https://orcid.org/0000-0001-7512-185X
Abstract
Background
Poor survival rates in different cancer types are sometimes blamed on diagnostic and treatment delays, and it has been suggested that such delays might be related to sociodemographic factors such as education and ethnicity. We examined associations of the wait time from diagnosis to surgery and survival in endometrial cancer (EC) and explored patient and tumour factors influencing the wait time.
Material and methods
In this historical population-based cohort study, The Swedish Quality Registry for Gynaecologic Cancer (SQRGC) was used to identify EC patients who underwent primary surgery between 2010 and 2018. Factors associated with a wait time > 32 d were analysed with logistic regression. The 32-d time point was defined in accordance with the Swedish Standardisation Cancer Care programme. Adjusted Poisson regression analyses were used to analyse excess mortality rate ratio (EMRR).
Results
Out of 7366 women, 5535 waited > 32 d for surgery and 1098 > 70 d. The overall median wait time was 44 d. The factors most strongly associated with a wait time > 32 d were surgery at a university hospital (adjusted odds ratio [OR] 1.34, 95% confidence interval [CI] 1.08–1.66) followed by country of birth (OR 1.31, 95% CI 1.10–1.55) and year of diagnosis. There were no associations between wait time and histology or age. A wait time < 15 d was associated with higher mortality (adjusted EMRR 2.29,95% CI 1.36–3.84) whereas no negative survival impact was seen with a wait time of 70 d. Age, tumour stage, histology and risk group were highly associated with survival, whereas education, country of origin and hospital level did not have any impact on survival.
Conclusions
Surgery within the first two weeks after EC diagnosis was associated with worsened survival. A prolonged wait time did not seem to have any significant adverse effect on prognosis.
Surgery within the first two weeks after diagnosis of endometrial cancer (EC) was associated with poorer survival.
A prolonged wait time to surgery did not worsen prognosis.
Delay in time to surgery was associated with sociodemographic factors.
Highlights
Background
Poor survival rates in different cancer types are sometimes blamed on diagnostic and treatment delays [Citation1] and it has been suggested that such delays might be related to sociodemographic factors [Citation2,Citation3]. A common belief is that the prognosis would improve if delay in cancer treatment could be avoided [Citation4] but there is no international consensus about recommended wait times and the literature is contradictory. The influence of wait times to treatment on survival has been studied in different cancer diseases and the results are ambiguous [Citation5]. Some studies found that longer wait times caused significantly worsened survival in breast, bladder, rectal cancer and melanoma [Citation6–8] whereas no such relation could be shown for some other forms of cancer [Citation9,Citation10].
In endometrial cancer (EC), a treatment delay might theoretically lead to deep myometrial invasion and consequently to both lymph node and distant metastasis. The 5-year survival rate for women diagnosed with an EC confined to the uterus is 90%, but as low as 20% if diagnosed when the cancer has spread distantly, and there is some evidence that long wait times might have a negative impact on survival [Citation6,Citation11–15].
In many countries, both professional societies and authorities have set standards for wait times to treatment in different cancers. Fast tract programmes with the aim of minimising time to surgery have been started in Canada, USA, and Europe [Citation16]. The time to cancer surgery is similarly a highly prioritised health issue in Sweden. In 2017 the Standardisation Cancer Care (SVF) programme was implemented as one of the key commitments of the Swedish Government’s Cancer Strategy. The programme aimed to improve treatment outcomes by standardising the workup as well as shortening the wait time from diagnosis to definitive treatment in different types of cancer. The current recommendation for treatment of EC is to perform the primary surgery within 32 d from diagnosis. Taken together, the evidence for an association between a shorter wait time to surgery and improved oncologic outcome in EC patients is still unclear.
This study aimed to investigate whether the wait time from diagnosis of EC to primary surgery was associated with survival. A secondary aim was to elucidate predictors of delay for primary surgery.
Material and methods
Study design
This is a historical population-based, nationwide cohort study comprising patients with EC in Sweden diagnosed between 2010 and 2018 who had surgery as the primary treatment.
Data sources
The Swedish National Cancer Registry (NCR) was founded in 1958 and reporting is mandatory for both pathologists and clinicians, and the registry has over 95% coverage for all malignant tumours, of which 99% are histologically verified [Citation17]. The Swedish Quality Registry for Gynaecologic Cancer (SQRGC) started registration of EC in 2010. The registration is web-based and includes information on patient and tumour characteristics, treatment details, and follow-up. Reporting to the SQRGC is performed prospectively by all hospitals and clinics in the six Swedish health care regions. Quality control is continuously performed by registrars at the regional cancer centres, who monitor entered data. Through the personal identification numbers allocated to all citizens in Sweden, the SQRGC continuously receives data on date of death from the Population Registry, enabling coverage control compared to the NCR and life-long follow-up of patients. The coverage between the SQRGC and the NCR was verified and showed agreement in 97–100%. The validity of SQRGC data has been assessed, with 70–100% agreement between registered data and the original case files for selected variables [Citation18]. A patient can choose to opt out of registration in the SQRGC.
The following data were extracted from the SQRGC: age at diagnosis, date of diagnosis (date of biopsy by endometrial aspiration or dilation and curettage), date of surgery, type of treating hospital (county hospital or tertiary/university hospital), stage, preoperative risk group), and calendar year at diagnosis.
Sociodemographic data were provided by Statistics Sweden (https://www.scb.se/en/). Education was categorised after the highest level of education into elementary school, secondary school, and college/university; country of birth for subjects and their parents was dichotomised as Sweden or one parent immigrant versus foreign-born or both parents’ immigrants.
Study population
The SQRGC was used to identify patients with EC (ICD-10 code C54) International Federation of Gynaecology and Obstetrics (FIGO) stages I–IV diagnosed from 1 January 2010 through 31 December 2018. Inclusion criteria were: age ≥ 18 years, histologically verified primary EC with a histology of endometrioid-, serous-, mucinous-, clear cell carcinomas, or carcinosarcoma with surgery as primary treatment (n = 13,300). Exclusion criteria consisted of age not registered, histology not specified, missing date of surgery, having the hysterectomy on or before the diagnosis date, primary surgery with palliative intention, hysterectomy performed more than 1 year after diagnosis, and neoadjuvant chemotherapy.
Staging was performed according to the International Federation of Gynaecology and Obstetrics (FIGO) classification from 2009 [Citation19]. In most patients, the diagnostic procedure (endometrial biopsy) was performed at an outpatient facility. After the pathology report, the patient was referred to the local or tertiary hospital and after additional examinations (vaginal sonography and computer tomography) classified to low- or high-risk group.
Preoperatively, a risk classification into low/high-risk groups was performed based on histology, FIGO grade, and DNA ploidy. High-risk was defined as non-endometrioid histology (serous, clear cell carcinoma, or carcinosarcoma), endometrioid adenocarcinoma FIGO grade 3, or non-diploid tumours. Preoperative low-risk group was defined as endometrioid histology Grade 1−2 and diploid tumours. In preoperative high-risk tumours, a lymphadenectomy of the pelvic and para-aortic regions was recommended in addition to hysterectomy and salpingo-oophorectomy. According to the National Guidelines for Endometrial Cancer 2011, patients in the high-risk group were referred to the tertiary centres [Citation19]. Postoperative high-risk patients in FIGO stages I–II were defined as those with non-endometrioid histology or those with endometrioid histology with two or more risk factors; grade 3, ≥50% myometrial invasion or non-diploid tumour. The postoperatively high-risk patients in FIGO stages I–II were recommended chemotherapy ± brachytherapy and those with positive nodes (FIGO stage III) or no lymphadenectomy were offered chemotherapy ± external radiotherapy. Women with preoperative signs of advanced disease (FIGO stage IV), who were considered operable, were surgically treated with the intension to obtain macroscopic radicality.
Wait times
We categorised wait times based on whether the wait time was within 14 d, 15–32 d, 33–50 d, 51–70 d, or more than 70 d from diagnosis to primary surgery. The 32-d time point was used as a reference because it constitutes the benchmark of wait time in the Swedish Standardisation Cancer Care programme for EC [Citation19]. According to this programme, implemented 2017, the time from endometrial biopsy to the histology report and information to the patient should not exceed 9 d. Farther, the referral to the hospital should occur in 2 d, the additional examinations (CT, MRI, and vaginal sonography), should be performed before the decision about the treatment which should be discussed with the patient within 14 d. The surgery should be performed within 7 d afterwards. It means that the time from the endometrial biopsy to the definitive surgery should not exceed 32 d.
Patients were followed for 5 years after the surgery or to emigration or death, whichever came first. The regional ethical review board in Gothenburg approved the study (Dnr 814-15).
Statistical analyses
Factors associated with a wait time >32 d were analysed with uni- and multivariable logistic regression and presented as odds ratio (OR) with a 95% confidence interval (CI).
Kruskal–Wallis equality-of-populations rank test was used for differences in wait time between groups.
Relative survival (RS) by wait time for all EC patients and according to preoperative risk group was estimated by the Pohar Perme method, which is a measure illustrating the excess mortality in a study population due to the disease, compared to the mortality in the general population. Excess mortality rate ratio (EMRR) was estimated by using uni- and multivariable Poisson regression models. The main independent variable was the wait time to primary surgery, defined as the number of days between date of diagnosis (date of biopsy by endometrial aspiration or dilation and curettage) and definitive surgery date (defined by surgery including at least a hysterectomy).
Covariates from SQGRC included age at diagnosis, level of treating hospital, FIGO stage, preoperative risk group, and year of diagnosis. Included sociodemographic variables from Statistics Sweden were level of education and country of birth for subjects and their parents. A p value of <.05 was considered statistically significant. All statistical analyses were carried out with Stata version 16.1 (Stata Corp, College Station, TX, USA).
Results
We identified 13,300 women in the SQRGC with EC who underwent primary surgery. Patients were excluded according to the exclusion criteria described above and finally 7366 patients were available for analysis (). Of the 7366 women, 6275 had endometrioid adenocarcinoma and 1091 non-endometrioid carcinoma. The median follow-up time was 4.7 years (IQR 2.8–6.6). Patient characteristics by the wait time from diagnosis to primary surgery are shown in . The median wait time to surgery was 44 d and only 24.9% of patients underwent surgery within 32 d from diagnosis. The largest group of patients (37.4%) waited between 33 and 50 d, and 14.9% over 70 d following diagnosis. Women with the lowest education level had a shorter median wait time (42 d) than those with university level education (47 d) (p < .001). Country of birth turned out to impact wait times, with foreign-born women waiting longer than Swedish-born women (48 against 43 d) (p < .001). Women diagnosed and treated at tertiary/university hospitals, had a median wait time of 50 d compared with 36 d in county hospitals and 49 d for those referred to university hospitals. Only 19.1% of women diagnosed at the university hospital underwent surgery within 32 d comparing to 37.3% treated at the county hospital. Among women diagnosed at the county hospital and referred to the university hospital 14.4% underwent surgery within 32 d. These differences are highly significant ().
Figure 1. Flow chart on study cohort.
Table 1. Characteristics of patients with endometrial cancer by wait time from diagnosis to primary surgery.
Table 2. Factors associated with wait time over 32 d from diagnosis to primary surgery.
Wait times varied substantially depending on the calendar year of diagnosis. Between 2010 and 2016 the median wait time increased from 35 to 47 d. In contrast, the median wait time went down from 47 to 41 d between 2016 and 2018 (). When exploring factors associated with a wait time <15 d (144 patients), the majority of patients were treated at county hospitals and diagnosed in the earlier years of the study (2010–2011) ().
Multivariable regression analysis
In the multivariable analysis, the factors most strongly associated with a wait time >32 d were country of birth (OR 1.31, 95% CI 1.10–1.55) and surgery at a tertiary/university hospital. To be referred to the university/tertiary hospital resulted in a significantly higher risk of a wait time >32 d compared to those diagnosed at the university hospital (OR 1.34, 95% CI 1.08–1.66). On the other hand, we found significantly lower risk of a wait time >32 d for surgery at the county hospital (OR 0.37 (95% CI 0.32–0.42). Furthermore, women with higher education had a significantly longer wait time compared with women with the lowest educational level (OR 1.23, 95% CI 1.05–1.44). FIGO stage IV was associated with a shorter wait time to surgery compared with stages I–III (OR 0.67, 95% CI 0.49–0.92). There was no difference in wait time between preoperative high and low-risk groups, between endometrioid and non-endometrioid tumours or between age groups. A significantly higher proportion of women were treated within 32 d in 2018 compared to any other year between 2012 and 2017, but the highest proportion was seen in 2010 (OR 0.68, 95% CI 0.53–0.88) ().
Survival analysis
In the survival analyses, we found that a wait time of 1–14 d was associated with significantly higher adjusted EMRR (2.29, 95% CI 1.36–3.84) compared with a wait time of 15–32 d. On the other hand, we found slightly lower mortality in the group waiting 33–50 d (adjusted EMRR 0.75, 95% CI 0.57–0.97) and no negative impact on survival with a wait time of 70 d. As expected, age, tumour stage, histology, and preoperative risk group were highly associated with survival, whereas education, country of origin, and hospital level did not have any impact on survival (). RS according to the wait time in the whole cohort is shown in . The impact of the wait time on survival was more prominent in the high-risk tumour group () than in the low-risk group ().
Figure 2. Five-year relative survival according to the wait time. CI: confidence interval.
Table 3. Factors associated with excess mortality. Excess mortality rate ratio (EMRR) estimated by uni- and multivariable Poisson regression.
Discussion
This is one of a few studies examining the effect of wait times to EC surgery on survival that has included a large and unselected population-based cohort with data on tumour characteristics and sociodemographic variables.
We noted a poorer survival for women operated within 2 weeks of diagnosis and the difference was more prominent in women with high-risk tumours. We did not see any adverse effect of a wait time of 70 d on RS when adjusting for relevant covariables. A longer wait time was associated with treatment at tertiary/university hospital, especially for patients referred from the county hospital but even those diagnosed and treated at the university hospitals had a significantly higher risk of a wait time >32 d than patients treated at the county hospitals. This fact may indicate that the preoperative work up process is slower at the university hospital compared with the county hospital. However, the prolonged wait time at university hospitals did not seem to impact survival. Other factors associated with significantly longer wait time were lower tumour stage, foreign country of origin, higher education, and specific year of diagnosis.
In a large population-based study of over 9000 women with EC in Ontario, Canada, Elit et al. [Citation11] found a worse overall survival in the group with a wait time to surgery <2 weeks. The authors suggested that it might be associated with a higher proportion of emergent surgeries (11.4%) in this group compared to the whole cohort (2%). Furthermore, Shalowitz et al. [Citation20] found that patients with the shortest time to surgery were more likely to have no insurance, to have advanced disease, to be Afro-Americans, to be diagnosed and treated in hospitals with the lowest case-volume quartile, and not to have undergone lymphadenectomy. In accordance with these studies, we found poorer survival for women with a wait time <15 d, even in the adjusted analyses and when emergency surgeries were excluded. However, in our study we found no difference between women with a wait time <15 d and those waiting longer regarding education level, age, country of birth, and tumour factors. On the other hand, more patients in this subgroup were treated at county hospitals and in the earlier years of the study period. These findings might indicate that the increased risk seen for surgery <2 weeks after adjustment for observable characteristics may indicate inadequacies in preoperative workup, clinical acuity or skills.
Contrasting with our results, some previous studies on large populations have demonstrated that a prolonged time to surgery in EC adversely impacts survival [Citation11,Citation20,Citation21]. Strohl et al. [Citation21] found that a wait time of 6 weeks from diagnosis to surgery had a negative impact on survival in EC. Furthermore, patient characteristics such as race and ethnicity, sociodemographic factors and insurance coverage were all associated with a longer time to surgery. Moreover, Shalowitz et al. [Citation20] demonstrated that delay in surgical treatment was an important risk factor for poorer survival but only in low-risk cancers.
We noted significant variations in the wait time over the study period. The variations coincided with changes in the Swedish National Guidelines over the years. The guidelines (2011) included staging with lymphadenectomy for high-risk tumours, which led to centralisation, and this is probably the explanation for the rise in the wait time during 2012–2016. It seemed that the decision on centralisation was not followed by corresponding allocation of resources to perform the surgery. Yun et al. [Citation5] have also shown that when cancer care is centralised to high-volume centres, it leads to hospital crowding and longer wait times to surgery and treatment. These findings are consistent with other studies [Citation12,Citation13,Citation21,Citation22]. In 2017, the programme of Standardised Cancer Care with a defined goal of primary surgery within 32 d was implemented, and consequently the wait time decreased in 2017 and 2018.
In a study based on the 284,499 patients registered in the National Cancer Database in USA between 2004 and 2013, AlHilli et al. [Citation23] demonstrated that the wait time increased during the study period. The largest impact on wait times was found for the year of diagnosis and stage. In our study, stage was less important than treating hospital, country of birth, and education level. As regards survival in the AlHillis study [Citation23], they found that delay in a wait time >6 weeks was associated with worse overall survival in type I EC in stages I and II but not III and IV. We could not confirm this in our study, possibly because there were very few differences in sociodemographic factors between the patients in our study cohort.
In a previous population-based study based on the Western Swedish Healthcare Region, involving 5579 women diagnosed with EC in the period 1995–2016, it was found that a lower education level was associated with markedly higher stage at diagnosis. It was suggested that this might be caused by patient delay or a longer wait time to surgery [Citation24]. In contrast, in this national study, women with the lowest education level had shorter median wait times than those with university level education.
Interestingly, we found that foreign origin was associated with a longer wait time but neither country of birth nor educational level had any impact on survival. A large proportion of women of foreign origin were born in other Nordic countries with a similar culture and tradition as Sweden and very few were born outside Europe, which could have had an impact on our results. It is also reasonable to believe that the Swedish health care system with public care and no private insurance alternatives for cancer treatment minimises the risk of sociodemographic differences in the surgical wait time. Even though no major impact on oncologic outcome was associated with treatment delay in our cohort, it should be emphasised that waiting for cancer treatment is anxiety-provoking for the patient and might influence patient satisfaction and quality of life [Citation15].
The strengths of our study are the large, nationwide population-based setting with high coverage and prospective registering, which minimised the risk of selection bias. Furthermore, the system of personal identification numbers, which is used to record data for every person living in Sweden, has the advantage that patients could only be lost to follow-up by emigration. Thus, we believe that our large register study including both tumour-specific and sociodemographic confounders in the multivariable analyses adds new knowledge about the impact of the wait time to treatment on survival, and about factors influencing this time. Some limitations should be considered in the present study. In total, 13,300 women with EC registered in the SQRGC underwent primary surgery. The main reason for exclusion was that the date of surgery was not specified (n = 3454). We do not know how many of these women fulfilled eligibility criteria. We do not know whether results differ among these women.
Data on adjuvant treatment were not complete in the SQRGC and thus they were not included in the multivariable analyses; however, since adjuvant treatment is decided postoperatively it could hardly impact on the wait time to surgery. Additionally, certain covariables such as comorbidity and body mass index (BMI) that could have an impact on the wait time and mortality are not registered in SQRGC. Furthermore, we lacked information about individual factors affecting the wait time such as patient’s preferences and extended workups due to risk factors for surgery.
In conclusion, in this large population-based cohort, we found that a prolonged wait time to surgery did not seem to have a significant adverse effect on prognosis. The changes in national guidelines with increased centralisation to university hospitals seem to increase the wait time without a negative impact on survival.
The lack of association between the wait time and survival should not be an incentive to abstain from prioritising this patient population, since a long wait time has a negative impact on other factors such as anxiety and quality of life.
Supplemental Material
Download MS Word (32.9 KB)Disclosure statement
The authors declare no competing interests.
Additional information
Funding
References
- Coleman MP, Forman D, Bryant H, et al. Cancer survival in Australia, Canada, Denmark, Norway, Sweden, and the UK, 1995–2007 (the international cancer benchmarking partnership): an analysis of population-based cancer registry data. Lancet. 2011;377(9760):127–138.
- Coleman MP, Babb P, Sloggett A, et al. Socioeconomic inequalities in cancer survival in England and Wales. Cancer. 2001;91(S1):208–216.
- Jack RH, Davies EA, Møller H. Breast cancer incidence, stage, treatment and survival in ethnic groups in South East England. Br J Cancer. 2009;100(3):545–550.
- Richards MA. The size of the prize for earlier diagnosis of cancer in England. Br J Cancer. 2009;101(2):S125–S129.
- Elit L. Wait times from diagnosis to treatment in cancer. J Gynecol Oncol. 2015;26(4):246–248.
- Yun YH, Kim YA, Min YH, et al. The influence of hospital volume and surgical treatment delay on long-term survival after cancer surgery. Ann Oncol. 2012;23(10):2731–2737.
- Neal RD, Tharmanathan P, France B, et al. Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review. Br J Cancer. 2015;112(1):S92–S107.
- Hanna TP, King WD, Thibodeau S, et al. Mortality due to cancer treatment delay: systematic review and Meta-analysis. BMJ. 2020;371:m4087.
- Symonds RP. Cancer biology may be more important than diagnostic delay. BMJ. 2002;325(7367):774.
- Umezu T, Shibata K, Kajiyama H, et al. Prognostic factors in stage IA-IIA cervical cancer patients treated surgically: does the waiting time to the operation affect survival? Arch Gynecol Obstet. 2012;285(2):493–497.
- Elit LM, O’Leary EM, Pond GR, et al. Impact of wait times on survival for women with uterine cancer. J Clin Oncol. 2014;32(1):27–33.
- Matsuo K, Opper NR, Ciccone MA, et al. Time interval between endometrial biopsy and surgical staging for type I endometrial cancer: association between tumor characteristics and survival outcome. Obstet Gynecol. 2015;125(2):424–433.
- Dolly D, Mihai A, Rimel BJ, et al. A delay from diagnosis to treatment is associated with a decreased overall survival for patients with endometrial cancer. Front Oncol. 2016;6:31.
- Pergialiotis V, Haidopoulos D, Tzortzis AS, et al. The impact of waiting intervals on survival outcomes of patients with endometrial cancer: a systematic review of the literature. Eur J Obstet Gynecol Reprod Biol. 2020;246:1–6.
- Robinson KM, Christensen KB, Ottesen B, et al. Diagnostic delay, quality of life and patient satisfaction among women diagnosed with endometrial or ovarian cancer: a nationwide Danish study. Qual Life Res. 2012;21(9):1519–1525.
- Johnson N, Miles T, Bailey D, et al. Delays in treating endometrial cancer in the South West of England. Br J Cancer. 2011;104(12):1836–1839.
- Barlow L, Westergren K, Holmberg L, et al. The completeness of the Swedish cancer register: a sample survey for year 1998. Acta Oncol. 2009;48(1):27–33.
- Rosenberg P, Kjolhede P, Staf C, et al. Data quality in the Swedish quality register of gynecologic cancer - a Swedish gynecologic cancer group (SweGCG) study. Acta Oncol. 2018;57(3):346–353.
- Swedish National Guidelines for Endometrial Cancer [Internet]. 2017. [cited November 10, 2020]. Available from: https://www.sfog.se/media/117538/natvp_endometricancer_nov2011.pdf
- Shalowitz DI, Epstein AJ, Buckingham L, et al. Survival implications of time to surgical treatment of endometrial cancers. Am J Obstet Gynecol. 2017;216(3):268.e 1–268.e18.
- Strohl AE, Feinglass JM, Shahabi S, et al. Surgical wait time: a new health indicator in women with endometrial cancer. Gynecol Oncol. 2016;141(3):511–515.
- Vandborg MP, Christensen RD, Kragstrup J, et al. Reasons for diagnostic delay in gynecological malignancies. Int J Gynecol Cancer. 2011;21(6):967–974.
- AlHilli MM, Elson P, Rybicki L, et al. Time to surgery and its impact on survival in patients with endometrial cancer: a national cancer database study. Gynecol Oncol. 2019;153(3):511–516.
- Svanvik T, Marcickiewicz J, Sundfeldt K, et al. Sociodemographic disparities in stage-specific incidences of endometrial cancer: a registry-based study in west Sweden, 1995–2016. Acta Oncol. 2019;58(6):845–851.