Family History of Cancer and Head and Neck Cancer Risk in a Chinese Population.

BACKGROUND
The aim of this study was to investigate whether family history of cancer is associated with head and neck cancer risk in a Chinese population.


MATERIALS AND METHODS
This case-control study included 921 cases and 806 controls. Recruitment was from December 2010 to January 2015 in eight centers in East Asia. Controls were matched to cases with reference to sex, 5-year age group, ethnicity, and residence area at each of the centers.


RESULTS
We observed an increased risk of head and neck cancer due to first degree family history of head and neck cancer, but after adjustment for tobacco smoking, alcohol drinking and betel quid chewing the association was no longer apparent. The adjusted OR were 1.10 (95% CI=0.80-1.50) for family history of tobacco-related cancer and 0.96 (95%CI=0.75-1.24) for family history of any cancer with adjustment for tobacco, betel quid and alcohol habits. The ORs for having a first-degree relative with HNC were higher in all tobacco/ alcohol subgroups.


CONCLUSIONS
We did not observe a strong association between family history of head and neck cancer and head and neck cancer risk after taking into account lifestyle factors. Our study suggests that an increased risk due to family history of head and neck cancer may be due to shared risk factors. Further studies may be needed to assess the lifestyle factors of the relatives.


Introduction
In 2012, approximately 599,600 head and neck cancer (HNC) cases were diagnosed and 324,000 deaths due to head and neck cancer occurred (Ferlay et al., 2013). While alcohol drinking and tobacco smoking are the major risk factors for HNC, family history of cancer may also play an important role in the risk of HNC (Negri et al., 2009). The International Head and Neck Cancer Epidemiology (INHANCE) consortium reported that family history of HNC increased the risk of HNC by 1.68-fold (95% CI 1.23-2.29; 9,025 cases and 13,739 controls) with genetics or shared lifestyle factors.
Previous case-control studies in China have investigated risk factors for HNC. A hospital based casecontrol study of 404 case-control pairs in Beijing included laryngeal cancer cases diagnosed between 1989-1990 and controls matched by age and gender (Zheng et al., 1990). A population-based case-control in Shanghai, China from 1988 to 1990, included 204 oral cancer cases and 414 controls (Zheng et al., 1992). They reported that of the oral squamous cell carcinoma cases, 34% were attributed to tobacco smoking and 24% were attributed to alcohol drinking. Family history of cancer was not investigated in these studies, to our knowledge. A more recent large case-control study of oral cancer reported that green tea consumption may be protective against oral cancer, particularly in men and in smokers (Fu et al., 2013). One study in the Liaoning province of China reported an OR of 2.0 (95%CI=1.3-3.2) for any family history of cancer among laryngeal cancer patients (288 cases and 298 controls) without adjustment and an OR of 2.3 (95%CI=1.2-4.5) for family history of malignancy after adjustment for various factors (Li et al., 2009).
The aim of our study is to investigate the association between family history of cancer and HNC risk in a Chinese population, with adjustments for shared lifestyle factors such as tobacco smoking and alcohol drinking.

Materials and Methods
This study in East Asia is a case-control study including eight centers (Beijing, Fujian, Henan, Jiangsu, Liaoning, Shanghai, Sichuan, and Taiwan). Between December 2010 to November 2013, 921 incident cases of HNC cases, including oral cavity, oropharynx hypopharynx larynx and 806 controls were recruited. The interview of both cases and controls were structured to obtain information on current and previous alcohol consumption, dietary habits, tobacco consumption and other lifestyle factors. Blood samples were collected from cases and controls whenever possible. Written consent for participation was obtained from all study participants. Ethical approval for human subject research was obtained at the University of Utah, Fujian, Henan, Shanghai, Sichuan, Taiwan, and Beijing.
The number of brothers and sisters, the number of first-degree relatives (parent, siblings, and children) with a history of cancer, the site of the cancer and the type of affected relative were included in information on family history. A subject with a family history for a given cancer was considered if at least one affected first-degree relative was reported by the subject.
Statistical Methods. The odds ratios (OR) and 95% confidence intervals (CI) were estimated using logistic regression. The adjustment variables included center, age (categories as shown in Table 1), sex, education (categories as shown in Table 1), cigarette smoking (categorical packyears, never smoker/1-19/20-39/≥40), alcohol drinking frequency (never, <2 drinks/day, ≥2 drinks/day), number of sisters, and number of brothers where appropriate. We assessed family history of HNC, tobacco-related cancers (lung, nasopharynx, nasal cavity, paranasal sinuses, esophagus, stomach, pancreas, liver, kidney, urinary bladder, uterine cervix, bone marrow (IARC, 2012), and any cancer. We additionally adjusted on years of betel quid use for the analysis of family history of head and neck cancer.

Results
The demographic characteristics of 921 cases and 806 controls in the case-control study are presented in Table  I. Taiwan contributed a large number of cases (482 cases, 52.3%). There were more male cases (726 cases, 78.8%) compared to female cases, and more cases with oral cavity cancer compared to other cancer sites.
In the crude model, we observed a 3.45-fold increase in the risk of HNC due to family history of HNC (Table 2). The strong association was not observed after we adjusted for center, age, sex, tobacco smoking, alcohol drinking, betel quid chewing and number of brothers and sisters. None of the female controls had family history of HNC, thus we did not estimate odds ratios for this subgroup. When comparing cancer sites, the risk due to family history of HNC was strong for oral cavity, hypopharynx cancers as well as cancers in overlapped HNC sites in the crude analysis but not after adjustment for lifestyle factors.
Family history of tobacco-related cancers conferred an OR of 1.37 (95%CI=1.04-1.81) in the crude model, but when adjusted for tobacco smoking, alcohol drinking and other factors, the association was no longer statistically significant (Table 3).
Also, the OR was 1.70 (95% CI=1.00-2.89) for family history of tobacco-related cancer due to relative with the cancer diagnosis <50 years in the crude model, but after adjustment for confounders, the association was not observed. There was an increased risk observed for siblings being the type of affected relative for family history of tobacco-related cancers even after adjustment for tobacco and alcohol (OR=1.65, 95%CI=1.02-2.68).
Family history of any cancer was not associated with HNC risk (Table 4) Figure 1. The OR, for having a first-degree relative with HNC was higher in tobacco/alcohol subgroups; increasing from 2.10 to 2.95 in users of tobacco only, and from 5.33 to 16.73 in alcohol and tobacco users. The p-value of interaction for family history of HNC with smoking and drinking habits was 0.5303. The corresponding p-values of interaction for tobacco, alcohol were 0.3309 with family history of tobacco related cancers and 0.9781 with family history of all cancers.

Figure 1. Odds Ratios for Family History of Head and Neck Cancers by Alcohol and Tobacco Consumption.
Adjusted for center, age, sex, number of sisters and brothers  Previous studies have reported that family history of HNC in first-degree relatives increased the risk of HNC. Our results support the association between family history of HNC in first-degree relatives and the risk of HNC, but with adjustment for tobacco, betel quid and alcohol habits the association was not persistent.
While the INHANCE consortium results showed that the HNC risk was higher for more distal sites (hypopharynx OR=2.28, 95%CI=1.4-3.54 and larynx OR=2.07, 95%CI=1.57-2.73), the crude estimates in our study suggest that the HNC risk was higher in the hypopharynx and oropharynx. However, our study sample size is smaller and thus the confidence intervals were fairly wide, making these comparisons difficult. In the Utah population, for first-degree relatives of a cancer patient, the RRs for cancer of the same site were 6.0 (95% CI 2.37-12.65) for tongue cancer, 7.9 (95% CI 3.46-15.67) for pharyngeal cancer, and 5.0 (95% CI 2.59-8.65) for laryngeal cancer (Teerlink et al., 2012). In the Swedish Family Cancer database, the standardized incidence ratios (SIRs) for upper aerodigestive tract cancer in offspring by parental cancer was 1.40 (95%CI=0.98-1.95) for head and neck cancers (Li et al., 2003). In the Icelandic population, for first-degree relatives of a cancer patient, the RRs for cancers of the same site were 5.04 (95% CI 2.75-9.52) for lip cancer, 3.02 (95% CI 1.06-6.65) for larynx cancer (Amundadottir et al., 2004). Differences in risk due to the head and neck cancer subsite may be difficult to discriminate, even in large-scale studies.
Although family history of tobacco-smoking-related cancer was associated with a modest increase in risk (OR= 1.11, 95% CI 1.01-1.23) in the INHANCE analysis, we did not observe a clear association with family history of tobacco-related cancers or with family history of cancer in general. In the Utah population, first degree relatives of had increased risks of laryngeal cancer (Negri et al., 2009). In the Icelandic population, first degree relatives of esophageal, lip, lung and pharyngeal cancer had an increased risk of laryngeal cancer; first degree relatives of breast, cervical, and lung cancers had a increased risk of pharyngeal cancer; and first degree relatives of anal, esophageal, lip and lung cancer had an increased risk of tongue cancer. Although family history of tobacco-related cancers may increase the risk of head and neck cancers, the relative risks are generally lower than those for family history of head and neck cancer. Our study may not have had an enough statistical power to detect an association.
In our study, the interaction analysis was suggestive of a higher risk of HNC due to family history of head and neck cancer among tobacco and alcohol users, although the p-value for interaction was not significant. In the INHANCE analysis, the odds ratio increased from 3.34 (95% CI 2.90-3.86) to 7.21 (95% CI 5.46-9.54) for both tobacco and alcohol users. In our study, the odds ratio increased from 5.33(95%CI=3.88-7.33) to ). Similar to the INHANCE analysis, we did not observe a statistically significant association between family history of HNC and HNC risk among never drinkers and never cigarette smokers.
There are several possible limitations in this study.
Recall bias is a potential limitation since patients had already been diagnosed with cancer and may have recalled family history more carefully. It seems unlikely that cases or controls would forget to report on any incident cases of family history of cancer in their first degree relatives.
Although we had information about first-degree relatives, we did not have information on cancer among second-or third-degree relatives. Due to the small number of subjects with family history of cancer, we had lower statistical power to detect moderate risks and we were unable to estimate some odds ratios in a stratified analysis. Having information on whether the relatives smoked tobacco or drank alcohol would have been of interest to further clarify whether associations with family history of cancer were due to shared lifestyle habits.
The key strengths of the study include the fairly large sample size of 921 cases and 806 controls in our case-control study. Another strength of our study is the detailed tobacco, betel quid and alcohol information, which allowed us to adjust on lifestyle factors that may contribute to the association between family history of cancer and cancer risk. We were also able to explore potential effect modification, i.e., the impact of family history of HNC stratified by tobacco and alcohol. Also, in this case-control study, we adjusted for the number of brothers and sisters in the assessment of the associations. Finally, to our knowledge, this is the first investigation of family history of HNC in an East Asian population.
In conclusion, we observed a four-fold increase in the risk of head and neck cancer due to family history of head and neck cancer but after adjustment for lifestyle factors the association was no longer observed. Family history of tobacco-related cancers and cancer in general were also not associated with an increased risk of head and neck cancer after adjustment for tobacco and alcohol habits. Our results support that shared lifestyle factors in a family are likely to play an important role for head and neck cancer risk due to family history of cancer. Further large scale studies in this population may be needed to further discern risk differences due to family history of head and neck cancer, for the head and neck cancer subsites and to assess effect modification due to tobacco and alcohol habits.