Table 1 shows the relative proportions for age group 50 to 59 years old were 35% (105/300), followed by 25.7% (77/300) for those aged 40 to 49 years old, 24.3% (73/300) for those 60 to 69 years old and 9.3% (28/300) for those aged below 40 years old. The mean age of cases and controls were 54.06 ± 10.94 and 54.24 ± 11.11 years, respectively. The proportion of males diagnosed with NPC was higher than the females (77.0 % vs 23.0% respectively). Chinese made up 71.0% of cases. Slightly more than half (52.7%; 158/300) of the cases had normal BMI while 20% of them were underweight and 20% were overweight. Although the majority of controls had normal BMI (n = 119; 39.7%), the proportion of controls who were overweight or obese were higher than in cases (n =100; 33.3% and n = 68; 22.7% respectively). The median BMI of cases was significantly lower than the controls (21.62 kg/m2 vs. 25.93 kg/m2). The majority of the cases significantly engaged in the highest intensity of physical activity (fourth quartile; > 1775 hours/ week) while majority of controls (n = 84; 24.0%) were engaged with 80 to < 1775 hours of physical activity per week (third quartile). The median physical activity frequency was significantly higher among cases compared to controls (85.3 hours/week vs. 69.7 hours/ week). A significantly higher proportion of cases (51.7%) ever smoked, compared with controls. The proportion of cases that never consumed alcohol was lower than the controls (54.3% vs 62.3% respectively) but this was not significant. There was no association between consumption of high vitamin D content food such as beef, pork, fish and dairy products and NPC; however, consumption of egg and soy products once a week or more were [significantly higher among cases compared to controls (8.7% vs. 2.7%, χ2 = 10.1, p = 0.001 and 16.7% vs. 7.7%, χ2 = 11.37, p =0.001 respectively). Significantly higher proportion of cases ever consumed salted fish (76.0% vs 57.7%, χ2 = 22.75, p <0.0001). The median amount of salted fish intake was significantly higher among cases compared to controls (1.88 g vs 0.66 gm/day). The proportion of cases with a family history of NPC was significantly higher among cases than controls (12.3% vs. 2.0%, p < 0.001) (Table 1).
The mean serum concentration of 25(OH) D was significantly lower among cases compared to controls (25.3 ± 7.7 ng/mL vs. 27.0 ± 9.2 ng/mL, p = 0.016) (Table 2) but was not affected by the stage of NPC (Table 3). There was a significant inverse association between serum concentration of 25(OH) D and odds of getting NPC after adjusting for confounding (AOR = 0.73, 95% CI = 0.57 - 0.94) (Table 2).
Table 3 shows the 25(OH) D serum concentration for controls and cases across stages of cancer. The concentration among the cases is lower in the control. The differences however not significant.
Fig 1 illustrates the correlations between BMI, physical activity intensity, consumption of salted fish per day and serum 25(OH) D. Serum 25(OH) D concentration was inversely correlated with BMI and consumption of salted fish per day, however the relationships observed were not statically significant (r= - 0.039; p = 0.338 for BMI and r= - 0.029; p = 0.484 for consumption of salted fish per day). There was a weak positive correlation between physical activity intensity and serum 25(OH) D concentration (rho =0.097, p = 0.017).
The multivariable analysis revealed that having insufficient level of serum 25(OH) D, being underweight, salted fish consumption and having family history of NPC independently increased the odds of NPC. Having family history of NPC was significantly directly associated with NPC more than six-fold (AOR = 6.59, 95% CI = 2.63, 16.56). NPC cases had five-time more likelihood of being underweight compared to controls (AOR = 5.82, 95% CI = 2.79, 12.14). Consumption of salted fish independently doubled the odds of NPC (AOR = 2.40, 95% CI = 1.58, 3.65). After adjusting for potential confounders, this study revealed that people with serum 25(OH) D insufficiency significantly increased the odd of NPC by 73% (AOR = 1.73, 95% CI = 1.10, 2.74). Lifestyle factors such as smoking and alcohol consumption showed no independent association with NPC (Table 4).
Fig 2 shows the risk of NPC using fractional polynomial, where the resulting function OR(x) = 0.0121x4 - 0.0791x3 + 0.2298x2 – 0.4645x + 1.2384, shows a non-linear association, setting those with sufficient serum 25(OH) D (> 30 ng/ml) from the categorical analysis as reference. The black squares and lines indicate the odd ratios and 95% confidence intervals of the ordinal variables. The size of the black square indicates the effect size of the association at every step. A good fit was shown between the curve and the odds ratio estimates of the categorical analysis (R2 = 0.9957). Insufficient serum 25 (OH) D status with accumulation of each risk factor showed a dose response association with odds of getting NPC. There was a significant association between insufficient serum 25 (OH) D status with accumulation of four risk factors and increased odds of getting NPC. Increased NPC odds was observed after accumulation of second risk factors with presence insufficient serum 25 (OH) D status (OR = 0.54, 95% CI = 0.27, 4.77, OR = 1.04, 95% CI = 0.64, 1.72, OR = 1.15, 95% CI = 0.73, 1.80, OR = 1.93, 95% CI = 1.13, 3.31, and OR = 5.55, 95% CI = 1.67, 10.3 respectively).
Fig 3 shows the Kaplan-Meier estimates for 5 years survival time of NPC patients across the levels of serum 25(OH) D. The 5-year survival rates were 51.0%, 57.0% and 43.0% for serum 25(OH) D deficiency, insufficiency and sufficient respectively. However, no significant association was found between levels of serum 25(OH) D and survival of NPC.