Search

To allow for non-linear exposure-response relationships, we applied flexible non-parametric smoothing techniques to models of time to lung cancer mortality in two occupational cohorts with skewed exposure distributions. We focused on three different smoothing techniques in Cox models: penalized splines, restricted cubic splines, and fractional polynomials. We compared standard software implementations of these three methods based on their visual representation and criterion for model selection. We propose a measure of the difference between a pair of curves based on the area between them, standardized by the average of the areas under the pair of curves. To capture the variation in the difference over the range of exposure, the area between curves was also calculated at percentiles of exposure and expressed as a percentage of the total difference. The dose-response curves from the three methods were similar in both studies over the denser portion of the exposure range, with the difference between curves up to the 50th percentile less than 1 per cent of the total difference. A comparison of inverse variance weighted areas applied to the data set with a more skewed exposure distribution allowed us to estimate area differences with more precision by reducing the proportion attributed to the upper 1 per cent tail region. Overall, the penalized spline and the restricted cubic spline were closer to each other than either was to the fractional polynomial., ((c) 2007 John Wiley & Sons, Ltd.)

Many approaches have been taken to adjust for smoking in modeling cancer risk. In case-control studies, these metrics are often used arbitrarily rather than being based on the properties of the metric in the context of the study. Depending on the underlying study design, hypotheses, and base population, different metrics may be deemed most appropriate. We present our approach to evaluating different smoking metrics. We examine the properties of a new metric, "logcig-years", that we initially derived from using a biological model of DNA adduct formation. We compare this metric to three other smoking metrics, namely pack-years, square-root pack-years, and a model in which smoking duration and intensity are separate variables. Our comparisons use generalized additive models and logistic regression to examine the relationship between the logit probability of cancer and each of the metrics, adjusting for other covariates. All models were fit using data from a lung cancer study of 1,275 cases and 1,269 controls that has focused on gene-smoking relationships. There was a very significant, linear relationship between logcig-years and the logit probability of lung cancer in this sample, without any need to adjust for smoking status. These properties together were not shared by the other metrics. In this sample, logcig-years captured more information about smoking that is important in lung cancer risk than the other metrics. In conclusion, we provide a general framework for evaluating different smoking metrics in studies where smoking is a critical variable.

Aims: To illustrate the contribution of smoothing methods to modelling exposure-response data, Cox models with penalised splines were used to reanalyse lung cancer risk in a cohort of workers exposed to silica in California's diatomaceous earth industry. To encourage application of this approach, computer code is provided., Methods: Relying on graphic plots of hazard ratios as smooth functions of exposure, the sensitivity of the curve to amount of smoothing, length of the exposure lag, and the influence of the highest exposures was evaluated. Trimming and data transformations were used to down-weight influential observations., Results: The estimated hazard ratio increased steeply with cumulative silica exposure before flattening and then declining over the sparser regions of exposure. The curve was sensitive to changes in degrees of freedom, but insensitive to the number or location of knots. As the length of lag increased, so did the maximum hazard ratio, but the shape was similar. Deleting the two highest exposed subjects eliminated the top half of the range and allowed the hazard ratio to continue to rise. The shape of the splines suggested a parametric model with log hazard as a linear function of log transformed exposure would fit well., Conclusions: This flexible statistical approach reduces the dependence on a priori assumptions, while pointing to a suitable parametric model if one exists. In the absence of an appropriate parametric form, however, splines can provide exposure-response information useful for aetiological research and public health intervention.

p21 (Waf1/Cip1) is a downstream target of p53. We evaluated the association between p21 polymorphism (codon 31), p53 polymorphism (codon 72) and their corresponding in vivo mRNA expression. In this study, p21 and p53 genetic polymorphisms (using standard PCR-RFLP techniques) and p21 and p53 gene expressions (using a radiolabelled ribonuclease protection assay (RPA) technique) were evaluated in the peripheral leukocytes of 84 individuals (63 with lung cancer). Log-transformed values of mRNA expression by RPA, which approximated a normal distribution, were analyzed. p53 genotypes did not correlate with p53 mRNA log-expression (P>0.05 for all comparisons), but the Pro allele variants of p53 were associated with a significant decrease in mRNA log-expression of its downstream target, p21. The variant Arg allele of p21 was also associated with a significant decrease in p21 mRNA log-expression. When individuals with at least one variant allele of both p53 and p21 (double-variants) were compared with all other genotype groups, these double-variants had significantly lower log-expression of p21 (P<0.005 by both t-tests (crude) and linear regression analyses (adjusted)). This is translated into an approximate 48% reduction in the geometric mean of the mRNA expression of the double-variants, when compared with all other groups. Results were consistent in both patients with lung cancer (n=63) and in normal controls (n=21). In conclusion, the presence of a p53 Pro allele and/or p21 Arg allele is associated with lower downstream target gene expression of p21.

XRCC1 (X-ray cross-complementing group 1) and ERCC2 (excision repair cross-complementing group 2) are two major DNA repair proteins. Polymorphisms of these two genes have been associated with altered DNA repair capacity and cancer risk. We have described statistically significant interactions between the ERCC2 polymorphisms (Asp312Asn and Lys751Gln) and smoking in lung cancer risk. In this case-control study of 1091 Caucasian lung cancer patients and 1240 controls, we explored the gene-environment interactions between the XRCC1 Arg399Gln polymorphism, alone or in combination with the two ERCC2 polymorphisms, and cumulative smoking exposure in the development of lung cancer. The results were analyzed using logistic regression models, adjusting for relevant covariates. Overall, the adjusted odds ratio (OR) of XRCC1 Arg399Gln polymorphism (Gln/Gln versus Arg/Arg) was 1.3 [95% confidence interval (CI), 1.0-1.8]. Stratified analyses revealed that the ORs decreased as pack-years increased. For nonsmokers, the adjusted OR was 2.4 (95% CI, 1.2-5.0), whereas for heavy smokers (>/=55 pack-years), the OR decreased to 0.5 (95% CI, 0.3-1.0). When the three polymorphisms were evaluated together, the adjusted ORs of the extreme genotype combinations of variant alleles (individuals with 5 or 6 variant alleles) versus wild genotype (individuals with 0 variant alleles) were 5.2 (95% CI, 1.7-16.6) for nonsmokers and 0.3 (95% CI, 0.1-0.8) for heavy smokers, respectively. Similar gene-smoking interaction associations were found when pack-years of smoking (or smoking duration and smoking intensity) was fitted as a continuous variable. In conclusion, cumulative cigarette smoking plays an important role in altering the direction and magnitude of the associations between the XRCC1 and ERCC2 polymorphisms and lung cancer risk.

Excision repair cross-complementing group 2 (ERCC2), a major DNA repair protein, is involved in nucleotide excision repair and basal transcription. The ERCC2 polymorphisms have been associated with altered DNA repair capacity. We investigated two ERCC2 polymorphisms, Asp312Asn and Lys751Gln, in 1092 Caucasian lung cancer patients and 1240 spouse and friend controls. The results were analyzed using generalized additive models and logistic regression, adjusting for relevant covariates. The overall adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were 1.47 (95% CI, 1.1-2.0) for the Asp312Asn polymorphism (Asn/Asn versus Asp/Asp) and 1.06 (95% CI, 0.8-1.4) for the Lys751Gln polymorphism (Gln/Gln versus Lys/Lys). Gene-smoking interaction analyses revealed that the adjusted ORs for each of the two polymorphisms decreased significantly as pack-years increased. When comparing individuals with Asn/Asn + Gln/Gln versus individuals with Asp/Asp + Lys/Lys, the fitted ORs (95% CIs) were 2.56 (95% CI, 1.3-5.0) in nonsmokers and 0.69 (95% CI, 0.4-1.2) in heavy smokers (80 pack-years; P < 0.01 for the interaction term). Consistent and robust results were found when models incorporated different definitions of cumulative cigarette smoking. A stronger gene-smoking interaction was observed for the Asp312Asn polymorphism than for the Lys751Gln polymorphism. In conclusion, cumulative cigarette smoking modifies the associations between ERCC2 polymorphisms and lung cancer risk.

N-acetyltrasferase-2 (NAT2) and microsomal epoxide hydrolase (mEH) are polymorphic genes that metabolize different tobacco carcinogens. Smaller studies found inconsistent relationships between NAT2 or mEH polymorphisms and lung cancer risk. To determine whether there is gene-environment interaction between NAT2 polymorphisms, alone or in combination with mEH polymorphisms, and cumulative smoking exposure in the development of lung cancer, we conducted a case control study of 1115 Caucasian lung cancer patients and 1250 spouse and friend controls. The results were analyzed using generalized additive models and logistic regression, adjusting for relevant covariates. There was no overall relationship between NAT2 genotype and lung cancer risk; the adjusted odds ratio (OR) of the rapid versus slow acetylator genotypes was 0.96 [95% confidence interval (CI), 0.79-1.16]. However, gene-environment interaction analyses revealed that the adjusted ORs increased significantly as pack-years increased. For nonsmokers, the fitted OR was 0.66 (95% CI, 0.44-0.99), whereas for heavy smokers (80 pack-years), the OR increased to 1.22 (95% CI, 0.89-1.67). When comparing the extreme genotype combinations of the NAT2 rapid acetylator, higher mEH activity genotype to the NAT2 slow acetylator, and very low mEH activity genotype, the corresponding ORs at 0 and 80 pack-years were 0.30 (95% CI, 0.14-0.62) and 2.19 (95% CI, 1.26-3.81), respectively. Results were similar with ORs derived from stratified models. In conclusion, NAT2 rapid acetylator genotypes are protective against lung cancer in nonsmokers but are risk factors in heavy smokers. The joint effects of NAT2 and mEH polymorphisms are consistent with an independent, additive effect of these two genes, modified by smoking history.

Traditionally, non-small cell lung cancer (NSCLC) has been evaluated as a unique entity in genotyping studies. However, recent biological data suggest that different NSCLC subtypes, specifically adenocarcinomas (AC) and squamous cell carcinomas (SCC), differentially alter cancer behavior. Several studies have associated a p53 polymorphism at codon 72 with NSCLC susceptibility. This study investigated whether different p53 genotypes altered the overall risk of developing AC versus SCC. Polymorphisms in metabolizing enzymes, together with prolonged exposure to tobacco carcinogens, can result in accumulation of DNA damage; these effects may potentiate the effects of subtle differences in p53 function. Thus, interactions between polymorphisms of p53 and either GSTM1 or GSTT1 were also evaluated. We analyzed 1168 incident lung cancer cases and 1256 control subjects using multiple logistic regression. Histological data were available for 1144 cases (98%): 585 with AC, 284 with SCC, and 275 with other histological subtypes (large cell, small cell, mixed, and other). An increase in the NSCLC risk posed by the p53 Pro allele (versus Arg/Arg) was seen in AC compared with controls [adjusted odds ratio (OR), 1.36; 95% confidence interval (CI), 1.1-1.7] but not in SCC (adjusted OR, 1.04; 95% CI, 0.8-1.4). Among AC and SCC cancer patients, individuals with the GSTM1-null genotype had an OR of 1.80 (95% CI, 1.1-2.8; case-only analysis) of having AC versus SCC if they also carried a p53 Pro allele. We conclude that different genotype combinations of p53 and GSTM1 increase the risk of developing specific histological subtypes of NSCLC.

Microsomal epoxide hydrolase (mEH) is involved in the metabolism of environmental and tobacco carcinogens. Smaller studies found inconsistent results in the relationship between mEH polymorphisms and lung cancer risk. We investigated the two polymorphisms of mEH in 974 Caucasian lung cancer patients and 1142 controls using PCR-RFLP techniques. The results were analyzed using generalized additive models and logistic regression, adjusting for relevant covariates. There was no overall relationship between mEH genotypes and lung cancer risk. The adjusted odds ratio (OR) of the very low activity genotype versus that of other genotypes combined was 1.00 [95% confidence interval (CI), 0.74-1.34]. However, gene-environment interaction analyses revealed that the ORs decreased as cumulative smoking (defined as square root of pack-years) increased. When pack-years = 0, the OR was 1.89 (95% CI, 1.08-3.28). When pack-years = 28.5, the OR was 1.00 (95% CI, 0.76-1.32), and when pack-years = 80, the OR decreased to 0.65 (95% CI, 0.42-1.00). When cases were stratified according to histological subtypes, the interaction between mEH genotype and cumulative smoking was statistically significant (P < 0.01) for the 222 squamous cell carcinoma cases, whereas it was not significant (P = 0.18) for the 432 adenocarcinoma cases. In conclusion, cumulative cigarette smoking plays a pivotal role in the association between mEH polymorphisms and lung cancer risk, altering the direction of risk (in the case of the very low activity genotype) from a risk factor in nonsmokers to a relatively protective factor in heavy smokers.

We conducted a hospital-based case-control study of 814 lung cancer patients and 1123 controls to examine the association of the NAD(P)H: quinone oxidoreductase 1 (NQO1) gene polymorphism with lung cancer susceptibility. Using PCR-RFLP genotyping assay techniques, we analyzed DNA samples to detect the variant forms of the NQO1 gene in exon 6 on chromosome 16q. We examined the relationship between lung cancer odds and NQO1 genotypes after adjusting for age, gender, and smoking behavior using generalized additive modeling. We found no overall association between NQO1 genotypes and lung cancer susceptibility, regardless of age, gender, family history of cancer, or histological cell type. However, our data demonstrated that in both former and current smokers, there was an association between NQO1 genotypes and lung cancer susceptibility that was dependent upon cigarette smoking duration and smoking intensity. For both current and former smokers, smoking intensity was more important in predicting cancer risk than smoking duration for all of the genotypes. Among former smokers, individuals with the T/T genotype were predicted to have a greater cancer risk than those with the C/C genotype for smoking durations up to 37 years. The predicted cancer risk for former smokers with the C/T versus T/T genotype depended on both smoking intensity and smoking duration. Our results support the concept that differential susceptibility to lung cancer is a function of both an inheritable trait in NQO1 metabolism and individual smoking characteristics.

Background: DNA adducts formed as a consequence of exposure to tobacco smoke may be involved in carcinogenesis, and their presence may indicate a high risk of lung cancer. To determine whether DNA adducts can be used as a "dosimeter" for cancer risk, we measured the adduct levels in nontumorous lung tissue and blood mononuclear cells from patients with lung cancer, and we collected data from the patients on their history of smoking., Methods: We used the 32P-postlabeling assay to measure aromatic hydrophobic DNA adducts in nontumorous lung tissue from 143 patients and in blood mononuclear cells from 54 of these patients. From the smoking histories, we identified exposure variables associated with increased DNA adduct levels by use of multivariate analyses with negative binomial regression models. RESULTS/ CONCLUSIONS: We found statistically significant interactions for variables of current and former smoking and for other smoking variables (e.g., pack-years [number of packs smoked per day x years of smoking] or years smoked), indicating that the impact of smoking variables on DNA adduct levels may be different in current and former smokers. Consequently, our analyses indicate that models for current and former smokers should be considered separately. In current smokers, recent smoking intensity (cigarettes smoked per day) was the most important variable. In former smokers, age at smoking initiation was inversely associated with DNA adduct levels. A highly statistically significant correlation (r=.77 [Spearman's correlation]; two sided P<.001) was observed between DNA adduct levels in blood mononuclear cells and lung tissue., Implications: Our results in former smokers suggest that smoking during adolescence may produce physiologic changes that lead to increased DNA adduct persistence or that young smokers may be markedly susceptible to DNA adduct formation and have higher adduct burdens after they quit smoking than those who started smoking later in life.