Decreasing incidence of cancer after liver transplantation-A Nordic population-based study over 3 decades.

Cancer remains one of the most serious long-term complications after liver transplantation (LT). Data for all adult LT patients between 1982 and 2013 were extracted from the Nordic Liver Transplant Registry. Through linkage with respective national cancer-registry data, we calculated standardized incidence ratios (SIRs) based on country, sex, calendar time, and age-specific incidence rates. Altogether 461 cancers were observed in 424 individuals of the 4246 LT patients during a mean 6.6-year follow-up. The overall SIR was 2.22 (95% confidence interval [CI], 2.02-2.43). SIRs were especially increased for colorectal cancer in recipients with primary sclerosing cholangitis (4.04) and for lung cancer in recipients with alcoholic liver disease (4.96). A decrease in the SIR for cancers occurring within 10 years post-LT was observed from the 1980s: 4.53 (95%CI, 2.47-7.60), the 1990s: 3.17 (95%CI, 2.70-3.71), to the 2000s: 1.76 (95%CI, 1.51-2.05). This was observed across age- and indication-groups. The sequential decrease for the SIR of non-Hodgkin lymphoma was 25.0-12.9-7.53, and for nonmelanoma skin cancer 80.0-29.7-10.4. Cancer risk after LT was found to be decreasing over time, especially for those cancers that are strongly associated with immunosuppression. Whether immunosuppression minimization contributed to this decrease merits further study.

Response to Intravenous Glucose-Tolerance Test and Risk of Cancer: A Long-Term Prospective Cohort Study.

BACKGROUND: Impaired glucose regulation, measured with an oral glucose-tolerance test, has been associated with the risk of cancer. Here, we explored whether the response to an intravenous glucose-tolerance test (IVGTT) is associated with the risk of cancer. METHODS: A cohort of 945 healthy men, aged 40-59years in 1972-75, was followed for 40years. An IVGTT was performed at baseline. Blood samples for glucose determinations were drawn immediately before glucose injection and thereafter every 10min for 1h. Associations were assessed with incidence rate ratios (IRR) and Cox models. FINDINGS: Cancer incidence was higher among men with 10-min glucose levels below the median than in men with levels above the median (IRR: 1.5, 95% CI: 1.2-1.9). This association remained significant after adjusting for relevant confounders (HR: 1.6, 95% CI: 1.3-2.1) and when excluding the first 10years of follow-up to minimize the possibility of reverse causality (HR: 1.5, 95% CI: 1.2-2.0). INTERPRETATION: Healthy middle-aged males that responded to an intravenous glucose injection with rapid glucose elimination during the first phase had an elevated risk of cancer during 40years of follow-up. First phase response to a glucose load might be related to cancer development.

Familial risk of non-Hodgkin lymphoma by sex, relationship, age at diagnosis and histology: a joint study from five Nordic countries.

We aimed to estimate stratified absolute (cumulative) and relative (standardized incidence ratios; SIRs) risks of non-Hodgkin lymphoma (NHL) in relatives of NHL patients. A cohort of 169 830 first-degree relatives of 45 406 NHL patients who were diagnosed between 1955 and 2010 in five European countries was followed for cancer incidence. The lifetime (0-79 year) cumulative risk of NHL in siblings of a patient with NHL was 1.6%, which represents a 1.6-fold increased risk (SIR=1.6, 95% confidence interval (CI)=1.2-1.9) over the general population risk. NHL risk among parent-offspring pairs was increased up to 1.4-fold (95% CI=1.3-1.5; lifetime risk 1.4%). The lifetime risk was higher when NHL was diagnosed in a sister (2.5% in her brothers and 1.9% in her sisters) or a father (1.7% in his son). When there were ⩾2 NHL patients diagnosed in a family, the lifetime NHL risk for relatives was 2.1%. Depending on sex and age at diagnosis, twins had a 3.1-12.9% lifetime risk of NHL. Family history of most of the histological subtypes of NHL increased the risk of concordant and some discordant subtypes. Familial risk did not significantly change by age at diagnosis of NHL in relatives. Familial risk of NHL was not limited to early onset cases.

Perinatal characteristics and breast cancer risk in daughters: a Scandinavian population-based study.

The in utero origins of breast cancer are an increasing focus of research. However, the long time period between exposure and disease diagnosis, and the lack of standardized perinatal data collection makes this research challenging. We assessed perinatal factors, as proxies for in utero exposures, and breast cancer risk using pooled, population-based birth and cancer registry data. Birth registries provided information on perinatal exposures. Cases were females born in Norway, Sweden or Denmark who were subsequently diagnosed with primary, invasive breast cancer (n = 1419). Ten controls for each case were selected from the birth registries matched on country and birth year (n = 14,190). Relative risks (RRs) and 95% confidence intervals (CIs) were estimated using unconditional regression models. Breast cancer risk rose 7% (95% CI 2-13%) with every 500 g (roughly 1 s.d.) increase in birth weight and 7% for every 1 s.d. increase in birth length (95% CI 1-14%). The association with birth length was attenuated after adjustment for birth weight, while the increase in risk with birth weight remained with adjustment for birth length. Ponderal index and small- and large-for-gestational-age status were not better predictors of risk than either weight or length alone. Risk was not associated with maternal education or age, gestational duration, delivery type or birth order, or with several pregnancy complications, including preeclampsia. These data confirm the positive association between birth weight and breast cancer risk. Other pregnancy characteristics, including complications such as preeclampsia, do not appear to be involved in later breast carcinogenesis in young women.

Metabolic risk factors and skin cancer in the Metabolic Syndrome and Cancer Project (Me-Can).

BACKGROUND: Little is known about the associations of metabolic aberrations with malignant melanoma (MM) and nonmelanoma skin cancer (NMSC). OBJECTIVES: To assess the associations between metabolic factors (both individually and combined) and the risk of skin cancer in the large prospective Metabolic Syndrome and Cancer Project (Me-Can). METHODS: During a mean follow-up of 12 years of the Me-Can cohort, 1728 (41% women) incident MM, 230 (23% women) fatal MM and 1145 (33% women) NMSC were identified. Most NMSC cases (76%) were squamous cell carcinoma (SCC) (873, 33% women). Hazard ratios (HRs) were estimated by Cox proportional hazards regression for quintiles and standardized z-scores (with a mean of 0 and SD of 1) of body mass index (BMI), blood pressure, glucose, cholesterol, triglycerides and for a combined metabolic syndrome score. Risk estimates were corrected for random error in the measurements. RESULTS: Blood pressure per unit increase of z-score was associated with an increased risk of incident MM cases in men and women [HR 1·17, 95% confidence interval (CI) 1·04-1·31 and HR 1·18, 95% CI 1·03-1·36, respectively] and fatal MM cases among women (HR 2·39, 95% CI 1·58-3·64). In men, all quintiles for BMI above the reference were associated with a higher risk of incident MM. In women, SCC NMSC risk increased across quintiles for glucose levels (P-trend 0·02) and there was a trend with triglyceride concentration (P-trend 0·09). CONCLUSION: These findings suggest that mechanisms linked to blood pressure may be involved in the pathogenesis of MM. SCC NMSC in women could be related to glucose and lipid metabolism.

Gene variations in sex hormone pathways and the risk of testicular germ cell tumour: a case-parent triad study in a Norwegian-Swedish population.

BACKGROUND: Testicular germ cell tumour (TGCT) is the most common cancer in young men, and an imbalance between the estrogen and androgen levels in utero is hypothesized to influence TGCT risk. Thus, polymorphisms in genes involved in the action of sex hormones may contribute to variability in an individual's susceptibility to TGCT. METHODS: We conducted a Norwegian-Swedish case-parent study. A total of 105 single-nucleotide polymorphisms (SNPs) in 20 sex hormone pathway genes were genotyped using Sequenom MassArray iPLEX Gold, in 831 complete triads and 474 dyads. To increase the statistical power, the analysis was expanded to include 712 case singletons and 3922 Swedish controls, thus including triads, dyads and the case-control samples in a single test for association. Analysis for allelic associations was performed with the UNPHASED program, using a likelihood-based association test for nuclear families with missing data, and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. False discovery rate (FDR) was used to adjust for multiple testing. RESULTS: Five genetic variants across the ESR2 gene [encoding estrogen receptor beta (ERß)] were statistically significantly associated with the risk of TGCT. In the case-parent analysis, the markers rs12434245 and rs10137185 were associated with a reduced risk of TGCT (OR = 0.66 and 0.72, respectively; both FDRs <5%), whereas rs2978381 and rs12435857 were associated with an increased risk of TGCT (OR = 1.21 and 1.19, respectively; both FDRs <5%). In the combined case-parent/case-control analysis, rs12435857 and rs10146204 were associated with an increased risk of TGCT (OR = 1.15 and 1.13, respectively; both FDRs <5%), whereas rs10137185 was associated with a reduced risk of TGCT (OR = 0.79, FDR <5%). In addition, we found that three genetic variants in CYP19A1 (encoding aromatase) were statistically significantly associated with the risk of TGCT in the case-parent analysis. The T alleles of the rs2414099, rs8025374 and rs3751592 SNPs were associated with an increased risk of TGCT (OR = 1.30, 1.30 and 1.21, respectively; all FDRs <5%). We found no statistically significant differences in allelic effect estimates between parental inherited genetic variation in the sex hormone pathways and TGCT risk in the offspring, and no evidence of heterogeneity between seminomas and non-seminomas, or between the Norwegian and the Swedish population, in any of the SNPs examined. CONCLUSIONS: Our findings provide support for ERß and aromatase being implicated in the aetiology of TGCT. Exploring the functional role of the TGCT risk-associated SNPs will further elucidate the biological mechanisms involved.

Metabolic syndrome and rare gynecological cancers in the metabolic syndrome and cancer project (Me-Can).

BACKGROUND: Risk factors for rare gynecological cancers are largely unknown. Initial research has indicated that the metabolic syndrome (MetS) or individual components could play a role. MATERIALS AND METHODS: The Metabolic syndrome and Cancer project cohort includes 288,834 women. During an average follow-up of 11 years, 82 vulvar, 26 vaginal and 43 other rare gynecological cancers were identified. Hazard ratios (HRs) were estimated fitting Cox proportional hazards regression models for tertiles and standardized z-scores [with a mean of 0 and a standard deviation (SD) of 1] of body mass index (BMI), blood pressure, glucose, cholesterol, triglycerides and MetS. Risk estimates were corrected for random error in the measurement of metabolic factors. RESULTS: The MetS was associated with increased risk of vulvar [HR 1.78, 95% confidence interval (CI) 1.30-2.41) and vaginal cancer (HR 1.87, 95% CI 1.07-3.25). Among separate MetS components, 1 SD increase in BMI was associated with overall risk (HR 1.43, 95% CI 1.23-1.66), vulvar (HR 1.36, 95% CI 1.11-1.69) and vaginal cancer (HR 1.79, 95% CI 1.30-2.46). Blood glucose and triglyceride concentrations were associated with increased risk of vulvar cancer (HR 1.98, 95% CI 1.10-3.58 and HR 2.09, 95% CI 1.39-3.15, respectively). CONCLUSION: The results from this first prospective study on rare gynecological cancers suggest that the MetS and its individual components may play a role in the development of these tumors.

Familial risks in nervous system tumours: joint Nordic study.

BACKGROUND: Familial nervous system cancers are rare and limited data on familial aspects are available particularly on site-specific tumours. METHODS: Data from five Nordic countries were used to analyse familial risks of nervous system tumours. Standardised incidence ratios (SIRs) were calculated for offspring of affected relatives compared with offspring of non-affected relatives. RESULTS: The total number of patients with nervous system tumour was 63 307, of whom 32 347 belonged to the offspring generation. Of 851 familial patients (2.6%) in the offspring generation, 42 (4.7%) belonged to the families of a parent and at least two siblings affected. The SIR of brain tumours was 1.7 in offspring of affected parents; it was 2.0 in siblings and 9.4 in families with a parent and sibling affected. For spinal tumours, the SIRs were much higher for offspring of early onset tumours, 14.0 for offspring of affected parents and 22.7 for siblings. The SIRs for peripheral nerve tumours were 16.3 in offspring of affected parents, 27.7 in siblings and 943.9 in multiplex families. CONCLUSION: The results of this population-based study on medically diagnosed tumours show site-, proband- and age-specific risks for familial tumours, with implications for clinical genetic counselling and identification of the underlying genes.

Association between serum 25(OH)D and death from prostate cancer.

Based on observations that for certain cancers, mortality varies according to sun exposure, vitamin D has been proposed to influence on disease progression. This study aims to investigate whether serum levels of 25(OH)D are associated with prognosis in patients with prostate cancer. In total, 160 patients with a serum sample in the JANUS serum bank were included. For 123 patients a pre-treatment serum sample was taken, whereas 37 of the patients had received hormone therapy prior to the blood collection. The serum level of 25(OH)D was classified as low (<50 nmol l(-1)), medium (50-80 nmol l(-1)) or high (>80 nmol l(-1)). A Cox proportional hazard regression model was used to assess the association between serum 25(OH)D and cancer mortality. During follow-up, 61 deaths occurred, of whom 52 died of prostate cancer. The median time of follow-up was 44.0 months (range, 1.2-154.6). Serum 25(OH)D at medium or high levels were significantly related to better prognosis (RR 0.33; 95% CI 0.14-0.77, RR 0.16; 95% CI 0.05-0.43) compared with the low level. Analysis restricted to patients receiving hormone therapy gave a stronger association. The serum level of 25(OH)D may be involved in disease progression and is a potential marker of prognosis in patients with prostate cancer.

Birth and parental characteristics and risk of neuroblastoma in a population-based Norwegian cohort study.

In this population-based Norwegian cohort study (2.1 million children), the impact of birth and parental characteristics on the risk of neuroblastoma (178 cases) was evaluated. In children below the age of 18 months, there was an increased neuroblastoma risk among those with congenital malformations and suggestion of increased risk when the mother had pre-eclampsia.

Body size and thyroid cancer in two million Norwegian men and women.

We investigated relations between measured body mass index (BMI) and stature and thyroid cancer (3046 cases) in a large Norwegian cohort of more than two million individuals. The risk of thyroid cancer, especially of the papillary and follicular types, increased moderately with increasing BMI and height in both sexes.

Improved short-term survival for advanced ovarian, tubal, and peritoneal cancer patients operated at teaching hospitals.

The aim of this study was to study the impact of hospital level and surgical skill on short-term survival of advanced ovarian, tubal, and peritoneal cancer patients in a prospective population-based study. All 198 women with a diagnosis of advanced epithelial invasive ovarian, tubal, and peritoneal cancer in Norway who underwent surgery during 2002 were included in this study. The data were derived from notifications to the Norwegian Cancer Registry and from medical, surgical, and histopathologic records. The hospitals were grouped into teaching and nonteaching hospitals (NTH), and the operating physicians were classified according to specialty (specialist gynecologist, gynecologist, and surgeon). The follow-up period was from 455 to 820 days. The short-term survival at 450 days was 79% for women operated at teaching hospitals (TH) and 62% at NTH (P= 0.02). After simultaneous adjustment for seven prognostic factors and residual disease, the risk of death within 600 days at NTH was unchanged compared to TH, hazard ratio 1.83. The women operated on by specialist compared to general gynecologists had a 20% increased short-term survival (P < 0.0001). TH and specialist gynecologists achieved better short-term survival of patients operated for advanced ovarian, tubal, and peritoneal cancer. Centralization and specialization of ovarian cancer surgery might improve the outcome for this patient group.

Surgery and pre-operative irradiation for locally advanced or recurrent rectal cancer in patients over 75 years of age.

OBJECTIVE: Reports of multimodal treatment regimens especially focusing on locally advanced or recurrent rectal cancer in the elderly, aged>75 years, are unavailable. We have tried to identify and evaluate pre- and peri-operative risk factors for morbidity and mortality and outcome after irradiation/surgery regimens in such patients. PATIENTS AND METHODS: Prospective registration of 86 consecutive patients aged>75 years undergoing elective surgery after irradiation 46-50 Gy for either primary locally advanced rectal cancer (n=51) or recurrent rectal cancer (n=35) from January 1991 to August 2003, 51 men and 35 women, median age 78 years (range 75-85 years) in a national cancer hospital. RESULTS: Multivisceral resections were needed in 63% of patients and 70% R0 resections were obtained in locally advanced cases and 46% in recurrent ones. Both in-hospital- and 30-day-mortality was 3.5%. Sixty-two postoperative complications occurred in 38 patients, three of them fatal. Both operation times over 5 h and transfusion of more than 3 SAG were prognostic factors regarding infections. Estimated five-year survival in R0 patients was 46%. Estimated five-year survival for patients with nonmetastatic tumours with locally advanced primary cancer was 29% and for locally recurrent rectal cancer 32%. Old males had a higher mortality rate the first year after surgery than females with only 65% relative survival compared to a matched normal population. The estimated five-year local recurrence rates were 24% for R0 resections and 54% for R1 resections (P=0.434 ns) and 24% and 45% for locally advanced and recurrent rectal cancer (P=0.248 ns), respectively. CONCLUSION: Thorough pre-operative evaluation and preparation and judicious surgery are important for achieving potentially curative treatment with acceptable morbidity in locally advanced and recurrent rectal cancer in patients over 75 years of age. We suggest that these patients should be evaluated and considered for treatment by multidisciplinary teams as younger patients.

Risk factors for testicular cancer--differences between pure non-seminoma and mixed seminoma/non-seminoma?

The origin of testicular germ cell cancer (TGCC) is believed to be carcinoma in situ cells developed in utero. Clinically, TGCCs are divided into two major histological groups, seminomas and non-seminomas, where the latter group includes non-seminomatous TGCCs with seminomatous components (mixed S/NS TGCC). Recent studies, however, have suggested that non-seminomas and mixed S/NS TGCCs could have certain differences in aetiology, and in this study the TGCCs were divided into three, rather than the conventional two histological groups. A large case-control study was undertaken on data on all live-born boys registered in the Medical Birth Registry of Norway during the period 1967-1998 (n=961 396). Among these were 1087 TGCC cases registered in the Cancer Registry of Norway until February 2004. We found several risk factors for TGCC, including low parity, low gestational age, epilepsy and retained placenta. Several of the variables studied seemed to be risk factors for specific histological groups, e.g. parity 0 vs. 2 and low gestational age being associated with increased risk of non-seminomas, but not of mixed S/NS TGCC, and low maternal age being associated with increased risk of mixed S/NS TGCC, but not of non-seminomatous TGCC. Therefore, our results might suggest that non-seminomas and mixed S/NS TGCCs have partially different risk factors, whose associations may be obscured by combining these two histological groups. The histological groups were not significantly different, however. Most of our findings on risk factors for TGCC are in agreement with at least some previous studies. An unexplainable exception is low birth weight being associated with reduced risk of TGCC in our study.

Height and body mass index in relation to cancer of the small intestine in two million Norwegian men and women.

The present study aimed at exploring the relations between body mass index (BMI, (weight in kilograms)/(height in meters)(2)) and stature and cancer of the small intestine (1162 cases) in a large Norwegian cohort (of two million) with measured height and weight. Elevated BMI in males and increasing height in both sexes were associated with a moderately increased risk of cancer of the small intestine.

Early life events and later risk of colorectal cancer: age-period-cohort modelling in the Nordic countries and Estonia.

BACKGROUND: A lowering of colorectal cancer risk for the birth cohorts born around World War II (WWII) has previously been observed in Norway, a country which suffered some 20% caloric restriction during the war. The purpose of the study was to conduct a similar kind of analysis in the other Nordic countries and Estonia, which were also subjected to various degrees of energy restriction during WWII. METHODS: All new cases of colorectal cancer in the Nordic countries and Estonia diagnosed between 40 and 84 years of age and born between 1874 and 1953, were collected from the national cancer registries. The incidence data were fitted to an age-period-cohort model. RESULTS: A transient drop in the estimated colorectal cancer incidence rate was observed for the birth cohorts born around WWII in Estonia, together with a tendency of decreased risk in Sweden and Denmark. CONCLUSION: The previously observed lowering of colorectal cancer risk for persons born during WWII in Norway also prevails in Estonia. Energy restriction is a possible explanation for these findings, since the countries suffered from varying nutritional conditions during the war. Exogenous factors acting during periods early in life may have an impact on later colorectal cancer risk.

Breast and other cancers in 1445 blood relatives of 75 Nordic patients with ataxia telangiectasia.

Epidemiological studies have consistently shown elevated rates of breast cancer among female blood relatives of patients with ataxia telangiectasia (AT), a rare autosomal recessive disease. A large proportion of the members of AT families are carriers of AT-causing gene mutations in ATM (Ataxia Telangiectasia Mutated), and it has been hypothesised that these otherwise healthy carriers are predisposed to breast cancer. This is an extended and enlarged follow-up study of cancer incidence in blood relatives of 75 patients with verified AT in 66 Nordic families. Blood relatives were identified through population registry linkages, and the occurrence of cancer was determined from cancer registry files in each country and compared with national incidence rates. The ATM mutation carrier probabilities of relatives were assigned from the combined information on location in family, consanguinity, if any, and supplementary carrier screening in some families. Among the 1445 blood relatives of AT patients, 225 cancers were observed, with 170.4 expected, yielding a standardised incidence ratio (SIR) of 1.3 (95% confidence interval (CI), 1.1-1.4). Invasive breast cancer occurred in 34 female relatives (SIR, 1.7; 95% CI, 1.2-2.4) and was diagnosed in 21 women before the age of 55 years (SIR, 2.9; 95% CI, 1.8-4.5), including seven mothers of probands (SIR, 8.1; 95% CI, 3.3-17). When the group of mothers was excluded, no clear relationship was observed between the allocated mutation carrier probability of each family member and the extent of breast cancer risk. We concluded that the increased risk for female breast cancer seen in 66 Nordic AT families appeared to be restricted to women under the age of 55 years and was due mainly to a very high risk in the group of mothers. The findings of breast cancer risk in mothers, but not other likely mutation carriers, in this and other studies raises questions about the hypothesis of a simple causal relationship with ATM heterozygosity.

Weak associations between sociodemographic factors and breast cancer: possible effects of early detection.

Differences in incidence and survival of breast cancer have been observed to vary with regard to sociodemographic factors. This might be related to variation in frequency of doctor consultation and in time of diagnosis, since sociodemographic factors appear to influence the individual's attention to cancer symptoms and susceptibility to participate in screening programmes. This study aimed to examine the variation in breast cancer incidence and case fatality in sociodemographic groups in Norway, and to discuss whether any variation can result from temporal variation in detection time. The study included 589 521 women with information on residential history, childbearing pattern, educational level and occupational physical activity. Analyses were conducted using Poisson and Cox regression models. Although all the associations were weak, breast cancer incidence was associated with residence in urban areas, high age at first childbirth and high level of education. The urban women also tended to have better survival compared with the rural women. Childlessness was associated with high incidence and high case fatality. A high educational level was associated with the lowest case fatality. This study may emphasize the importance of discussing potential effects of early cancer detection. This is particularly important in epidemiological studies revealing weak associations between sociodemographic factors and breast cancer. Mortality rates may be less influenced by problems associated with early detection and, thus, analyses of breast cancer-specific mortality could give additional information.

Symptoms and referral of women with epithelial ovarian tumors.

OBJECTIVE: To study referral to hospital units and the clinical characteristics of women with epithelial ovarian tumors in a prospective, population-based study. METHODS: Clinical information on all women diagnosed with epithelial invasive (n=486) and borderline ovarian tumors (n=137) in Norway during 2002 was derived from notifications to the Cancer Registry of Norway and medical, surgical and histopathological records. RESULTS: Sixty-one percent of women with invasive ovarian tumors were initially referred to gynecology units. The 38% of women referred to surgical and medical units were more likely to have symptoms as 'bowel irregularity', 'pain outside the abdominal cavity', 'persisting fatigue' and 'respiratory difficulties'. These women were older, had lower performance status and had a delay in treatment of 20 and 24 days respectively compared to 11 at gynecology units. CONCLUSION: Greater awareness of the symptoms of ovarian cancer might lead to earlier diagnosis and treatment and thus possibly improve survival.

Effect of hospital caseload on long-term outcome after standardization of rectal cancer surgery at a national level.

BACKGROUND: The purpose of this prospective study was to examine the influence of hospital caseload on long-term outcome following standardization of rectal cancer surgery at a national level. METHODS: Data relating to all 3388 Norwegian patients with rectal cancer treated for cure between November 1993 and December 1999 were recorded in a national database. Treating hospitals were divided into four groups according to their annual caseload: hospitals in group 1 (n = 4) carried out 30 or more procedures, those in group 2 (n = 6) performed 20-29 procedures, group 3 (n = 16) 10-19 procedures and group 4 (n = 28) fewer than ten procedures. RESULTS: The 5-year local recurrence rates were 9.2, 14.7, 12.5 and 17.5 per cent (P = 0.003) and 5-year overall survival rates were 64.4, 64.0, 60.8 and 57.8 per cent (P = 0.105) respectively in the four hospital caseload groups. An annual hospital caseload of less than ten procedures increased the risk of local recurrence compared with that in hospitals where 30 or more procedures were performed each year (hazard ratio 1.9 (95 per cent confidence interval (c.i.) 1.3 to 2.7); P < 0.001). Overall survival was lower for patients treated at hospitals with an annual caseload of less than ten versus hospitals with 30 or more (hazard ratio 1.2 (95 per cent c.i. 1.0 to 1.5); P = 0.023). CONCLUSION: The rate of local recurrence was higher for hospitals with a low annual caseload of less than ten procedures than for hospitals with a high treatment volume of 30 or more. Patients treated in small hospitals also had a shorter long-term survival than those treated in large hospitals.