• laryngeal cancer;
  • hypopharyngeal cancer;
  • survival;
  • smoking;
  • alcohol drinking;
  • diet


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Little information is available on the role of tobacco, alcohol and diet in the survival of upper aero digestive cancers. Our study analysed the survival of 931 laryngeal and hypopharyngeal cancer patients, enrolled in a population based case-control study conducted at 5 centres in southeast Europe during 1979–1982. Age at the time of diagnosis and site of origin of tumour were observed to be predictors of the survival. Cigarette smoking, and to a limited extent alcohol drinking, before the diagnosis of tumour seem to influence the overall survival whereas high intakes of vegetables and vitamin C were observed to favourably affect the prognosis. For mortality from upper aerodigestive cancer protective effects of high intakes of vegetables, fibres and vitamin C were observed. Our results support the hypothesis that there is a role for dietary intervention to improve survival of laryngeal and hypopharyngeal cancer patients. © 2005 Wiley-Liss, Inc.

The incidence and mortality from laryngeal/hypopharyngeal cancer is relatively high in southern Europe. Tobacco smoking, alcohol drinking and dietary factors have been linked to the development of upper aerodigestive cancer.1 It is reasonable to argue that these factors might also influence the survival. Little is known, however, about a possible role of lifestyle related factors on survival of laryngeal/hypopharyngeal cancer.

We address the role of tobacco, alcohol and dietary factors in the survival of laryngeal and hypopharyngeal cancer by using the data from multicentric population-based case-control study conducted in 5 southern European centres in early 80s. To our knowledge this is the first large-scale study to describe the role of smoking, alcohol and dietary factors on laryngeal/hypopharyngeal cancers prognosis. Two studies published previously on this topic are the analyses of same data set by 2 centres individually.2, 3 Such information should be useful for clinical and public health intervention to improve the survival.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

A multicentric population-based case-control study of laryngeal and hypopharyngeal cancer was carried out in 6 European regions covered by a cancer registry: the canton of Geneva (Switzerland), the department of Calvados (France), the city of Turin and the province of Varese (Italy) and the provinces of Navarra and Zaragoza (Spain). A common protocol was designed and interviews were conducted to define the role of tobacco, alcohol, diet and occupation in the development of these cancers. The details of our study design have been reported previously.4 The patient group enrolled in our study was followed up for the occurrence of second primary tumour and vital status. The follow-up was conducted by linkage with population, mortality and cancer registry files. The follow-up could not be carried out for the cases from Calvados. For each subject, the starting date of follow-up was defined as the date of diagnosis of laryngeal or hypopharyngeal cancer. The end of follow-up was set to 31 December 2000, or departure from study area or the date of death if these events occurred earlier. Of the total 997 laryngeal/hypopharyngeal cancer cases included in the case-control study, for 25 (2.5%) the vital status was not known. An additional 41 (4.1%) cases with incomplete information on risk factors were also excluded. We had complete information on 931 cases who form the cohort of our present report.

In the original study, tumours were classified according to whether they arose from endolarynx, epilarynx or hypopharynx.5 We used the same definition for the purpose of our analysis. Survival rates were estimated using the Kaplan-Meier method. The Cox proportion hazard model was used to estimate hazard ratios (HR) for mortality according to age, socio-economic groups, tumour site, average cigarette consumption in cigarettes/day, average amount of alcohol intake in g/day and dietary intake. Average daily cigarette consumption was categorized into 4 categories using never-smoker as a reference category. Alcohol drinking in g/day was categorized into 4 categories using non-drinkers and drinkers of <41 g of alcohol/day as a reference category. The details regarding the assessment of diet and the conversion of foods into nutrient intake have been described previously.6 Dietary factors were analysed according to quartile of food items and nutrients, adjusting for age, gender, centre, site of origin of tumour, tobacco smoking, alcohol drinking and calories intake without alcohol. The food groups that are thought to be correlated and associated with survival were adjusted for each other. To test the equality of survivor function across groups, we used stratified log-rank test that is an exponential score test. Analyses were done separately for all causes of death (overall mortality) and considering only deaths from upper aerodigestive cancers (cause specific mortality, ICD9-140-150 and 161) as the event of interest. In the latter analysis, we treated deaths due to other causes as censored events. Analyses were carried out using STATA (STATA Corp. 2003, release 8.0).7


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Patients were followed up for maximum of 21 years and the mean duration of follow-up was 8 years. Table I displays baseline characteristics of laryngeal/hypopharyngeal cancer cases. Follow-up included 931 cases of which 879 (94.4 %) were males and 52 (5.6%) were females. For 73.8% of cases site of origin of tumour was larynx whereas for 10.0% and 16.2% cases site of origin of tumour was epilarynx and hypopharynx, respectively. A total of 755 patients (81.1%) died during the follow-up, and for 473 of them (62.6%) the cause of death was upper aerodigestive cancer. Overall survival was observed to be 79% (95% confidence interval [CI] = 76–81), 52% (95% CI = 48–55), 37% (95% CI = 34–40), 28% (95% CI = 25–31) and 17% (95% CI =14–19) at 1, 5, 10, 15 and 21 years of observation, respectively. The corresponding values for cause specific survival were 80% (95% CI = 78–83), 59% (95% CI = 56–62), 51% (95% CI = 48–55), 46% (95% CI = 43–50) and 40% (95% CI = 36–44).

Table I. Baseline Characteristics of Larynx/Hypopharynx Cancer Cases
CharacteristicsCases, nPercentage
  • 1

    Kaplan-Meier survival estimates at Fifth year of observation.

Age at diagnosis
 < 5016217.4
 ≥ 7017018.3
Occupational group
 Unskilled workers23725.9
 Skilled worker/professional67874.1
Site of cancer
Average cigarette smoking (cigarette/day)
 ≥ 2622424.2
Average alcohol drinking (g/day)
 ≥ 12124726.5

Table II presents the effect of selected risk factors on overall mortality. As expected, an increased mortality was observed with advancing age for all patients except for hypopharynx cancer patients. No difference in survival was observed according to gender and occupational groups. Cigarette smoking was observed to adversely affect the survival, in particular for the cases with tumour originating at endolarynx. Number of non-smokers among cases with tumour at epilarynx and hypopharynx was small (<5). Higher alcohol drinking was observed to worsen the survival, the effect being strongest for tumour originating at epilarynx. Patients with tumour originating at hypopharynx or epilarynx had a significantly poorer survival than those with origin of tumour at endolarynx. We examined interactive effects of smoking and alcohol drinking on survival. Interaction was not significant at 5% level.

Table II. Hazard Ratios of Overall Mortality According to Selected Risk Factors Among Patients with Laryngeal/Hypopharyngeal Cancer
DeathsHR195% CIDeathsHR195% CIDeathsHR95% CIDeathsHR295% CI
  • 1HR, hazard ratios; NA, not applicable. Models include term for age, sex, centre, occupational group, cigarette smoking, alcohol drinking.

  • 2

    Models include term for site of origin of tumour in addition to above variables.

  • 3

    Reference category.

Age (years)
 < 503671.0101.0261.01041.0
 > 701204.73.4–6.5163.31.4–8.0271.20.7–2.21633.32.6–4.3
 Stratified log rank test, p  < 0.001  0.35  0.45  < 0.001
 Stratified log rank test, p  0.69  0.31  0.42  0.37
Occupational group
 Unskilled workers31321.0 261.0361.01941.0
 Skilled worker/professional3850.90.7–1.1531.40.8–2.51080.90.6–1.35470.90.7–1.0
 Stratified log rank test, p  0.7  0.23  0.16  0.59
Average cigarette smoking (cigarette/day)
 ≥ 261232.11.2–3.9181.70.2–11.8410.30.05–1.41841.81.1–3.1
 Stratified log rank test, p  0.002  0.58  0.99  0.005
Average alcohol drinking (g/day)
 ≥ 1211271.21.0–1.6303.51.5–8.1570.90.5–1.52161.31.0–1.6
 Stratified log rank test, p  0.47  0.38  0.30  0.62
Tumour site
 Endolarynx3 NA  NA  NA 5251.0
 Epilarynx NA  NA  NA 811.51.2–1.9
 Hypopharynx NA  NA  NA 1473.02.5–3.7
 Stratified log rank test, p           < 0.001

When we repeated the Cox regression analysis, presented in Table II, taking into account mortality from upper aerodigestive cancers only, the results were only slightly different from those presented above. Age was a still strong predictor of survival and patients with tumour from the hypopharynx or the epilarynx had a significantly poorer survival than those with a tumour from the endolarynx. The effect of cigarette smoking and alcohol drinking was no longer statistically significant (HR for highest level of consumption 1.5; 95% CI = 0.7–3.1 and 1.2; 95% CI = 0.9–1.6, respectively).

Table III reports HR for mortality according to intake of selected food groups and nutrients. High intakes of poultry, vegetables and vitamin C were observed to confer protection in their highest quartiles. When we repeated the analysis for vitamin C after adjusting for vegetables as well, it was still protective (odds ratio [OR] = 0.8, 95% CI = 0.6–1.0). High intakes of meat, fresh fish, animal and vegetable proteins and fibres were observed to non-significantly improve the survival. For fat intake we used the ratio of polyunsaturated to saturated fats (P/S ratio). We could not observe any relationship between survival and consumption of eggs, cheese, potatoes, olive oil, butter, citrus fruits, β thiamine, retinol, carotene, vitamin E and P/S ratio. With reference to minerals, an inverse relationship was observed between survival and potassium intake (HR = 0.6; 95%CI = 0.5–0.9, in highest versus lowest quartile of intake), whereas no relationship emerged for other minerals (sodium, calcium, manganese, phosphorus, iron, copper and zinc). Except for iron intake, data on other mineral intake were not available for Geneva centre (results not shown in detail).

Table III. Hazard Ratios of Overall Mortality According to Intake of Selected Food Groups and Nutrients Intake Among Patients with Laryngeal/Hypopharyngeal Cancer
Dietary items1Upper limit of quartile2Deaths, nHR395% CIStratified log rank test, p
  • 1

    Reference category is Group I.

  • 2


  • 3

    Models are adjusted for age, gender, centre occupational group, site of primary tumour, alcohol drinking cigarette smoking and caloric intake without alcohol.

  • 4

    Meat and poultry intake adjusted for each other.

  • 5

    Fruit and vegetable intake adjusted for each other.

  • 6


Food groups
 Fresh fruits5
 Fresh fish
 Animal protein
 Vegetable protein
 Vitamin C6

When we repeated Cox regression analysis to study effects of diet on cause-specific mortality, results were similar to those for overall mortality. Significant protective effects of poultry, vegetables, and vitamin C were similarly observed in their highest quartiles. Significant protective effect of fibre in their highest quartile was also observed (HR = 0.7; 95% CI = 0.5–0.9), whereas the effects of ratio of polyunsaturated fats to saturated fats (P/S ratio) were also almost significant at the highest quartile (HR = 0.8; 95% CI = 0.6–1.0).


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The data from the case series of patients from 5 cancer registries in south Europe were analysed to study the determinants of survival of laryngeal and hypopharyngeal cancer patients. The population-based nature of our study, the high response rate and the minimal drop out during follow-up provide limited scope for selection bias. Information on lifestyle factors including tobacco, alcohol and diet was collected shortly after the diagnosis of tumour. One limitation of our study is the absence of information on TNM staging and therapy. If the risk factors under observations were related with the stage of disease, i.e., to disease progression, then our estimates would be confounded. A similar effect would result if the choice of therapy were affected by risk factors under observation (less severe therapy for heavy alcohol drinkers for example).

The survival rates observed in our study are comparable to those studies conducted in European cancer registries.8, 9, 10 Age at the time of diagnosis was observed to be a strong predictor of survival, for overall as well as for cause specific mortality, particularly in the age group of 70 years and above. The low rate of survival among the elderly is likely to be due to poor general health and difficulties in undergoing radical cancer therapies. The site of origin of tumour was also found to be the major determinant of survival. The tumours originated in hypopharynx had almost 3 times poorer survival than those originating in endolarynx, whereas tumours originating in epilarynx had an intermediate survival. These findings might reflect confounding by staging.

Tobacco smoking seems to have affected the survival of the patients in our study. Because we lack information on smoking habits after diagnosis or treatment, when it is likely that several of these patients stopped smoking or limited their tobacco use, effect of tobacco smoking on survival might have been underestimated. The finding of our present study confirms the results of partial analyses of the same data set.2, 3 Alcohol drinking seems to affect survival, albeit to a limited extent; the effect seems strongest for the tumour originating at epilarynx. When we restricted the analysis to mortality due to upper aerodigestive cancers only, we did not observe a significant effect of cigarette smoking and alcohol drinking on survival, suggesting that effect of smoking and alcohol drinking on survival may be due to the excessive mortality of heavy smokers and alcohol drinkers due to causes other than laryngeal/hypopharyngeal cancer.

Our study suggests a role of dietary factors in survival. The significant improved survival was observed for patients with high intake of vegetables and vitamin C. There are not many studies in literature regarding the role of diet on survival. The findings related to diet are in agreement with those reported by etiological studies on diet and epithelial neoplasm.11, 12 The beneficial effect of vegetables can be related to their content of several micronutrients, such as carotenoids, vitamin C and E, flavonoid and phytosterols, known to have strong antioxidant and anticarcinogenic properties. The absence of favourable effect of any specific micronutrient except vitamin C, suggest that effect of vegetables may be due to combined effect of several micronutrients or may be due to micronutrients other than those considered separately. The protective effect of vitamin C, even after adjusting for intake of vegetables suggest that vitamin C in particular may be related with survival. Several studies have suggested that vitamin C is a protective factor for laryngeal cancer.13, 14 The effect of dietary items on survival should be interpreted with caution. We could not exclude the possibility that the observed association is due to residual confounding of socio-economic and behavioural factors especially as they may strongly confound association between vitamins and disease.15 We analysed 28 dietary variables, therefore we could not exclude the possibility that some of the observed effects of dietary factors on survival are generated by chance.

To our knowledge this is the first large study that suggests a role of cigarette smoking and dietary factors in survival of laryngeal/hypopharyngeal cancer. Our results provide evidence that dietary factors may interfere with the progression of disease and thus chemoprevention may have a role in improving the survival of these patients.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Dr. R. Dikshit worked on our study under the tenure of an American Cancer Society Beginning Investigators award from UICC. The authors acknowledge the contribution of Drs. J. Esteve, E. Riboli, G. Pequignot, L. Raymond, B. Terracini, F. Berrino and the late Drs. A. Tuyns and W. Lehmann in the design and conduct of the original case-control study. We also acknowledge the contribution of M. Chiusolo and L. Nonnato.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Boffetta P, Trichopoulos D. Cancer of the lung larynx and pleura. In: AdamiHO, HunterD, TrichopoulosD, eds. Text book of cancer epidemiology. New York: Oxford University Press, 2002. p 24880.
  • 2
    Crosignani P, Russo A, Tagliabue G, Berrino F. Tobacco and diet as determinant of survival in male laryngeal cancer patients. Int J Cancer 1996; 65: 30813.
  • 3
    Boffetta P, Merletti F, Faggiano F, Migliaretti G, Ferro G, Zanetti R, Terracini B. Prognostic factors and survival of laryngeal cancer patients from Turin, Italy. Am J Epidemiol 1997; 145: 11005.
  • 4
    Riboli E, Pequignot G, Repetto F, Axerio M, Raymond L, Boffetta P, Zubiri A, Moral AD, Esteve J, Tuyns AJ. A comparative study of smoking drinking and dietary habits in population samples in France, Italy, Spain and Switzerland. Study design and dietary habits. Rev Epidemiol Sante Publique 1988; 36: 15165.
  • 5
    Tuyns AJ, Esteve J, Raymond L, Berrino F, Benhamou E, Blanchet F, Boffetta P, Crosignani P, Moral AD, Lehmann W, Merletti F, Pequignot G, et al. Cancer of the larynx/hypopharynx, tobacco and alcohol: IARC international case-control study in Turin and Varesse (Italy), Zaragoza and Navarra (Spain), Geneva (Switzerland) and Calvados (France). Int J Cancer 1988; 41: 48391.
  • 6
    Esteve J, Riboli E, Pequignot G, Terracini B, Merletti F, Crosignani P, Ascunce N, Zubiri L, Blanchet F, Raymond L, Repetto F, Tuyns A. Diet and cancers of the larynx and hypopharynx: the IARC multi-center study in southwestern Europe. Cancer Causes Control 1996; 7: 24052.
  • 7
    STATA Corp. Stata statistical software: release 8.0. College Station, TX: STATA Corporation, 2003.
  • 8
    Berrino F, Sant M, Verdecchia A, Capocaccia R, Hakulinen T, Esteve J. Survival of cancer patients in Europe the EUROCARE study. vol 132. Lyon: IARC, 1995.
  • 9
    Berrino F, Sant M, Verdecchia A, Capocaccia R, Hakulinen T, Esteve J. Survival of cancer patients in Europe the EUROCARE-2 study. vol 151. Lyon: IARC, 1999.
  • 10
    Ecimovic P, Pompe-Kirn V. Second primary cancer in laryngeal cancer patients in Slovenia, 1961-1996. Eur J Cancer 2002; 38: 125460.
  • 11
    Wu X, Hu Y, Lippman SM. Upper aerodigestive tract cancers. In: NeugutA, MeadowsA, RobinsonE, eds. Multiple primary cancer. Philadelphia: Lippincott Williams & Wilkins, 1999. p 31946.
  • 12
    IARC. Fruits and vegetables. Handbook of cancer prevention. vol. 8. Lyon: IARC, 2003.
  • 13
    Negri E, Franceschi S, Bosetti C, Levi F, Conti B, Parpinel M, La Vecchia C. Selected micronutrients and oral and pharyngeal cancer. Int J Cancer 2000; 86: 1227.
  • 14
    Bosetti C, LA Vecchia C, Talamini R, Negri E, Levi F, Maso LD, Franceschi S. Food groups and laryngeal cancer risk: a case-control study from Italy and Switzerland. Int J Cancer 2002; 100: 35560.
  • 15
    Lawolar DA, Smith DG, Bruckdorfer KR, Devi K, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet 2004; 363: 17247.