Open access
Note
16 October 2023

Nested Case–Control Studies Investigating Serum Perfluorooctanoate and Perfluorooctane Sulfonate Levels and Pancreatic Ductal Adenocarcinoma in Two Cohorts

Publication: Environmental Health Perspectives
Volume 131, Issue 10
CID: 107702

Introduction

Per- and polyfluoroalkyl substances (PFAS), including perfluorooctanoate (PFOA) and perfluorooctane sulfonate (PFOS), are persistent organic pollutants that can be ingested from contaminated drinking water and food. PFAS have long half-lives1 and are potential carcinogens, with epidemiologic evidence strongest for PFOA and cancers of the kidney and testis.1 PFOA has been shown to induce pancreatic acinar cell tumors in male rats2; however, the epidemiologic evidence for an association of PFAS with pancreatic cancer (PC) is limited. We investigated associations between prediagnostic PFOA and PFOS serum levels and pancreatic ductal adenocarcinoma (PDAC, the most common PC type) in two independent nested case–control studies within the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC) and the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO).

Methods

We conducted two PC nested case–control studies within the ATBC (251 matched pairs), including male Finnish smokers 50–69 years of age at baseline (1985–1988), and within the PLCO (360 matched pairs), involving American men and women, mostly nonsmokers, 55–74 years of age at baseline (1993–2001). The cohorts have been previously described.3,4 The ATBC was approved by the institutional review boards (IRBs) of the U.S. National Cancer Institute (NCI) and the National Public Health Institute in Finland. The PLCO was approved by the IRBs of the NCI and 10 study centers. All participants provided informed consent. Overnight fasting serum samples in the ATBC and nonfasting serum samples in the PLCO were collected before randomization using study-specific protocols and stored at 70°C. Demographics, diabetes history, and lifestyle habits (e.g., smoking, dietary intake) were self-reported using questionnaires at enrollment in each cohort. Height and weight were measured by trained staff in the ATBC and were self-reported in the PLCO.3,4 Incident primary PDAC cases [International Classification of Diseases for Oncology, 3rd Edition5 (ICD-O-3): C25.0–C25.3, C25.7–C25.9] were ascertained via linkage to the Finnish Cancer Registry in the ATBC (through December 2011) and by annual mail-in surveys, cancer registries or the National Death Index in the PLCO (up to 15 May 2010). Control participants were incidence-density sampled and matched to case participants on age at blood draw (ATBC: ±5 y, PLCO: frequency-matched 5-y blocks), date of blood draw (ATBC: within 30 d; PLCO: frequency-matched 2-month blocks), and sex and race (PLCO only).
Once thawed, the serum samples were sent to Metabolon, Inc. (Durham, North Carolina) on dry ice for metabolite measures that included PFAS: 251 ATBC-matched pairs (2013/2014) and 360 PLCO-matched pairs (2017/2018). Samples from matched case–control sets were handled consistently and placed consecutively in each batch. Relative PFOA and PFOS levels within each population were measured using untargeted ultra-performance liquid chromatography–tandem mass spectrometry or gas chromatography mass spectrometry using either the Orbitrap Elite or Orbitrap Q-Exactive platforms.6 Peak intensity of a metabolite was normalized by dividing values by the median on each run-day. Detailed laboratory analysis has been described elsewhere.6 No sample had metabolite values below the limit of detection. Replicate quality controls (10%) were placed across all batches. The coefficients of variation for PFOA and PFOS were 15.1% and 8.1% in the ATBC and 7.9% and 11.4% in the PLCO.
We used conditional logistic regression models, which inherently adjust for the matching factors to calculate cohort-specific odds ratios (ORs) and 95% confidence intervals (CIs) per standard deviation (SD) increase of log10-transformed PFOA or PFOS levels and PDAC and analyte categories defined using cohort-specific quintiles as cut points based on the distribution of the controls. Models were additionally adjusted for age at blood draw (continuous), smoking (ATBC: years smoked and cigarettes per day; PLCO: never smoker, former smoker quit 15 y, former smoker quit <15 y, current smoker, or missing), diabetes (no, yes, or missing), and body mass index (<25.0kg/m2, 25.030.0kg/m2, 30.0kg/m2, or missing) because these are putative risk factors for PDAC and potential confounders. We used cohort-specific smoking variables given the participant characteristics and how smoking history was queried. To address potential reverse causation and sex differences, sensitivity analyses were conducted stratified by follow-up time (<5, 5–10, or 10 y) and sex (PLCO only) and tested for heterogeneity across strata by comparing a Wald test statistic to a χ2 distribution with 2 or 1 degree of freedom, respectively. All analyses were conducted using R (version 4.2.3; R Development Core Team). All statistical tests were two-tailed.

Results

In both studies, case participants smoked more pack-years and were more likely to report diabetes than control participants (Table 1); neither risk factors were associated with serum PFAS, except for lower PFOA in participants with diabetes in the PLCO (p=0.012). The ATBC participants were younger at blood draw (median=56 vs. 65 years of age), had longer follow-up time [median (range): 12 (0–24) vs. 9 (0–18) y], and consumed more alcohol and fish than those in the PLCO.
Table 1 Baseline characteristics of nested case–control study participants from the Alpha-Tocopherol, Beta-Carotene Study (ATBC, 1985–1988) and Prostate, Lung, Colorectal and Ovarian Cancer Screen Trial (PLCO, 1993–2001) cohorts.
CharacteristicsATBCPLCO
Cases (n=251)Controls (n=251)p-ValueaCases (n=360)Controls (n=360)p-Valuea
Age (y)      
 Blood draw56 (53, 60)56 (53, 60)1.0065 (61, 69)65 (61, 68)0.78
 Diagnosis70 (54, 87)  73 (57, 89)  
Time to diagnosis (y)12 (0, 24)  9 (0, 18)  
Sex    1.00
 Males251 (100)251 (100) 211 (58.6)211 (58.6) 
 Females0 (0)0 (0) 149 (41.4)149 (41.4) 
Racial-ethnic group    1.00
 Non-Hispanic White251 (100)251 (100) 320 (88.9)320 (88.9) 
 Non-Hispanic Black0 (0)0 (0) 14 (3.9)14 (3.9) 
 Other0 (0)0 (0) 26 (7.2)26 (7.2) 
Smoking    <0.01
 Never smoker0 (0)0 (0) 146 (40.6)191 (53.1) 
 Former smoker quit 15 y0 (0)0 (0) 85 (23.6)100 (27.8) 
 Former smoker quit <15 y0 (0)0 (0) 62 (17.2)40 (11.1) 
 Current smoker251 (100)251 (100) 65 (18.1)23 (6.4) 
 Missing0 (0)0 (0) 2 (0.6)6 (1.7) 
Years smoked36 (31, 42)36 (30, 41)0.75
Cigarettes per day20 (15, 25)20 (15, 25)0.04
Smoking pack-years38 (27, 46)35 (23, 43)0.0314 (0, 43)0 (0, 23)<0.01
Diabetes  0.06  0.03
 No234 (93.2)244 (97.2) 315 (87.5)328 (91.1) 
 Yes17 (6.8)7 (2.8) 45 (12.5)29 (8.1) 
 Missing0 (0)0 (0) 0 (0)3 (0.8) 
BMI (kg/m2)b26.2 (24.0, 28.7)26.0 (23.5, 28.1)0.2326.6 (24.0, 29.8)26.7 (24.0, 29.2)0.76
BMI category (kg/m2)  0.24  0.72
<25.087 (34.7)100 (39.8) 122 (33.9)120 (33.3) 
25.0 to <30.0115 (45.8)115 (45.8) 150 (41.7)163 (45.3) 
30.049 (19.5)36 (14.3) 84 (23.3)74 (20.6) 
 Missing0 (0)0 (0) 4 (1.1)3 (0.8) 
Alcohol intake (g/d)b11.3 (3.1, 30.5)10.1 (2.2, 24.5)0.171.5 (0.3, 11.8)1.6 (0.3, 10.1)0.67
Fish intake [g/(d ×1,000 kcal)]b12.8 (8.0, 19.2)12.1 (7.1, 18.3)0.426.1 (2.7, 11.7)5.9 (2.2, 12.0)0.70
Note: Values are median (full range) for age at diagnosis and time to diagnosis, median (interquartile range) for other continuous variables, and number (percentage) for categorical variables. All the characteristics except age at diagnosis and time to diagnosis were collected at baseline in each cohort. —, not applicable; BMI, body mass index.
a
p-Values for differences for continuous and categorical variables were derived from the Wilcoxon’s rank-sum test (excluding missing data) and the chi-square test (including missing data as a missing category), respectively.
b
Missing values of alcohol and fish intake: 11 cases and 8 controls in the ATBC and 33 cases and 16 controls in the PLCO.
Serum PFOA level was positively associated with PDAC (OR=1.27; 95% CI: 1.04, 1.56 per SD increase) in the ATBC (Table 2). In contrast, associations between PFOA and PDAC in the PLCO were consistently null (OR=0.97; 95% CI: 0.82, 1.15). PFOS was not associated with PDAC in either cohort. There was no heterogeneity by follow-up time overall or, in the PLCO, by sex.
Table 2 Odds ratios (95% confidence intervals) for the associations of relative serum PFOA and PFOS levels with risk of PDAC in the Alpha-Tocopherol, Beta-Carotene Study (ATBC) and the Prostate, Lung, Colorectal and Ovarian Cancer Screen Trial (PLCO) cohorts nested case–control studies.
MetaboliteATBC (251 case–control pairs)PLCO (360 case–control pairs)
Ncases/controlsOR (95% CI)aNcases/controlsOR (95% CI)a
PFOA
 Quintileb
  130/511.00 (Ref)62/731.00 (Ref)
  255/501.94 (1.05, 3.59)78/711.26 (0.78, 2.04)
  341/501.45 (0.77, 2.72)81/721.43 (0.88, 2.31)
  463/502.27 (1.19, 4.33)78/711.30 (0.79, 2.13)
  562/502.37 (1.24, 4.51)61/730.95 (0.57, 1.59)
  ptrendc 0.01 0.87
  Per SD increased251/2511.27 (1.04, 1.56)360/3600.97 (0.82, 1.15)
 Stratified by follow-up time
  <5 y33/331.35 (0.76, 2.41)110/1101.10 (0.82, 1.48)
  5 to <10 y56/561.09 (0.71, 1.68)139/1390.89 (0.68, 1.17)
  10 y162/1621.35 (1.03, 1.76)111/1110.97 (0.68, 1.37)
  pheterogeneitye 0.70 0.59
 Stratified by sex
  Male251/2511.27 (1.04, 1.56)211/2110.91 (0.73, 1.14)
  Female149/1491.09 (0.84, 1.40)
  pheterogeneitye  0.31
PFOS
 Quintileb
  122/261.00 (Ref)80/731.00 (Ref)
  231/261.57 (0.69, 3.57)65/720.86 (0.51, 1.44)
  318/260.77 (0.32, 1.86)72/711.03 (0.63, 1.70)
  423/260.89 (0.38, 2.11)75/711.05 (0.65, 1.70)
  536/261.82 (0.82, 4.03)68/730.88 (0.53, 1.48)
  ptrendc 0.34 0.88
  Per SD increased130/1301.13 (0.88, 1.45)360/3600.97 (0.83, 1.14)
 Stratified by follow-up time
  <5 y21/213.09 (0.63, 15.19)110/1100.95 (0.72, 1.26)
  5 to <10 y29/291.11 (0.58, 2.13)139/1390.94 (0.71, 1.25)
  10 y80/801.03 (0.76, 1.40)111/1110.97 (0.71, 1.33)
  pheterogeneitye 0.42 0.99
 Stratified by sex
  Male130/1301.13 (0.88, 1.45)211/2110.88 (0.71, 1.10)
  Female149/1491.12 (0.87, 1.43)
  pheterogeneitye  0.16
Note: —, not applicable; CI, confidence interval; OR, odds ratio; PDAC, pancreatic ductal adenocarcinoma; PFOA, perfluorooctanoate; PFOS, perfluorooctane sulfonate; Ref, reference; SD, standard deviation.
a
OR and 95% CI calculated using conditional logistic regression models adjusted for age at blood draw (continuous), smoking (ATBC: years smoked and cigarettes per day; PLCO: never smoker, former smoker quit 15 y, former smoker quit <15 y, current smoker, or missing), diabetes (no, yes, or missing), and body mass index (<25.0kg/m2, 25.030.0kg/m2, 30.0kg/m2, or missing) in each cohort. None of the samples had metabolite values below the limit of detection.
b
Quintiles of the log10 relative metabolite levels were defined using cohort-specific quintile cut points among the controls in 2013 and 2014 respectively in the ATBC, and 2017 and 2018 respectively in the PLCO.
c
Trend across quintiles was tested by multivariable adjusted conditional logistic regression models using quintile-specific median values of the metabolite among the controls as a continuous independent variable.
d
Standard deviations of the log10 metabolite levels were 0.19 in the ATBC and 0.24 in the PLCO for PFOA, and 0.23 in both the ATBC and the PLCO for PFOS.
e
p-Values for heterogeneity of associations across strata by follow-up time and sex were obtained by comparing a Wald test statistic to a χ2 distribution with 2 or 1 degree of freedom, respectively.

Discussion

We observed a positive association between serum PFOA level and PDAC in the ATBC, particularly among those diagnosed 10 y after blood collection, whereas the PLCO findings were null. PFOS was not associated with PDAC in either population. Weak, statistically nonsignificant associations between PFOA and PC were previously observed in a case–cohort study within a Danish cohort and an occupational cohort in Minnesota, whereas null findings were observed in two studies within the Mid-Ohio Valley involving exposures to PFOA in contaminated drinking water.710 It is unclear why the results from the ATBC and the PLCO differ; possible reasons may involve factors related to the younger age at serum collection, diet, residual or confounding from unmeasured factors among the ATBC participants, or they may be due to chance.
Strengths of our study include the relatively large number of cases in two populations and prediagnostic PFAS levels. A limitation is the relative quantification of PFOA and PFOS in each population, precluding direct comparisons of measured concentrations between these and other populations. Relative serum PFAS measurements from Metabolon have shown strong correlations with targeted absolute concentration measurements from the Centers for Disease Control and Prevention laboratory (Spearman correlation coefficient=0.76 for PFOA and PFOS),11 supporting the validity of relative PFAS measurements for epidemiologic analyses. Although absolute PFAS concentrations have not been measured in the ATBC, previous PLCO studies have reported that serum PFOA and PFOS concentrations in the control participants were comparable to those observed among the general population in the U.S. National Health and Nutrition Examination Survey during 1999–2000.12 Future studies using targeted analyses are needed for more direct comparisons. Long-term sample storage may affect measurement, although the samples from matched pairs had similar storage time and processing. Both studies were limited to PFAS levels at a single time point; evidence from the PLCO involving serially collected samples suggests that one-time measurements are informative as surrogates of long-term levels.12 Finally, our findings might not be generalizable to younger people, more diverse populations, or populations in regions other than Finland or the United States.
In conclusion, we observed a positive association between serum PFOA level and PDAC within the ATBC study of male Finnish smokers, whereas our PLCO findings in American men and women, mostly nonsmokers, were null. Additional studies using absolute PFAS quantification are needed to confirm this association before considering public health recommendations in relation to PDAC risk.

Acknowledgments

This work was funded by the Intramural Research Program, Division of Cancer Epidemiology and Genetics of the U.S. National Cancer Institute (NCI), National Institutes of Health and by the U.S. Army Medical Research Acquisition Activity through the Peer Review Cancer Research Program Discovery Award under award W81XWH-12-1-0369 (R.Z.S.).
Ethical restrictions on human subjects’ data prevent our posting the data used for this analysis. Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) data can be shared via a request and approval at https://cdas.cancer.gov/learn/plco/instructions/ and after appropriate data transfer agreement. A small portion of the PLCO cancer data collected via certain state cancer registries have not permitted individual data sharing outside the NCI. The Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC) data are maintained by the NCI, Division of Cancer Epidemiology and Genetics, and are available to bona fide researchers upon submission and approval of a research proposal, and subsequent completion of a data transfer agreement. Proposals can be submitted at https://atbcstudy.cancer.gov/ptsa/. The software code will be made available upon request to the corresponding authors.

Article Notes

*
These authors are joint last senior authors.
The authors declare they have nothing to disclose.

References

1.
International Agency for Research on Cancer. 2017. Perfluorooctanoic Acid (PFOA). https://monographs.iarc.fr/wp-content/uploads/2018/06/mono110-01.pdf [accessed 17 March 2023].
2.
Biegel LB, Hurtt ME, Frame SR, O’Connor JC, Cook JC. 2001. Mechanisms of extrahepatic tumor induction by peroxisome proliferators in male CD rats. Toxicol Sci 60(1):44–55. https://www.ncbi.nlm.nih.gov/pubmed/11222872, https://doi.org/10.1093/toxsci/60.1.44.
3.
ATBC Cancer Prevention Study Group. 1994. The Alpha-Tocopherol, Beta-Carotene Lung Cancer Prevention Study: design, methods, participant characteristics, and compliance. Ann Epidemiol 4(1):1–10. https://www.ncbi.nlm.nih.gov/pubmed/8205268, https://doi.org/10.1016/1047-2797(94)90036-1.
4.
Prorok PC, Andriole GL, Bresalier RS, Buys SS, Chia D, Crawford ED, et al. 2000. Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial. Control Clin Trials 21(6 suppl 1):273S–309S. https://www.ncbi.nlm.nih.gov/pubmed/11189684, https://doi.org/10.1016/s0197-2456(00)00098-2.
5.
Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, et al. 2013. International Classification of Diseases for Oncology, 3rd Edition. https://iris.who.int/bitstream/handle/10665/96612/9789241548496_eng.pdf [accessed 1 March 2023].
6.
Stolzenberg-Solomon R, Derkach A, Moore S, Weinstein SJ, Albanes D, Sampson J. 2020. Associations between metabolites and pancreatic cancer risk in a large prospective epidemiological study. Gut 69(11):2008–2015. https://www.ncbi.nlm.nih.gov/pubmed/32060129, https://doi.org/10.1136/gutjnl-2019-319811.
7.
Eriksen KT, Sørensen M, McLaughlin JK, Lipworth L, Tjønneland A, Overvad K, et al. 2009. Perfluorooctanoate and perfluorooctanesulfonate plasma levels and risk of cancer in the general Danish population. J Natl Cancer Inst 101(8):605–609. https://www.ncbi.nlm.nih.gov/pubmed/19351918, https://doi.org/10.1093/jnci/djp041.
8.
Barry V, Winquist A, Steenland K. 2013. Perfluorooctanoic acid (PFOA) exposures and incident cancers among adults living near a chemical plant. Environ Health Perspect 121(11–12):1313–1318. https://www.ncbi.nlm.nih.gov/pubmed/24007715, https://doi.org/10.1289/ehp.1306615.
9.
Raleigh KK, Alexander BH, Olsen GW, Ramachandran G, Morey SZ, Church TR, et al. 2014. Mortality and cancer incidence in ammonium perfluorooctanoate production workers. Occup Environ Med 71(7):500–506. https://www.ncbi.nlm.nih.gov/pubmed/24832944, https://doi.org/10.1136/oemed-2014-102109.
10.
Vieira VM, Hoffman K, Shin HM, Weinberg JM, Webster TF, Fletcher T. 2013. Perfluorooctanoic acid exposure and cancer outcomes in a contaminated community: a geographic analysis. Environ Health Perspect 121(3):318–323. https://www.ncbi.nlm.nih.gov/pubmed/23308854, https://doi.org/10.1289/ehp.1205829.
11.
Rhee J, Loftfield E, Albanes D, Layne TM, Stolzenberg-Solomon R, Liao LM, et al. 2023. A metabolomic investigation of serum perfluorooctane sulfonate and perfluorooctanoate. Environ Int 180:108198. https://www.ncbi.nlm.nih.gov/pubmed/37716341, https://doi.org/10.1016/j.envint.2023.108198.
12.
Rhee J, Barry KH, Huang WY, Sampson JN, Hofmann JN, Silverman DT, et al. 2023. A prospective nested case-control study of serum concentrations of per- and polyfluoroalkyl substances and aggressive prostate cancer risk. Environ Res 228:115718. https://www.ncbi.nlm.nih.gov/pubmed/36958379, https://doi.org/10.1016/j.envres.2023.115718.

Information & Authors

Information

Published In

Environmental Health Perspectives
Volume 131Issue 10October 2023
PubMed: 37844029

History

Received: 21 April 2023
Revision received: 15 August 2023
Accepted: 29 September 2023
Published online: 16 October 2023

Authors

Affiliations

Ting Zhang
Metabolic Epidemiology Branch, Division of Cancer Epidemiology and Genetics (DCEG), National Cancer Institute (NCI), National Institutes of Health (NIH), Department of Health and Human Services (DHHS), Rockville, Maryland, USA
Sheng Fu
Biostatistics Branch, DCEG, NCI, NIH, DHHS, Rockville, Maryland, USA
Kai Yu
Biostatistics Branch, DCEG, NCI, NIH, DHHS, Rockville, Maryland, USA
Demetrius Albanes
Metabolic Epidemiology Branch, Division of Cancer Epidemiology and Genetics (DCEG), National Cancer Institute (NCI), National Institutes of Health (NIH), Department of Health and Human Services (DHHS), Rockville, Maryland, USA
Steven C. Moore
Metabolic Epidemiology Branch, Division of Cancer Epidemiology and Genetics (DCEG), National Cancer Institute (NCI), National Institutes of Health (NIH), Department of Health and Human Services (DHHS), Rockville, Maryland, USA
Mark P. Purdue*
Occupational and Environmental Epidemiology Branch, DCEG, NCI, NIH, DHHS, Rockville, Maryland, USA
Rachael Z. Stolzenberg-Solomon*
Metabolic Epidemiology Branch, Division of Cancer Epidemiology and Genetics (DCEG), National Cancer Institute (NCI), National Institutes of Health (NIH), Department of Health and Human Services (DHHS), Rockville, Maryland, USA

Notes

Address correspondence to Rachael Z. Stolzenberg-Solomon, Metabolic Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute. Email: [email protected]

Metrics & Citations

Metrics

About Article Metrics


Citations

Download citation

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click DOWNLOAD.

View Options

View options

PDF

View PDF

Get Access

Restore your content access

Enter your email address to restore your content access:

Note: This functionality works only for purchases done as a guest. If you already have an account, log in to access the content to which you are entitled.

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media