饮酒与喉癌：整体和饮用量-风险关系的系统性回顾和荟萃分析

出处：Oral Oncol. 2010 Nov;46(11):802-10. Epub 2010 Sep 15.
作者：Islami F, Tramacere I, Rota M, Bagnardi V, Fedirko V, Scotti L, Garavello W, Jenab M, Corrao G, Straif K, Negri E, Boffetta P, La Vecchia C.
PMID：20833578
译者： F1000因子：0 评级：必读

    饮酒是喉癌的一个已知危险因素。然而，目前关于轻度饮酒与喉癌风险的资料很少。为了阐明这一问题，我们采用两种方法进行荟萃分析：（1）通过随机效应模型重建饮酒类别，使用Hamling法根据预先设定的饮酒量计算风险估计值；（2）随机效应多元回归模型。在PubMed数据库中检索所有关于饮酒与喉癌风险的相关英文病例对照研究或队列研究。40项报告了至少3个饮酒水平的研究（38项病例对照研究，2项队列研究）纳入分析。总体而言，饮酒者相对于非饮酒者，喉癌的风险大约增加2倍（RR=1.90; 95％置信区间：1.59-2.28）。虽然轻度饮酒（≥1次 /日）与喉癌风险无任何显著相关性（12项研究 RR= 0.88; 95％CI： 0.71-1.08），但适量饮酒（每天饮酒>1次、 <4次）喉癌风险增加1.5倍（35项研究RR= 1.47; 95％CI 1.25-1.72）。酗酒（4次/日）使喉癌风险增加2.5倍（33项研究 RR= 2.62，95％CI  2.13-3.23）。校正了主要潜在混杂因素（年龄、性别、吸烟）的亚组分析和几项进一步亚组分析的结果类似，证实了结果的可靠性。 

Alcohol drinking is a known risk factor for laryngeal cancer. However, little information is available on the risk associated with light alcohol drinking. To address this issue, we conducted a meta-analysis using two methods: (i) random-effects models with reconstruction of alcohol consumption categories and calculation of risk estimates associated with predefined consumption levels using Hamling method and (ii) random-effects meta-regression models. The PubMed database was searched for all case-control or cohort studies published in the English language on the association between alcohol consumption and risk of laryngeal cancer. Forty studies (38 case-control, 2 cohort) reporting on at least three levels of consumption were included. Overall, alcohol drinking versus non-drinking was associated with an approximately 2-fold increase in risk of laryngeal cancer (RR=1.90; 95% CI: 1.59-2.28). While light alcohol drinking (≥1 drink/day) did not show any significant association with risk of laryngeal cancer (12 studies. RR=0.88; 95% CI: 0.71-1.08), moderate drinking (>1 to <4drinks/day) was associated with a 1.5-fold increase in risk (35 studies. RR=1.47; 95% CI: 1.25-1.72) and heavy drinking (⩾4drinks/day) was associated with a 2.5-fold increased risk (33 studies. RR=2.62; 95% CI: 2.13-3.23). Subgroup analyses for studies that adjusted for main potential confounding factors (age, sex, and tobacco use) and several further subgroup analyses showed similar results, which suggest the robustness of the results.

专家评价： This interesting meta-analysis including 40 selected publications provides convincing statistical evidence confirming the quantitative relationship between alcohol drinking and development of laryngeal cancer suggested by individual reports. The study provides clear figures expressing the relative risk of developing laryngeal cancer related to the amount of alcoholic drinks per day – for light alcohol drinking (/=4 drinks) – that are easily applicable in consulting individual patients and patient populations. The association between alcohol and the risk of laryngeal cancer has been suggested for over 1 century. Various reports confirmed this association but only a few investigated the dose-risk association between alcohol consumption and laryngeal cancer. Using a meta-analytic approach, the authors retained 40 original high-quality studies out of 362 publications retrieved from an extensive literature research. Using a thorough epidemiologically sound analysis, the authors concluded that alcohol drinking in general was associated with an approximately 2-fold increase in risk of laryngeal cancer. Light alcohol drinking is not associated with an elevated risk, whereas moderate and heavy drinking is associated with a 1.5-fold and 2.5-fold risk, respectively. 

Strengths: 
This thorough meta-analysis correcting as much as possible for confounding effects summarizes very nicely the results of the included studies so far performed. The dose-effect relationship and the high prevalence of the risk factor strongly support a true causal relationship. 

Weaknesses: 
Most of the included articles (38/40) are case-control studies, which widely differ in study design and variables that results were controlled for. Consequently, subgroup analyses concerning relative risks for main potential confounding factors (age, sex and tobacco use) are based on a rather limited number of studies. Also, data available for light alcohol drinking were derived from only 12 out of 40 studies, whereas data concerning moderate and high intake were retrieved from 35/40 and 33/40 studies, respectively, and no distinction is made between non-drinkers, sporadic alcohol consumers and people who regularly drink alcohol but no more than up to one drink per day. One other weakness remains the chosen outcome 'laryngeal cancer', and the study does not differentiate between the subsites (supraglottis versus glottis and subglottis). It is well known that the subsites in the larynx are to a very different degree exposed to alcohol in patients using alcohol and, consequently, the carcinogenetic effect differs among subsites. In this respect, glottic cancers (65% of laryngeal cancer) have much less carcinogenetic effect (none at all in the study of Brugere et al. {1}), whereas for supraglottic cancers (30% of laryngeal cancers) the effect is much higher. 

Future prospective large-scaled studies would be highly valuable to confirm the study findings, to separate the effect on glottic cancer from that on supraglottic cancer and to further investigate the impact of different potential confounding factors. References:  {1} Brugere et al. Cancer 1986, 57:391-5 [PMID:3942973].  

专家评价：

    这是一项有趣的荟萃分析，纳入了40项研究，提供了令人信服的统计证据证实了单个研究所报道的饮酒和喉癌发生的定量关系。这项研究提供了明确数据表明每日饮酒量与喉癌发生的相对风险——轻度饮酒（每日/ = 4次）——这在个别患者和患者群中是适用的。
饮酒与喉癌风险之间存在关联这一观点的提出已有百年。多个报告证实了这一关联，但是只有少数研究探讨了饮酒量和喉癌风险的关系。使用荟萃分析方法，本文作者在广泛文献检索得到的362篇文献中抽提出40篇高品质论著。通过深入流行病学分析，作者认为总体而言饮酒使喉癌风险增加大约2倍。轻度饮酒不增加喉癌风险，而中度饮酒使喉癌风险增加1.5倍，重度饮酒则增加2.5倍。


优势：

这一透彻的荟萃分析校正了尽可能多的混杂影响，对纳入的迄今进行的研究进行了非常好的总结。剂量效应关系和危险因素高度流行强烈支持饮酒与喉癌间具有真正的因果关系。


劣势：

所纳入的大部分论文（38/40）是病例对照研究，这些研究设计变异大，研究结果中的对照变量差异大。因此，关于主要潜在混杂因素（年龄、性别和吸烟）相对风险的亚组分析是基于数量有限的研究。此外，轻度饮酒的数据来自于40项研究中的12项，中度饮酒的数据来自35项研究，重度饮酒的数据来自33项研究，不饮酒者、偶尔饮酒和规律饮酒但每日最多不超过一次者的喉癌风险无差异。另外一个局限性仍然是所选择的结果“喉癌”，这项研究并没有区分部位（声门、声门下与声门上）。众所周知，对于饮酒者而言，喉的不同部位接触酒精的程度是显著不同的，因此，饮酒对不同部位的致癌作用不同。饮酒对声门癌（占喉癌的65％）的致癌作用较小（在Brugere等人的研究中无致癌作用{1}），而对声门上型癌（占喉癌的30％）的致癌作用大。


将来进行大型前瞻性研究对于证实这些研究发现将是非常有价值的，以区分饮酒对声门癌和声门上型癌的不同效应，并进一步探讨不同潜在混杂因素的影响。

参考文献

{1} Brugere et al. Cancer 1986, 57:391-5 [PMID:3942973].