N Cooper

Trends and socioeconomic inequalities in cancer survival in England and Wales up to 2001.

We examined national trends and socioeconomic inequalities in cancer survival in England and Wales during the 1990s, using population-based data on 2.2 million patients who were diagnosed with one of the 20 most common cancers between 1986 and 1999 and followed up to 2001. Patients were assigned to one of five deprivation categories (from ‘affluent’ to ‘deprived’) using characteristics of their electoral ward of residence at diagnosis. We estimated relative survival up to 5 years after diagnosis, adjusting separately in each deprivation category for background mortality by age, sex and calendar period. We estimated trends in survival and in the difference in survival between deprivation categories (‘deprivation gap’) over the periods 1986–90, 1991–95 and 1996–99. We used period analysis to examine likely survival rates in the near future. Survival improved for most cancers in both sexes during the 1990s, and appears likely to continue improving for most cancers in the near future. The deprivation gap in survival between rich and poor was wider for patients diagnosed in the late 1990s than in the late 1980s. Increases in cancer survival in England and Wales during the 1990s are shown to be significantly associated with a widening deprivation gap in survival.

Cancer survival in England and Wales at the end of the 20th century

Survival has risen steadily since the 1970s for most cancers in adults in England and Wales, but persistent inequalities exist between those living in affluent and deprived areas. These differences are not seen for children. For many of the common adult cancers, these inequalities in survival (the ‘deprivation gap’) became more marked in the 1990s. This volume presents extended analyses of survival for adults diagnosed during the 14 years 1986–1999 and followed up to 2001, including trends in overall survival in England and Wales and trends in the deprivation gap in survival. The analyses include individual tumour data for 2.2 million cancer patients. This article outlines the structure of the supplement – an article for each of the 20 most common cancers in adults, followed by an expert commentary from one of the leading UK clinicians specialising in malignancies of that organ or system. The available data, quality control and methods of analysis are described here, rather than repeated in each of the 20 articles. We open the discussion between clinicians and epidemiologists on how to interpret the observed trends and inequalities in cancer survival, and we highlight some of the most important contrasts in these very different points of view. Survival improved substantially for adult cancer patients in England and Wales up to the end of the 20th century. Although socioeconomic inequalities in survival are remarkably persistent, the overall patterns suggest that these inequalities are largely avoidable.

Survival from cancer of the rectum in England and Wales up to 2001

E Mitry, B Rachet, MJ Quinn, N Cooper and MP Coleman* Département d’Hépatogastroentérologie et Oncologie Digestive, Centre Hospitalo-Universitaire Ambroise-Paré, 9 avenue Charles de Gaulle, Boulogne, France; Cancer Research UK Cancer Survival Group, Non-Communicable Disease Epidemiology Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, Keppel Street, London, UK; Social and Health Analysis and Reporting Division, Off ice for National Statistics (Room FG/114), 1 Myddelton Street, London EC1R 1UW, UK

No socioeconomic inequalities in colorectal cancer survival within a randomised clinical trial

There is strong evidence that colorectal cancer survival differs between socioeconomic groups. We analysed data on 2481 patients diagnosed during 1989–1997 and recruited to a randomised controlled clinical trial (AXIS, ISRCTN32414363) of chemotherapy and radiotherapy for colorectal cancer. Crude and relative survival at 1 and 5 years was estimated in five categories of socioeconomic deprivation. Multiple imputation was used to account for missing data on tumour stage. A multivariable fractional polynomial model was fitted to estimate the excess hazard of death in each deprivation category, adjusting for the confounding effects of age, stage, cancer site (colon, rectum) and sex, using generalised linear models. Relative survival in the trial patients was higher than in the general population of England and Wales. The socioeconomic gradient in survival was much smaller than that seen for colorectal cancer patients in the general population, both at 1 year −3.2% (95% CI −7.3 to 1.0%, P=0.14) and at 5 years −1.7% (95% CI −8.3 to 4.9%, P=0.61). Given equal treatment, colorectal cancer survival in England and Wales does not appear to depend on socioeconomic status, suggesting that the socioeconomic gradient in survival in the general population could well be due to health-care system factors.

Survival from cancer of the larynx in England and Wales up to 2001.

Cancer of the larynx is one of the more common malignancies in England and Wales (ranking 20th in both sexes combined). Approximately 1800 new cases are diagnosed each year, 80% of them in men. Incidence rates are approximately 6.2 and 1.3 per 100 000 per year in men and women, respectively. Incidence has fallen by approximately 5% in men over the last decade, but little change has occurred in women. Laryngeal cancer is rare under the age of 40, but the risk rises rapidly with age. There is a marked socioeconomic gradient, with risk twice as high in the most deprived groups as in the most affluent groups (Quinn et al, 2001). Geographic variation in risk is also wide, with incidence less than 70% of the United Kingdom and Ireland average in southwest England and parts of the southeast, but 50% or more above the average in much of Scotland and in the main urban areas of northwest and northeast England. The combined effect is a striking regional disparity in the socioeconomic profile of the disease. In the Oxford region, for example, 50% of cases occur in affluent groups, although in the West Midlands and the northwest, that proportion is approximately 20%, with 65% of cases among the most deprived (data not shown). The annual death rate of laryngeal cancer in England and Wales is approximately 2.3 per 100 000 in men (570 deaths a year) and 0.6 in women (150 deaths a year). The main risk factors for laryngeal cancer are alcohol and tobacco, and their effects are synergistic (Tuyns and Audigier, 1976; Tuyns et al, 1988). Tobacco dominates the risk for cancers of the vocal cords and glottis, whereas alcohol is more prominent for cancers of the supraglottis. This has a direct impact on survival in men and women for all laryngeal cancers combined, because the main causal exposures and the most common anatomic location of tumours within the larynx differ between the sexes, as do their diagnosis, treatment and outcome. Glottal cancers are more common in men; they give rise to hoarseness when the tumour is still small. They can often be treated surgically and are responsive to radiotherapy. They tend to have higher survival than supraglottic tumours. Cancers of the supraglottis are more common in women and do not give rise to early symptoms of hoarseness. Diagnosis from dysphagia or sore throat is often later than for cancers of the glottis, curative radiotherapy and surgery may be less successful, and survival is lower. Survival analyses are reported here only for men. Some 20 000 men were diagnosed with a first, primary, invasive malignancy of the larynx in England and Wales during the period 1986–1999, and followed up to the end of 2001, approximately 89% of those eligible for analysis. Approximately 2% were excluded because their vital status was unknown on 5 November 2002, when the data were extracted for analysis; 4% because the laryngeal cancer was not their first primary cancer and another 4% because their survival was zero or unknown, most of whom were registered from a death certificate only. Half (49%) of the laryngeal tumours diagnosed in men during the 1990s arose in the glottis (endolarynx), including the vocal cords. The increase of approximately 5% since the 1980s is matched by a similar drop in the proportion of tumours of unspecified subsite (down to 31%), suggesting gradual improvement in diagnostic precision. Approximately 16% arose in the supraglottis (epilarynx). Tumours of the larynx below the cords (subglottis) remained rare (1.3%). Almost 85% of laryngeal tumours diagnosed during the 1990s were squamous carcinomas, an increase of 6% since the 1980s, matched by a similar fall in the proportion coded as carcinoma without further specification (down to 7%), again suggesting improved precision of pathology. Verrucous carcinoma was specified as often in the 1990s alone (125 cases, 1%) as in the earlier two decades combined (130 cases, 0.4%), but adenocarcinoma remains rare (0.3%).

Survival from cancer of the breast in women in England and Wales up to 2001

Breast cancer is the most common cancer in women worldwide (Parkin, 2001). In England and Wales at the end of the 1990s, approximately 36 000 new cases were diagnosed each year, representing 30% of the annual total of 120 000 cases. The number of cases exceeds the combined total for the second and third most common cancers in women, those of the large bowel (14 700 cases, 12%) and lung (12 600 cases, 10%). Breast cancer in men is comparatively uncommon, and is not considered further here. During the 1970s and 1980s, age-standardised incidence increased on average by approximately 2% each year (Quinn et al, 2001). The NHS breast-screening programme began in 1988, covering women aged 50–64 years and using single mediolateral oblique view mammography and a 3-year interval between screens. During the prolonged prevalence round, overall breast cancer incidence increased by approximately 20%; it subsequently declined, but then rose again, and by the late 1990s, incidence was some 7% higher than at the peak in the early 1990s (Office for National Statistics, 2003). Part of the increase, possibly as much as 20% in post-menopausal women, may have resulted from the increasing use of hormone replacement therapy (HRT) in the 1990s (Beral et al, 1997). After the exclusion of in situ tumours, we analysed the data for 382 277 women diagnosed with a first, primary, malignant neoplasm of the breast in England and Wales during the 14-year period 1986–1999 and followed up to 31 December 2001, some 89% of those eligible for analysis. Approximately 6.1% of women who were otherwise eligible were excluded with zero recorded survival (date of diagnosis same as date of death): most of these women will have been registered from a death certificate only (DCO), and their duration of survival is unknown, but they could not be reliably distinguished in these data from women with true zero survival. The proportion of women excluded from analysis as DCO records was very similar in all socioeconomic groups. A further 2.4% of women were excluded because it was not their first primary cancer: for one-third of these women, the previous cancer was also a breast cancer, diagnosed before 1986. Major shifts in the morphological distribution of breast cancers have occurred over the last 30 years. The proportion described as an adenocarcinoma has been falling steadily, from 36% in the early 1970s (Coleman et al, 1999) to 6% in these data by the late 1990s. In the early 1970s, ductal, lobular and medullary tumours comprised less than 10% of all breast cancers, but by the late 1990s this figure had reached 75%, ductal carcinomas alone comprising 60% of all breast cancers. The proportion of tumours with unspecified morphology has continued to fall, from 47% in 1971–1975 to 30% in 1986–1990 and less than 5% in 1996–1999. Even this massive improvement in data quality cannot account for the overall increase in ductal, lobular and medullary tumours.

Survival from cancer of the colon in England and Wales up to 2001

Colon cancer is one of the most common malignancies in adults worldwide (Parkin, 2001). About 17 000 new cases are diagnosed each year in England and Wales, where it is the second and third most frequent cancer in women and men, respectively. Over the last 25 years, incidence has risen slightly in men, but has remained fairly stable in women (Quinn et al, 2001). Of the improvements in survival reported in Europe during the 1990s, one of the largest absolute increases was for large bowel cancer (colon and rectum combined). Survival increased more slowly in the UK during this period than in other western European countries (Sant et al, 2001). Substantial deprivation gradients in colon cancer survival have been reported, both in England and Wales (Coleman et al, 1999) and in other countries (Auvinen, 1992; Monnet et al, 1993). These have been attributed in part to differential access to chemotherapy (Corazziari et al, 2004) or specialised treatment centres (Herbert et al, 2002; Satagopan et al, 2004).

Survival from cancer of the ovary in England and Wales up to 2001

The incidence of ovarian cancer varies widely around the world, with the highest rates in the developed countries of Europe and North America, and the lowest in Africa and Asia (Parkin et al, 2002). In England and Wales, ovarian cancer is the most common gynaecological cancer, and the fourth most common after cancers of the breast, large bowel and lung, representing some 5% of all cancers in women. It is rare in premenopausal women, with less than 10% of cases arising under the age of 45. The pattern of incidence for ovarian cancer with age is similar to that of breast and uterine cancers, and they share some reproductive risk factors. Late menarche, high parity, early menopause and long-term use of combined oral contraceptives all confer a lower risk of ovarian cancer, probably by reducing the number of ovulatory cycles. Incidence in England and Wales has increased gradually since the early 1970s, especially in older women (Coleman et al, 1993), but the upward trend in younger women appears to have reached a plateau by the late 1990s (Quinn et al, 2001). Incidence is 5–10% higher among women in more affluent groups than in the most deprived group, and it increased further in these groups during the 1990s (data not shown). In the late 1990s, about 6000 women were diagnosed with ovarian cancer in England and Wales each year. Ovarian cancer accounts for 4000 deaths a year in England and Wales, about 6% of all cancer deaths in women. In contrast to the increasing trend in incidence, overall age-standardised mortality has fallen slightly over the last decade, but a 20–30% fall in mortality among women under 65 years has been balanced by an increase of about 10% in mortality among older women. Survival from ovarian cancer is the lowest among the gynaecological cancers, because it is often at an advanced stage when diagnosed. Some 81 600 women were registered with an ovarian tumour in England and Wales during the 14-year period 1986–1999, but more than 6000 of these tumours were benign, of uncertain behaviour, or metastatic to the ovary from a primary malignancy elsewhere. Of the 75 800 women resident in England and Wales who were registered with a primary, malignant tumour of the ovary, some 63 800 were included in the analyses (84% of those eligible). One percent of women were excluded because their vital status was not known on 5 November 2002, when the data were extracted for analysis and 10% because their recorded survival time was zero (mainly death certificate only (DCO) cases whose survival time was unknown). A further 3% were excluded because it was not the womans first primary, invasive cancer, a previous malignancy having been registered for the same woman at some time since 1971. The proportion of cases excluded from analysis as DCOs did not vary by year of diagnosis or by deprivation category. Tumours of the ovary have usually been grouped with those of the Fallopian tube, broad ligament and other uterine adnexa (ICD-9 183, ICD-10 C56–57). Ovarian tumours were only assigned a separate three-digit rubric in the International Classification of Diseases with the introduction of the tenth revision in 1995 (ICD-9 183.0; ICD-10 C56), but in any case almost 99% of the tumours were coded to the ovary, with only about 1% to the Fallopian tube (183.2, C57.0); tumours coded to the broad ligament and other adnexa were extremely rare. As is conventional, therefore, these tumours were included with ovarian cancers in the survival analyses, for consistency in the interpretation of long-term trends. Adenocarcinomas accounted for most tumours that were assigned to a specific morphology, with 39% coded to serous, papillary or mucinous cystadenocarcinoma and 34% to other specific types of adenocarcinoma; 20% were poorly specified carcinomas.

Survival from cancer of the pancreas in England and Wales up to 2001

Cancer of the pancreas accounts for some 3% of cancer in both sexes combined, with about 6000 new cases a year (Cooper et al, 2005). The only established risk factor for pancreatic cancer is tobacco smoking (Lowenfels, 1984). Incidence has fallen some 10–15% since the mid-1970s in men but has risen slightly in women, and annual incidence in both sexes is now similar at about 10 per 100 000 (Quinn et al, 2001). Both incidence and mortality rates may be biased, however, by under-registration of incident cases and an over-certification of deaths (Remontet et al, 2003). In this context, the survival of patients who were registered with a diagnosis of pancreatic cancer in life assumes particular importance as a measure of outcome. Pancreatic cancer has one of the worst prognoses: the European mean relative 5-year survival rate for patients diagnosed during 1990–1994 was less than 4% (Sant et al, 2003). We analysed the data for 62 815 patients registered with pancreatic cancer in England and Wales during the period 1986–1999, only 74% of those were apparently eligible. More than a fifth (22%) of all cases had to be excluded from survival analysis because their recorded survival was zero (date of diagnosis same as date of death): some will in fact have died on the day of diagnosis, but most were registered from a death certificate only (DCO), and their survival time was unknown. Some 3% of cases in the national data were known to have been DCOs, but the proportion varied by registry, and they could not be reliably distinguished from cases with true zero survival in the national data. The proportion of cases whose recorded survival was zero rose from 12 to 14% in the 1970s to 19–22% in the 1990s (Coleman et al, 1999). As they represent such a large proportion of patients who were otherwise eligible for survival analysis, and who may have shorter than average survival (Berrino et al, 1995), the impact of their exclusion on observed trends and inequalities in survival needs to be considered. Nationally, the proportion of such cases fluctuated slightly within the range 19–25%. Trends in the proportion of such cases, however, varied very widely among regions: stable and relatively low (6–11%) in Northern and Yorkshire, East Anglia and Oxfordshire; initially high but falling rapidly (Thames, Wales); or initially low but rising rapidly (South and West, West Midlands). These patterns are hard to interpret, as the efficiency and timeliness of registration have increased over this period, and the proportion of other cancers registered solely from a death certificate has generally declined. Overall estimates of survival may thus be slightly high. In contrast, the impact of exclusions for zero survival on trends in the deprivation gap in survival during the 1990s is likely to have been small or negligible, as there was no systematic difference in the proportion of such cases among socioeconomic groups (data not shown). The vital status of 1.4% of patients was unknown at 5 November 2002, and a further 2.8% were also excluded because pancreatic cancer was not their first primary malignancy.

Survival from brain tumours in England and Wales up to 2001

Brain tumours comprise 2% of all malignant neoplasms in adults in England and Wales, with some 4000 new cases and 3000 deaths each year. Incidence has increased by approximately 25% since 1971 in England and Wales (Quinn et al, 2001). Similar increases have been seen in other western countries (Muir et al, 1994). Brain tumours are 20–50% more common in men. Incidence was approximately 25% higher in the most affluent fifth of the population in England and Wales than in the most deprived fifth during 1991–1993 (Quinn et al, 2001). The cause of most brain tumours is unknown. Ionising radiation is the only established cause, although some nitrosamines may be a risk factor (Preston-Martin, 1996). Inherited syndromes may account for 5% of cases. Acquired immunosuppression may increase the risk of cerebral lymphoma. Occupations that have been linked to increased risk include the petrochemical and rubber industries, agricultural work, the nuclear industry and work involving exposure to electromagnetic fields (Preston-Martin et al, 2006). Only tumours of the brain explicitly coded as primary, malignant (behaviour code 3) tumours were included here. Patients previously registered with another primary malignancy at any time since 1971 were excluded. Brain tumours coded as benign (behaviour code 0) or of uncertain or unspecified behaviour (behaviour code 1) are often considered together with malignant tumours (Muir et al, 1994), and approximately 10 900 such brain tumours were recorded in the National Cancer Registry during the period 1986–1999, approximately 20% of all recorded brain tumours in all regions of England and Wales (data not shown). These tumours were not considered eligible for analysis. It can be difficult to distinguish primary tumours of the brain from metastases of primary tumours in other organs, which are common, but tumours coded as metastatic (behaviour code 6) were excluded. Malignant tumours of the cranial nerves and spinal cord were also excluded. The survival analyses reported here are based on the data for 37 917 adults (aged 15–99 years) who were registered with a first, primary, malignant tumour of the brain in England and Wales during the period 1986–1999 and followed up to the end of 2001. These patients represented approximately 86% of those eligible for analysis. For 1.9%, the vital status was unknown on 5 November 2002 when the data were extracted for analysis, and they were excluded. Patients whose brain tumour was not their first primary malignancy (1.8%) were also excluded. Most of the other exclusions were for a recorded survival time of zero (date of diagnosis same as date of death; 9.8% of cases). Some of these patients may have died on the day of diagnosis, but many are likely to have been registered from a death certificate only (DCO) (data not shown), and the two groups could not be distinguished in these data. As the date of diagnosis and thus the duration of survival are not available from a death certificate, both categories were excluded. Such patients may well have shorter than average survival (Berrino et al, 1995); hence, their exclusion may have caused slight overestimation of overall survival. However, the proportion of DCO cases was similar between deprivation groups and stable over time; hence, their exclusion is unlikely to have caused bias in the estimates of trends in survival, or of socioeconomic gradients. During the 1990s, some 50% of brain tumours were specified as arising in the frontal, parietal, temporal or occipital lobes, and only 3% arose in the cerebellum or brain stem, but the site was poorly specified or unspecified for 42% of tumours. The proportions are similar to those for the 1980s (Coleman et al, 1999). Gliomas represented 87% of the tumours, mainly malignant glioma (52%) and astrocytoma (30%). Morphology was ill-defined or undefined for approximately 11%.