Nasopharyngeal carcinoma and smoking : evidence from observational and genetic studies

Abstract

Nasopharyngeal carcinoma (NPC) is rare globally but has high incidence and mortality in North Africa and Southeast Asia, especially in Southern China, such as Hong Kong and Guangzhou. NPC is a highly malignant, fast progressing and asymptomatic cancer deeply located in the centre of the head. These characteristics make it difficult to detect at early stages with over 70% of NPC in high-risk regions diagnosed at advanced stages (stage III and IV), creating a heavy disease burden to society.

Tobacco is a well-known causal factor for head and neck cancers, but not for NPC. In both high- and low-risk regions, most studies on the association between smoking and NPC were case-control studies and prospective cohort studies were few. Significant positive associations between smoking and NPC were mainly observed in low-risk regions, but findings in high-risk regions were mixed. The distinct geographic distribution of high- and low-risk regions supports that the association between NPC and smoking should better be examined separately in high- and low-risk regions. Therefore, the present thesis focuses on high-risk regions.

After conducting a comprehensive critical literature review, I found insufficient high-level evidence to support smoking as a causal factor for NPC. Hence, I conducted a series of epidemiological (observational) studies from low- to high-level of evidence to examine the association in high-risk regions. The Hong Kong Multicentre Case-control Study with 821 NPC incident cases and 1,868 controls (1,512 hospital controls and 356 community controls) was analysed to evaluate the association of active smoking, passive smoking and smoking cessation with the risk of NPC. To understand the mechanism of how smoking causes NPC, genetic analysis of the nicotine dependence single nucleotide polymorphisms- rs1051730 was also conducted. The Guangzhou Biobank Cohort Study (GBCS), a large prospective study (n=26,399) in the NPC high-risk region was analysed to evaluate the association of smoking and NPC risk in elderly adults (aged 50 or above) with 51 NPC incident cases and NPC deaths. The individual participant data (IPD) meta-analysis is considered to be the ‘gold standard’ of systematic review and can provide the strongest observational evidence to support causality. In the present thesis, the IPD meta-analysis on cohort studies of NPC in high-risk regions was the first IPD meta-analysis to study the association between smoking and NPC.

The results from the present Hong Kong case-control study, GBCS and the IPD meta-analysis found sufficient evidence that active smoking is causally associated with NPC in high-risk regions. Passive smoking is probably associated with an increased risk of NPC. Smoking cessation is probably associated with a lower risk of NPC. The harms of active and passive smoking in NPC and the benefit of smoking cessation highlight the importance of implementing strict tobacco control strategies in high-risk regions.