. Br J Cancer 2001; 85:944–52.
18. Nguyen HN, Averette HE, Hoskins W, Penalver M, Sevin BU, Steren A. The impact of physician's specialty on patients' survival. Cancer 1993;72:3663–70.
19. Jacksson S, Murdoch J, Howe K, Bedfod C, Sanders T, Prentice A. The management of cervical carcinoma within the South West region of England. Br J Obstet Gynaecol 1997;104:140–4.
20. Trimbos JB, Hellebrekers BWJ, Kenter GG, Peters LAW, Zwinderman KH. The long learning curve of gynecological cancer surgery: An argument for centralization. Br J Obstet Gynaecol 2000;107:19–22.
21. Berec JS, Hacker NF, Lagasse LD, Leuchter RS. Lower urinary tract resection as part of cytoreductive surgery for ovarian cancer. Gynecol Oncol 1982;13:87–92.
22. Burghardt E, Pickel H, Lahousen M, Stettner H. Pelvic lymphadenectomy in operative treatment of ovarian cancer. Am J Obstet Gynecol 1986;155:315–9.
23. Eisenkop SM, Spirtos NM, Montag TW, Nalick RH, Wang HJ. The impact of subspecialty training on the management of advanced ovarian cancer. Gynecol Oncol 1992;47:203–9.
24. Blythe JG, Wahl TP. Debulking surgery: Does it increase the quality of survival? Gynecol Oncol 1982;14:396–408.
© 2003 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.