A Prospective Study of Organochlorines in Adipose Tissue and Risk of Non-Hodgkin Lymphoma

Background: Exposure to organochlorines has been examined as a potential risk factor for non-Hodgkin lymphoma (NHL), with inconsistent results that may be related to limited statistical power or to imprecise exposure measurements. Objective: Our purpose was to examine associations between organochlorine concentrations in prediagnostic adipose tissue samples and the risk of NHL. Methods: We conducted a case–cohort study using a prospective Danish cohort of 57,053 persons enrolled between 1993 and 1997. Within the cohort we identified 256 persons diagnosed with NHL in the population-based nationwide Danish Cancer Registry and randomly selected 256 subcohort persons. We measured concentrations of 8 pesticides and 10 polychlorinated biphenyl (PCB) congeners in adipose tissue collected upon enrollment. Associations between the 18 organochlorines and NHL were analyzed in Cox regression models, adjusting for body mass index. Results: Incidence rate ratios and confidence intervals (CIs) for interquartile range increases in concentrations of dichlorodiphenyltrichlorethane (DDT), cis-nonachlor, and oxychlordane were 1.35 (95% CI: 1.10, 1.66), 1.13 (95% CI: 0.94, 1.36), and 1.11 (95% CI: 0.89, 1.38), respectively, with monotonic dose–response trends for DDT and cis-nonachlor based on categorical models. The relative risk estimates were higher for men than for women. In contrast, no clear association was found between NHL and PCBs. Conclusion: We found a higher risk of NHL in association with higher adipose tissue levels of DDT, cis-nonachlor, and oxychlordane, but no association with PCBs. This is the first study of organochlorines and NHL using prediagnostic adipose tissue samples in the exposure assessment and provides new environmental health evidence that these organochlorines contribute to NHL risk.


Methods S2
Extraction and gas chromatography-mass spectroscopy S2 Quality Assurance/Quality Control of PCB Analyses S2 :   Table S1: Recovery, intra-day and inter-day precision for the individual organochlorines S4 Table S2: Spearmann rank correlation between the adipose tissue concentrations of individual PCB congeners and organochlorine pesticide analytes S5 Table S3: Incidence rate ratio for non-Hodgkin lymphoma and 95% confidence interval (CI) in association with adipose concentrations of p,p´-DDT, cis-nonachlor and oxychlordane within the early and late years of follow-up S6 Table S4: Sensitivity analysis, crude incidence rate ratio for non-Hodgkin lymphoma and 95 % confidence interval (CI) in association with adipose concentrations of DDT, cis-Nonachlor and oxychlordane including samples under LOD S7

List of Tables
S2

Extraction and gas chromatography-mass spectroscopy
The extracts were analyzed on a gas chromatography-mass spectroscopy instrument from (Agilent Technologies (Hewlett-Packard; Palo Alto, CA) model 6890/5973) using a DB-XLB capillary column. (Agilent Technologies; 60 m long, 0.25 mm inner diameter and 0.25 µm film thickness). The measurement of ions generated after negative chemical ionization was performed in selective ion mode. Peak areas were calculated relative to labeled internal standards (PCB 141-13 C 12 , PCB 153-13 C 12, PCB 180-13 C 12, hexachlorobenzene-13 C 6 , α-hexachlorohexane-13 C 6, p,p'-DDE-13 C 12 , oxychlordane-13 C 10, trans-nonachlor-13 C 10 and endrin-13 C 12 ). Samples (3 µl) were injected in the pulsed splitless mode. The temperature program was as follows: 2 min at 100 °C followed by an increase to 200 °C at a rate of 20 °C min −1 , increase to 245 °C at a rate of 1.5 °C min −1 hold 10 minutes, increase to 280 °C at a rate of 20 °C min -1 hold 5 minutes and finally an increase to 330 °C at a rate of 30 °C min -1 hold 15 minutes. The total run time was 70.42 minutes.
The linearity of the six-point curve was evaluated during the validation of the analytical method, but due to time consumption considerations and to maintain a high throughput production, the quantification was based on a single mid-point calibration.

Quality Assurance/Quality Control of PCB Analyses
Determination of analytical uncertainties found in adipose samples is crucial for the interpretation of data. In this study, each batch consisted of 16 samples, one calibration standard, one procedural blank and one sample of internal reference material. The internal reference material was cod liver oil containing all organochlorines analyzed, and was provided by the National usually by a combination of robust statistics and the use of data from expert laboratories. Where a determinant was present at low concentration and no assigned value could be set, the concentrations were given as indicative values. The z-scores were determined for each compound and when the zscore was less than 2, the determinant was considered satisfactory. If the z-score was between 2 and 3, the determinant was considered questionable and if it was greater than 3, the determinant was unsatisfactory. Results of the inter-comparison exercises mentioned for our laboratory were all deemed satisfactory (the best score).  * The cut-off points between the exposure groups were the 25 th , 50 th , 75 th and 90 th percentiles ** Age as time scale and model stratified according to gender. Participants with OC levels below LOD were assigned to the lowest exposure category if the LOD for the actual sample was below the 25th percentile (i.e., the cutoff point between the two lower exposure categories) because only these belonged for certain to the lowest exposure category).

S4
The same samples were included in the linear analyses. A separate dummy 'exposure' variable for the rest of the nondetects with LOD => 25 th percentile was included.
# Linear estimate per inter-quartile range based on pesticide concentration analyzed as a continuous variable in a linear model. A separate dummy 'exposure' variable for the rest of the non-detects with LOD => 25 th percentile was included.