IMJ
About IMJ
Disclaimer
Contact IMJ
Register as Reviewer
Register As Author
View IMJ Authors
View IMJ Volumes
View Supplement
Issue Archive 1980-1996
Subscription Detail 2010

IMJ Search

 

Advanced Search
 
 
Lung Cancer Survival in Northern Ireland   Back Bookmark and Share
AT Gavin

Lung cancer causes more deaths than any other cancer in Northern Ireland. Survival, and factors which could influence survival, were examined for the 4,458 patients diagnosed with lung cancer from 1992 to 1996. Overall five-year relative survival was low, but better for females (9%) than males (7%). Survival improved each year although this could not be attributed to treatment changes. Survival was better for men under 65 years, than older men (p<0.01), while survival from non-small cell type lung cancer was better than small cell (p<0.001). The proportion of patients having bronchoscopy and CT scan increased over the study period. Although the proportion of patients receiving surgery or radiotherapy remained unchanged over the period there was an increase in the proportion receiving chemotherapy (8.7% vs. 9.4%, P<0.05). Survival rates, although slightly better than in the rest of the United Kingdom, were poorer than in several other countries, including the Republic of Ireland and the United States of America. High mortality rates from lung cancer and, at best, modest improvements in survival, point to the need for increased efforts to reduce levels of lung cancer by tackling the main cause, smoking.

Author : AT Gavin, P Wilkinson

Abstract

Lung cancer causes more deaths than any other cancer in Northern Ireland. Survival, and factors which could influence survival, were examined for the 4,458 patients diagnosed with lung cancer from 1992 to 1996. Overall five-year relative survival was low, but better for females (9%) than males (7%). Survival improved each year although this could not be attributed to treatment changes. Survival was better for men under 65 years, than older men (p<0.01), while survival from non-small cell type lung cancer was better than small cell (p<0.001). The proportion of patients having bronchoscopy and CT scan increased over the study period. Although the proportion of patients receiving surgery or radiotherapy remained unchanged over the period there was an increase in the proportion receiving chemotherapy (8.7% vs. 9.4%, P<0.05). Survival rates, although slightly better than in the rest of the United Kingdom, were poorer than in several other countries, including the Republic of Ireland and the United States of America. High mortality rates from lung cancer and, at best, modest improvements in survival, point to the need for increased efforts to reduce levels of lung cancer by tackling the main cause, smoking.

Introduction
Lung cancer causes almost 800 deaths in Northern Ireland each year, more than any other cancer, whilst over 900 patients are diagnosed with lung cancer annually. Marked international variations in the treatment and outcome of lung cancer have been reported over time1. Northern Ireland has a stable population and detailed information has been available on patients developing lung cancer since October 1991.

This study examines lung cancer survival, and changes in survival over the five-year study period, 1992 to 1996, investigates whether any improvements could be related to changes in clinical practice, and compares survival in Northern Ireland with other European countries and the United States of America (US).

Methods
Dataset

The Northern Ireland Cancer Registry (NICR) records electronically, all incident cases of cancer in Northern Ireland diagnosed since 1993. Data were extracted from the NICR database on all 3,576 malignant lung tumours (International Classification of Diseases, 9th Revision (ICD9 162))2 over the period January 1993 to December 1996. Data on 882 patients were also available from an earlier study that spanned a one-year period from October 1991 to September 1992. This one-year period will be referred to as 1992 hereafter. The study was a prospective study using multisource notification3. Demographic, tumour, treatment, date and cause of death information was extracted from these datasets.

Based on pathology reports, the cases were classified according to cell type:
  1. Non small cell (including squamous, adenocarcinoma, large cell, other microscopically verified),
  2. Small cell,
  3. Unknown no microscopic verification (NMV).

Data on all deaths in the Northern Ireland population for all causes including cancer were obtained from the General Registers Office (GRO). The incident data were cross-matched with the GRO mortality data to ascertain if a person had died and, if so, his or her date of death. The censoring date for deaths was 31 December 1999. Those patients not matched with a date of death were assumed to be still alive on the censoring date. Active follow-up of these survivors was undertaken, by seeking confirmation of their survival status from their hospital of treatment and/or general practitioner. Survival times were calculated for each patient, defined as the time elapsed between diagnosis and death or censoring, whichever occurred earlier.

Two hundred and nine patients (4.7%) were first diagnosed with lung cancer at or after death, resulting in zero or negative survival times. These patients were excluded from the survival dataset, as inclusion of such cases would have biased survival downwards. This is standard procedure of cancer registries in calculating survival4.

Statistical Analysis
To allow for deaths from diseases other than lung cancer, survival was expressed as a relative survival rate (RSR). Relative survival is the ratio of the observed survival divided by the survival that the patient would have experienced if s/he had the same probability of dying as the general population having the same age and sex, and thus adjusts for background mortality from all causes. The relative survival rates were calculated using the SURV2 relative survival software developed by the Finnish Cancer Registry5.

Linear regression analyses were carried out on the age-standardised incidence and mortality rates separately for males and females to investigate if there were significant trends over the study period, 1992 to 1996. Additionally, Coxs proportional hazards regression models were fitted to the data, in order to investigate if year of diagnosis, age at diagnosis, sex, cell type and treatment, and interactions between these factors were significant predictors of survival. Relative survival rates were then calculated for each of the significant predictor variables. The survival rates for Northern Ireland lung cancer patients were compared with the rates for the Republic of Ireland6, Scotland4, England and Wales7, other European registries1 and the USA8.

Results
Incidence and Mortality

During the study period, there were 2,899 male and 1,559 female incident lung cancers, and 2,563 male and 1,326 female lung cancer deaths (Table 1). The incidence and mortality from lung cancer fell in men and rose in women, but the changes in incidence and mortality were not statistically significant (P>0.05).

Table 1 Incidence and mortality of lung cancer (ICD9 162) in the Northern Ireland population
Year 199219931994199519961992 to 1996 Significance of Tend P(trend)
MaleIncident Cases 589574626547563P=0.26
EASR* per 100,00084.380.689.575.274.7
(EASR 95% CI)(79.8 - 88.8)(73.9 - 87.3)(82.4 - 96.6)(68.8 - 81.5)(68.5 - 80.9)
FemaleIncident Cases293320314321311P=0.74
EASR* per 100,00031.234.733.133.632.6
(EASR 95% CI)(28.3 - 34.1)(30.8 - 38.7)(29.2 - 36.9)(29.8 - 37.5)(28.8 - 36.4)
MaleNo. deaths512539518489505P=0.05
EASR* per 100,00073.675.672.967.066.6
(EASR 95% CI)(67.1 - 80.0)(69.1 - 82.1)(66.5 - 79.3)(60.9 - 73.0)(60.7 - 72.5)
FemaleNo. deaths259266261266274P=0.24
EASR* per 100,00020.228.525.927.827.4
(EASR 95% CI)(16.6 - 23.3)(25.9 - 32.1)(22.6 - 29.2)(24.3 - 31.3)(24.0 - 30.9)
* EASR = European Age-Standardised Rate

Age and Sex
The age profiles for males and females diagnosed with lung cancer during the study period were similar; their mean (standard error) ages being 68.8 (0.2) and 68.4 (0.3) years respectively. The study population was elderly (median age at diagnosis being 69 years) and predominantly male (65%).

Cell Type
Over the study period, 2,630 (59%) lung cancers were recorded as non-small cell type, 535 (12%) were small cell and 1,293 (29%) were categorised NMV. In 1992 data on cell type were available for 750 (85%) patients compared with 656 (75%) in 1996. Significantly more incident lung cancers were recorded as non-small in 1992 626 (71%) compared to 533 (61%) in 1996, P<0.001).

Investigations
Data on bronchoscopy and CT scanning rates were available for 1992 and 1996 data. Bronchoscopies were performed in 494 (56%) lung cancer patients in 1992 and this increased to 585 (67%) in 1996, while CT scanning was used in 220 patients (52%) in 1992 rising to 586 (67%) in 1996.

Treatment Effects
The overall number of patients having surgery was higher in 1996 (Table 2), although not significantly (P>0.05). There was no difference in the proportions of patients having radiotherapy (P>0.05), however, more patients received chemotherapy in 1996 than in 1992 (P<0.05). Differences in treatment may account for variations in survival. A smaller proportion of patients in Northern Ireland received surgery (P<0.001) and chemotherapy (P<0.01) than in the Republic of Ireland6, whilst the proportion of patients undergoing radiotherapy was similar.

Table 2 Treatment of lung cancer surgery, chemotherapy and radiotherapy in Northern Ireland and Republic of Ireland by period of diagnosis
TreatmentSurgeryChemotherapyRadiotherapy
NINIROI*NINIROI*NINIROI*
19921996 19921996 19921996 
No. cases72862197577207304261418
%8.410.5158.79.413.835.331.830.9
NI: Northern Ireland
*Average 1994-1998 figures for ROI (Republic of Ireland)

Survival
The year of diagnosis, age at diagnosis, sex and cell type were found to be significant predictors of the risk of death from lung cancer (Table 3). There were no statistically significant interactions between the factors (P>0.05). Although the proportion of patients receiving chemotherapy was higher in 1996 than in 1992, it was not a statistically significant factor in the model (P>0.05). Radiotherapy did not have a significant effect (P>0.05) on the hazard of lung cancer death. For each year (1993 to 1996), the hazard for lung cancer death was significantly lower than the hazard of lung cancer death if diagnosed in 1992 (P<0.01). For example, the hazard for lung cancer death diagnosed in 1996 was 69% of the hazard for patients diagnosed in 1992. For women, the hazard for lung cancer death was 83% that of men (P<0.05). The hazard of lung cancer death for small cell type was 166% of that for non-small cell (P<0.001), whilst the hazard of lung cancer death for NMV was about 120% of that for non-small cell (P>0.05).

Table 3 Hazard ratio, 95% confidence interval, and P-value for change in survival expectancy per unit change in predictor variable,
Predictor Variable
Change in survival
95% Confidence Interval
P-value
Sex
0.83
0.71 - 0.97
<0.05
Female vs. male
Year of Diagnosis
<0.01
1993 vs. 1992
0.75
0.60 - 0.94
<0.05
1994 vs. 1992
0.79
0.63 - 0.99
<0.05
1995 vs. 1992
0.69
0.54 - 0.87
<0.01
1996 vs. 1992
0.69
0.54 - 0.87
<0.01
Age (years)
1.01
1.00 - 1.02
<0.01
<65 vs. 65
Cell Type
<0.001
Non-small vs. small
1.66
1.31 - 2.10
<0.001
Non-small vs. NMV
1.20
0.98 - 1.47
>0.05

Sex
Women had significantly better survival from lung cancer than men (P<0.05). The percentages of men and women alive after five years of follow-up were 7% and 9% respectively (Table 4).

Table 4 Lung cancer relative survival [95% Confidence Interval] by sex, age at diagnosis and cell type
 Relative survival(%) [ 95% confidence interval]
Sex
Age/Cell Type
No. Cases
1-year
2-year
3-year
4-year
5-year
Male
All Ages
2745
23 [22-25]
12 [11-13]
9 [8-10]
8 [6-9]
7 [6-8]
Male
<65yrs
833
28 [25-31]
14 [12-17]
12 [10-15]
10 [8-13]
9 [7-11]
Male
65yrs
1912
21 [19-23]
11 [10-13]
8 [7-10]
6 [5-8]
6 [4-7]
Female
All Ages
1466
25 [23-27]
14 [12-16]
11 [10-13]
10 [9-12]
9 [8-11]
Female
<65yrs
469
28 [24-32]
15 [12-18]
12 [9-15]
10 [8-13]
10 [7-12]
Female
65yrs
997
24 [21-26]
13 [11-16]
11 [9-13]
10 [8-13]
9 [7-11]
Male
Non Small
1663
29 [27-31]
16 [14-18]
12 [10-14]
10 [8-12]
9 [7-11]
Male
Small
316
17 [13-22]
6 [3-8]
4 [1-6]
2 [0-4]
0
Male
NMV
766
12 [10-14]
7 [5-8]
5 [5-9]
4 [2-5]
4 [2-6]
Female
Non Small
806
31 [27-34]
19 [16-22]
16 [13-19]
15 [12-17]
14 [11-16]
Female
Small
207
21 [15-27
5 [2-8]
0
0
0
Female
NMV
453
16 [13-20]
9 [7-11]
7 [4-9]
5 [3-8]
4 [1-6]

Age
Men diagnosed with lung cancer before 65 years had significantly higher survival rates than those diagnosed at older ages (9% five-year RSR vs. 6% respectively, P<0.01). However, for women there was no statistically significant difference between the survival rates of the two age groups (P>0.05) (Table 4).

Cell Type
Overall, patients with non-small cell type lung cancer had significantly better survival rates than those with small cell and NMV (P<0.001) (Table 4). The percentages of men alive after five years of follow-up were 9% and 4% for non-small cell and NMV respectively. For women, the corresponding five-year relative survival rates were 14% and 4% respectively. Only four of the 523 small cell patients survived at least five years after diagnosis.

National and International Comparisons
Relative survival information was available for the Republic of Ireland6, England and Wales10, Scotland4, other European Registries1 and the US8. Five-year relative survival rates for Northern Ireland patients in1992-96, of 7% in males and 9% in females were better than those in England (1991-93): males 5%, females 5%. They were, however, slightly lower than those in the Republic of Ireland (1994-98): males 8%, females 11%, and in several other European countries, including Finland (1985-89): males 11%, females 12%; Holland (1985-89): males 14%, females 12% and were also poorer than US survival rates (1993): males 13%, females 18%.

Discussion
The period 1992 to 1996 for which the trend was examined is short and the fall in incidence and mortality in men and rise in incidence and mortality in women most likely relate to patterns described elsewhere which reflect tobacco consumption9,10.

Survival rates for lung cancer although poor at 9% for females and 7% for males improved between 1992 and 1996, with the hazard of death from lung cancer decreasing year on year. The improvement in survival between 1992 and 1996 was most marked in young men. Overall, survival rates were higher in Northern Ireland than in the rest of the UK. It is likely that almost all cases were identified during the study period and the active follow up of cases ensures that the survival figures are unlikely to be significantly overestimated. Higher survival levels in Europe and the USA of up to 14% and 18% respectively indicate that there is potential for future improvement in survival. Moreover, some of these high survival rates relate to periods earlier than this study and may have improved even further for more recent time periods.

The higher survival reported in the USA for lung cancer is similar to differences reported for other cancer sites and may reflect more aggressive investigation. The microscopic verification rate for lung cancer in the USA is 91%8 compared to 71% in Northern Ireland (1993-1996). Other explanations include under-reporting of advanced disease with resultant improved survival rates, genetic variations in the populations, historic smoking patterns and associated diseases which can adversely affect outcomes. Another factor to consider is variation in treatment which may account for some of the observed variations in survival. Caution also needs to be exercised when comparing population based registries such as those of the UK, Ireland, Finland and Denmark with countries such as the US where only a proportion of the population is covered, as it is possible that the populations covered by regional registries may differ from the general population, in terms of factors such as socio-economic position and access to health services.

Although, over the study period, there was an increase in the number of patients who had surgical procedures and chemotherapy, the improvement in survival could not be explained by these factors. This may be because this study was not powerful enough to discriminate the effect of change. The survival of lung cancer patients in Northern Ireland is less than that of the Republic of Ireland, where a higher proportion of patients were recorded as receiving chemotherapy and tumour directed surgery6. This apparent difference in treatment could reflect the methods of data collection employed by each Registry and this is worthy of closer scrutiny. (NICR operates electronic notification with selected note review while the National Cancer Registry of Ireland obtain data through extraction from hospital records by trained Tumour Registration Officers).

The higher level of microscopic verification in 1992 may reflect the prospective multisource method of data collection undertaken by a clinician compared with the retrospective electronic data collection which is routine for the NICR. The level of microscopic verification in Northern Ireland 1996 was comparable at 75% with that reported as high (74.1%) in Scotland 199511.

A limitation of the study was the insufficient availability of information on disease and stage. There is a need for standardised information regarding stage of lung cancer at presentation to supplement the current data. Also, a five-year period is too short to robustly identify trends in survival; we await future data for this analysis.

The use of CT scanning and bronchoscopy had improved in Northern Ireland for CT scan from 52% of patients in 1992 to 67% in 1996 which is better than that of 47% reported in Scotland 199511. The proportion of patients undergoing investigation by bronchoscopy at 67% was identical to that in Scotland in 1995 and had improved from 56% in 1992. It may be that better targetting of treatment by enhanced selection of patients for surgery explains, at least in part, the year on year improvement in survival.

It is likely that survival in lung cancer in Northern Ireland can be improved in the future. Positron Emission Tomography (PET) promises to improve the selection of patients for lung cancer surgery and hence outcomes. New radiotherapeutic techniques and increasing numbers of patients (with both small cell and non-small cell) receiving chemotherapy also promises to improve survival. In Scotland where services are organised in a similar way to Northern Ireland it was found that management by a respiratory physician, oncologist or thoracic surgeon was an independent predictor of access to potentially curative treatment and of better survival11. This will be further studied in Northern Ireland as part of the evaluation of the implementation of a review of cancer services introduced by the Campbell Report12.

Despite the improvements reported in this study the poor overall survival, alongside the increase in incidence and mortality from lung cancer, points to the need for continued action to prevent tobacco use in the population.

Funding:
The N. Ireland Cancer Registry is funded by the Department of Health, Social Services and Public Safety, Northern Ireland.

Correspondence:
AT Gavin,
Director, N.Ireland Cancer Registry,
Department of Epidemiology and Public Health,
Queens University Belfast,
Mulhouse Building, Grosvenor Road,
Belfast BT12 6BJ.
Tel: 028 9026 3136.
Fax: 028 9024 8017.
E-mail:[email protected]

References
  1. Berrino, F., Sant, M., Verdecchia, A., Capocaccia, R., Hakulinen, T., Estve, J. Survival of Cancer Patients in Europe. The EUROCARE Study. IARC Scientific Publications No. 132, Lyon, IARC, 1995.
  2. World Health Organisation. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death. Ninth Revision. WHO, Geneva 1978.
  3. Wilkinson P. Lung cancer in Northern Ireland 1991 - 1992 [MD dissertation]. The Queens University of Belfast; 1995.
  4. Black, R., Brewster, D., Brown, H., Lesley, F., Harris, V., Kidd, J., Stewart, A., Stockton, D., Weir, J. Trends in Cancer Survival in Scotland 1971-1995. Edinburgh: ISD Publications 2000.
  5. Voutilainen, E.T., Dickman, P.W., Hakulinen, T. Surv: Relative Survival Analysis Program, version 2.0b, Helsinki.
  6. Report of the National Cancer Registry Ireland. Cancer in Ireland. Incidence, Mortality, Treatment and Survival. Cork: National Cancer Registry Board, September 2001.
  7. Office for National Statistics. Cancer Survival in England and Wales 1991-1993. Health Statistics Quarterly No.6, June 2000.
  8. Ries, L.A.G., Eisner, M.P., Kosary, C.L., Hankey, B.F., Miller, B.A., Clegg, L., Edwards, B.K. SEER Cancer Statistics Review, 1973-1998. National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/Publications/CSR1973-1998/, 2001.
  9. Coleman, M.P., Estve J., Damiecki P., Arslan A., Renard H. Trends in cancer incidence and mortality. Lyon: International Agency for Research on Cancer, 1993.
  10. Office for National Statistics. Cancer Statistics: Registrations. England and Wales Series MB1. London: Stationery Office.
  11. Gregor A, Thomson CS, Brewster DH, et al. Management and survival of patients with lung cancer in Scotland diagnosed in 1995: results of a national population based study. Thorax 2001; 56:212-17.
  12. Cancer Services. Investing for the Future. Report of the Cancer Working Group. DHSS, 1996.
Author's Correspondence
No Author Comments
Acknowledgement
No Acknowledgement
Other References
No Other References
   
© Copyright 2004 - 2009 Irish Medical Journal