El efecto sinérgico del consumo de alcohol y tabaco en el riesgo de

carcinoma de células escamosas: Meta-análisis

R2GE Ylse Gutiérrez Grobe

Am J Gastroenterol 2014; 109:822–827

Autores

• Anoop Prabhu

• Kenneth O. Obi

• Joel H. Rubenstein

Anoop Prahbu

Ca de esófago

• 5a y 9a causa de mortalidad en hombres y mujeres respectivamente

• Incidencia 4.4 casos por 100, 000

• China 100 casos por 100, 000

Mortalidad por ca de esófago

• Factores de riesgo:

– Exposición al cigarro

– Exposición al alcohol

• > 90% casos atribuibles a estos factores

• Se sabe poco de cómo interactúan estos dos factores

• Modificando riesgo general

Antecedentes

• Estudio entre 1992 y 1997

• Casos y controles

• 275 hombres

• Con cáncer de esófago de reciente diagnóstico

• Cuestionario de características sociodemográficas, antropometría y hábitos tabáquicos y consumo de alcohol

SMOKING, TYPE OF ALCOHOLIC BEVERAGE AND SQUAMOUS-CELL

OESOPHAGEAL CANCER IN NORTHERN ITALY

Paola ZAMBON1, Renato TALAMINI

2, Carlo LA VECCHIA3,4, Luigino DAL MASO

2, Eva NEGRI3, Sandro TOGNAZZO

1, Lorenzo SIMONATO1

and Silvia FRANCESCHI2*

1Servizio di Epidemiologia dei Tumori, Registro Tumori del Veneto, Padua, Italy2Servizio di Epidemiologia, Centro di Riferimento Oncologico, Aviano, Italy3Istituto di Ricerche Farmacologiche “Mario Negri”, Milan, Italy4Istituto di Statistica Medica e Biometria, Universita degli Studi di Milano, Milan, Italy

Between 1992 and 1997, we conducted a case-controlstudy of oesophageal cancer in 3 areas of northern Italy.Caseswere 275 men, ages39–77 years(median age 60), witha first incident squamous-cell carcinoma of the oesophagus.Controls were 593 men, ages 36–77 years (median age 60)admitted for acute illnesses, unrelated to tobacco and alco-hol, to major hospitals of the areas under surveillance. Num-ber of daily cigarettes wasstrongly associated with risk [oddsratio (OR) for > 25 cigarettes/day 7.0)]. Long-durationsmoking showed particularly elevated ORs(OR 6.4 for > 35years), and excess risk declined after smoking cessation(OR 1.5 after > 10 years). Oesophageal cancer risk steeplyrose with increasing level of alcohol consumption. ORs were6.2 for 35–55 drinksand 24.5 for 84 drinksor more per week.No trend in risk emerged for duration of alcohol drinking orage at start of drinking. The risk in the highest joint level ofalcohol drinking and current smoking wasincreased 130 folds(i.e., compatible with a multiplicative model). Excess risk indrinkers chiefly derived from wine. In conclusion, alcoholdrinking and cigarette smoking were both important, but theroles of dose and duration of exposure differed. The associ-ation with alcohol wasstronger than the one with smoking byexposure intensity, but apparently unaffected by duration orother temporal variables. Int. J. Cancer 86:144–149, 2000.© 2000 Wiley-Liss, Inc.

Alcohol and tobacco have long been suspected of increasing therisk of oesophageal cancer (Munoz and Day, 1996). In Italy (Negriet al., 1992), as in other countries of Europe and the Americas(Munoz and Day, 1996), more than 80% of oesophageal cancercases in men can be attributed to alcohol and tobacco. In otherregions of the world, dietary factors and other exogenous agentshave a larger role, but alcohol and tobacco (including tobaccochewing) are increasingly important (Munoz and Day, 1996).

The degree of risk and the interaction between alcohol andcigarette smoking have been studied in some detail. In regionssuch as northern France (Tuyns et al., 1977) and northern Italy(Baron et al., 1993), where heavy alcohol consumption is com-mon, more elevated risks can be seen at the extreme of consump-tion range for alcohol than for tobacco. A multiplicative model forthe combined risks fits the data well. Elevated risks have beenreported also in smokers who do not drink alcohol (Tavani et al.,1996) and alcohol drinkers who abstain from smoking (Tavani etal., 1994).

In addition to the relative importance of the 2 agents in differentpopulations, a few aspects of alcohol and tobacco carcinogenesis atthe level of the oesophagus need to be better elucidated. Theseinclude time-related variables, such as duration and age at startingand stopping smoking and drinking and the role of specific alco-holic beverages. Several groups have compared the effect of dif-ferent alcoholic beverages (Barra et al., 1990; Grønbaek et al.,1998; Launoy et al., 1997a; La Vecchia et al., 1999; Castellsagueet al., 1999), but more accurate information on the relative con-tribution of different sources of alcohol is still needed.

We present data on the role of alcohol drinking and tobaccosmoking in cancer of the oesophagus in men from 3 areas ofnorthern Italy, i.e., where oesophageal cancer incidence rates inmen (7 to 9 per 100,000, world-standardised) are 3- to 4-fold

higher than in the south of the country, and similar to those foundin intermediate- and high-risk areas of central Europe and SouthAmerica (Parkin et al., 1997).

MATERIAL AND METHODS

A case-control study of oesophageal cancer was conductedbetween January 1992 and December 1997 in 3 Italian areas (theprovinces of Padua and Pordenone in northeast Italy, and thegreater Milan area in north Italy). Cases were men below age 79years with a first histologically confirmed diagnosis of incidentsquamous-cell carcinoma of the oesophagus. A total of 275 caseswith squamous-cell carcinoma (median age 60 years, range 39–77) were included in the present analyses. Conversely, 21 menwith adenocarcinoma of the oesophagus were excluded.

Men admitted for acute illnesses to major hospitals of theareas where cases lived were eligible controls. None of themhad been admitted for malignant tumours. To avoid over-representation of smokers and heavy drinkers, compared to thesource population, hospital admissions due to diseases relatedto tobacco smoking (e.g., chronic obstructive bronchitis andcardiovascular diseases) or alcohol abuse (e.g., liver cirrhosisand pancreatitis) were excluded. By and large, all chronicconditions that may lead to life-style modifications (e.g., dia-betes mellitus) were not eligible causes of hospital admission.Conversely, co-morbidity from the above conditions did notrepresent an exclusion criterion. A total of 593 male controls(median age 60 years, range 36–77) frequency-matched withcases by age quinquennia and area of residence were inter-viewed. They belonged to the following diagnostic categories:traumas, mostly sprains and fractures (27%); other orthopaedicdisorders, such as low-back and disc disorders (22%); acutesurgical conditions (29%); eye diseases (13%); and other ill-nesses, such as ear, nose, skin and dental conditions (9%). Allinterviews were carried out in a hospital setting. Interviewingnurses were introduced to patients by attending clinical staff,resulting in a low proportion of refusal (about 5% in both casesand controls).

The questionnaire included information on socio-demographiccharacteristics, body size measures at various ages and a validatedfood-frequency section, including 78 food groups or recipes and 15questions on general eating patterns. Weekly number of drinks ofthe 5 most common alcoholic beverages or groups of beverageswas investigated. Taking into account the different ethanol con-centrations, 1 drink corresponded to approximately 125 ml ofwine, 330 ml of beer and 30 ml of hard liquor (i.e., about 12 g of

Grant sponsor: Italian Association for Research on Cancer.

*Correspondence to: Servizio di Epidemiologia, Centro di RiferimentoOncologico, Istituto Nazionale Tumori, Via Pedemontana Occ.le 12,I-33081 Aviano (PN), Italy. Fax: 39.0434.659222.E-mail: epidemiology@ets.it

Received 8 September 1999; Revised 9 November 1999

Int. J. Cancer: 86, 144–149 (2000)

© 2000 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

Int J Cancer 2000;86:144–9

Tabaco

• > 25 cigarros al día OR 7

• > 35 años de duración OR 6.4

• > 10 años de suspender OR 1.5

Alcohol

• 35 – 55 bebidas a la semana OR 6.2

• > 84 bebidas a la semana 24.5Los dos factores de riesgo OR 130

Int J Cancer 2000;86:144–9

BMC Cancer 2008, 8:221

• 202 casos de CE

– 160 casos escamoso

– 42 casos adenoca

– Valencia, España

– Parte del estudio PANESOES

– Determinar los efectos independientes de diferentes bebidas alcohólicas y tipo de tabaco

Alcoholismo previo OR 16

25 – 74g/d OR 8

75g/d OR 23.2

Cerveza OR 7.98

Años de consumo 1-19 8.21

< 5 años de suspender OR 5.89

Fumador actual OR 2.28

>30 cigarros OR 5.82

Años de tabaquismo OR 2.03

IT > 50 OR 3.93

Hipótesis

• La exposición a la combinación de alcohol y tabaco tiene efectos sinérgicos en el riego de desarrollar carcinoma de células escamosas

Objetivo

• Resumir los datos disponibles en diversos estudios para estimar dicha sinergia

Métodos

• Protocolo

– Búsqueda sistemática de literatura en MEDLINE, EBM, EMBASE, ISI web of knowledge y BIOSIS

– Identificar estudios que evaluaran la presencia de ca de del escamosas en relación al uso de alcohol y tabaco

– Términos: neoplasia esofágica, cáncer escamoso, tabaco, mariguana, tabaquismo, alcoholismp, abuso de alcohol, bebidas alcoólicas

Métodos

• Selección de los estudios y extracción de datos

– Criterios de inclusión:

• Diagnóstico basado en histología o como cáncer esofágico en regiones de alta prevalencia de ca escamoso (Asia, África)

• El estatus de alcoholismo o tabaquismo estuviera bien definido

• Datos presentados como RR, OR o HR

• Estudios de cohorte o casos y controles

Métodos

• Recolección de datos en Microsoft access

• Dos autores revisaron

• Modelo de efectos máximos de manera binaria: USO vs no uso

• Y luego por separado de acuerdo a la intensidad

Métodos

• Calidad de los estudios

– Escala Newcastle-Ottawa

– 0-9

– Evaluado por los autores principales

• Estudios que cumplían con criterios de elegibilidad incluidos en el MA primario con análisis de sensibilidad

Análisis

• Se calculó el factor de sinergia para estimar la interacción en una escala multiplicativa entre el tabaco y el alcohol en el desenlace de carcinoma escamoso

OR T+A+

SF =OR T+A- (OR T-A+)

SF > 1 indica interacción entre los factores

Análisis

• MIX 2.0 Pro software

• Se estimo heterogeneidad de las estimaciones con Q de Cochrane, p < 0.10

• Se calculó I2 para medir grado de heterogeneidad

• Se realizaron modelos de efectos aleatorios

Resultados

• De los 5 estudios

– 3 casos y controles

– Todos de alta calidad de acuerdo a la escala de Newcastle- Ottawa ( 7 a 9 puntos)

Referencias identificadas

N= 9,668

Duplicados removidosN= 7,629

Excluidos N= 7,218

Evaluados

N= 411

Excluidos N= 406

Incluidos

N= 5

Tabaco solo

OR (1.2 – 2.3)

OR 1.36

I 2 0%

Alcohol solo

OR (0.8 – 4.79)

OR 1.21

I 2 34.4 %

Dos factores

OR (2.1 – 11.4)

OR 3.28

I 2 55.3 %

Sinergia(SF)

0.73 – 3.78

Total

1.99

OR 3.2895% CI = 2.11, 5.08

Discusión

• Es la primera revisión sistemática y meta-análisis que estima los efectos combinados del alcohol y tabaco

• Existe una fuerte sinergia entre el tabaco y el alcohol, más allá de los efectos multiplicativos

• Aunque la explicación de este efecto es poco clara

Discusión

• Un estudio que mide la concentración de acetaldehído en la saliva mostro que la adición de tabaco y alcohol, eleva los niveles hasta 7 veces comparado con el alcohol únicamente, sugiriendo un efecto sinérgico en la carcinogénesis del tracto GI superior

Limitaciones

• Todos los estudios son de Asia

• Diferentes tipos de cigarros y diferente contenido de carcinógenos, factor no controlable

• Solo un estudio midió intensidad

Fortalezas

• Búsqueda extensa de la literatura

• La estimación del factor de sinergia