Selection of studies
We searched MEDLINE using PubMed, Google Scholar and Science Direct for association studies as of August 17, 2019. The terms used were “Lysyl oxidase”, “LOX”, “polymorphism” and “cancer” as medical subject headings and text. References cited in the retrieved articles were also screened manually to identify additional eligible studies. Inclusion criteria were (i) case-control studies evaluating the association between LOX polymorphisms and cancer risk and (ii) sufficient genotype frequency data presented to calculate the odds ratios (ORs) and 95% confidence intervals (CIs). The exclusion criteria were as follows: (i) reviews; (ii) studies whose control frequencies deviated from the Hardy-Weinberg equilibrium (HWE); (iii) articles that were not case-control studies; and (iv) studies with unusable genotype data.
Data extraction
Two investigators (RM and NP) independently extracted data and arrived at consensus. The following information was obtained from each publication: first author’s name, the year of publication, the country of origin, ethnicity, cancer type, study design, studies that matched their controls with cases and the criteria used, sample sizes and genotype frequencies.
Methodological quality of the studies
We used the Clark-Baudouin (CB) scale to evaluate the methodological quality of the included studies [19]. The CB scale is based mainly on statistical (P-values, statistical power, and corrections for multiplicity) and genetic (genotyping methods and the HWE) criteria. In this scale, low, moderate and high quality have CB scores of < 5, 5-6 and ³ 7, respectively.
Data distribution and power calculations
Data distribution was assessed with the Shapiro-Wilks test using SPSS 20.0 (IBM Corp., Armonk, NY, USA). Gaussian (normal) distribution (P > 0.05) warranted the use of the mean ± standard deviation (SD). Otherwise, the median (with interquartile range) was used. Using the G*Power program [20], we evaluated statistical power as its adequacy bolsters the level of associative evidence. Assuming an OR of 1.5 at a genotypic risk level of α = 0.05 (two-sided), power was considered adequate at ≥ 80%.
HWE
HWE was assessed with the application in https://ihg.gsf.de/cgi-bin/hw/hwa1.pl. A P-value of < 0.05 indicated deviation from the HWE. Deviations were found in six studies [14-18, 21] and were thus excluded from the analysis (Table S1). Table S2 includes a column that details the nonsignificance (P > 0.05) of the HWE-compliant studies.
Data synthesis
Cancer risks (ORs and 95% CIs) were estimated for each study using the following genetic models: (i) homozygous [H], (ii) recessive [R], (iii) dominant [D], and (iv) codominant [C]. Comparing the effects on the same baseline, we calculated pooled ORs and 95% CIs. In addition to the overall analysis, we examined two subgroups: Asians (3,834 cases/4,061 controls) and cancer type. The latter was stratified into gastrointestinal (GI) (1,453 cases/1,546 controls) and breast (935 cases/923 controls) cancer. The strength of evidence was assessed with four indicators: First, the magnitude of effects are higher or lower when the pooled ORs are farther from or closer to the OR value of 1.0 (null effect), respectively [22]. Second, associative P-values (Pa) with more zero decimals (e.g., 0.00001) are considered to be stronger than those closer to 0.05. Third, statistical significance found across the comparisons and genetic models indicates the consistency of effects. Fourth, homogeneity is preferred to heterogeneity, but heterogeneity is unavoidable [23]. This is because conclusions in the milieu of homogeneity (or at least non-heterogeneity) have greater evidential strength than those that are heterogeneously derived. Nevertheless, the presence of heterogeneity between studies was estimated with the c2-based Q test [24] at a threshold of significance set at Pb < 0.10. Heterogeneity was quantified with the I2 statistic, which measures variability between studies [25]. I2 values of > 50% indicate more variability than those £ 50%, with 0% indicating zero heterogeneity (homogeneity). Similarities in population features of the studies warranted using the fixed-effects (Fe) model [26]; otherwise, the random-effects (Re) model [27] was used. Sensitivity analysis, which involves the omission of one study at a time and the recalculation of the pooled OR, was used to test the robustness of the summary effects. Publication bias was not assessed because none of the comparisons had ≥ 10 studies. Less than this number presents low sensitivity of the publication bias tests [28]. Except for heterogeneity estimation [24], two-sided P-values of £ 0.05 were considered significant. All associative outcomes were Bonferroni corrected. Data for the meta-analysis were analyzed using Review Manager 5.3 (Cochrane Collaboration, Oxford, England), SIGMASTAT 2.03, and SIGMAPLOT 11.0 (Systat Software, San Jose, CA).