- Split View
-
Views
-
Cite
Cite
Thomas P Ahern, Logan G Spector, Per Damkier, Buket Öztürk Esen, Sinna P Ulrichsen, Katrine Eriksen, Timothy L Lash, Henrik Toft Sørensen, Deirdre P Cronin-Fenton, Medication–Associated Phthalate Exposure and Childhood Cancer Incidence, JNCI: Journal of the National Cancer Institute, Volume 114, Issue 6, June 2022, Pages 885–894, https://doi.org/10.1093/jnci/djac045
- Share Icon Share
Abstract
Human phthalate exposure is widespread through contact with myriad consumer products. Exposure is particularly high through medications formulated with phthalates. Phthalates disrupt normal endocrine signaling and are associated with reproductive outcomes and incidence of some cancers. We measured associations between gestational and childhood medication-associated phthalate exposures and the incidence of childhood cancers.
We identified all live births in Denmark between 1997 and 2017, including both children and birth mothers. Using drug ingredient data merged with the Danish National Prescription Registry, we measured phthalate exposure through filled prescriptions for mothers during pregnancy (gestational exposure) and for children from birth until age 19 years (childhood exposure). Incident childhood cancers were ascertained from the Danish Cancer Registry, and associations were estimated with Cox regression models.
Among 1 278 685 children, there were 2027 childhood cancer cases diagnosed over 13.1 million person–years of follow-up. Childhood phthalate exposure was strongly associated with incidence of osteosarcoma (hazard ratio [HR] = 2.78, 95% confidence interval [CI] = 1.63 to 4.75). We also observed a positive association with incidence of lymphoma (HR = 2.07, 95% CI = 1.36 to 3.14), driven by associations with Hodgkin and non-Hodgkin lymphoma but not Burkitt lymphoma. Associations were apparent only for exposure to low-molecular phthalates, which have purportedly greater biological activity.
Childhood phthalate exposure was associated with incidence of osteosarcoma and lymphoma before age 19 years. Lingering questions include which specific phthalate(s) are responsible for these associations, by what mechanisms they occur, and to what extent childhood cancer cases could be avoided by reducing or eliminating the phthalate content of medications and other consumer products.
Nearly 16 000 cases of cancer are expected to be diagnosed among US children and adolescents in 2021 (1), and an estimated 11.5 million disability-adjusted life-years were lost to childhood cancer globally in 2017 (2). Although incidence of childhood malignancy has slowly increased in recent decades, few exogenous causes have been identified for the majority of pediatric cancers (3,4). It is especially unclear to what extent maternal and childhood environmental exposures modify risk and, therefore, to what extent childhood cancers could be preventable through these factors. A survey by the Children’s Cancer Group found that the potential etiologic role of environmental exposures in a child’s disease was the most prominent concern expressed by parents of childhood cancer patients (5).
Phthalates are plastic additives that increase flexibility, transparency, durability, and longevity (6). They are found in children’s toys, medical tubing, food storage containers, building materials, furniture, and clothing (6). Because they are not covalently bound with other constituents, phthalates readily leach out of products (7). Human exposure is widespread through ingestion, inhalation, and absorption of leached phthalate residues. The US National Health and Nutrition Examination Survey found that more than 75% of US adults had detectable levels of at least 1 urinary phthalate metabolite (8). Phthalate metabolites are also detectable in human amniotic fluid (9). Phthalates mimic hormones and interfere with normal endocrine signaling pathways (10) and may therefore affect hormonally mediated health outcomes including fertility, fetal and child development, and cancer (6,11).
Phthalates are used in the formulation of many pharmaceuticals, particularly those requiring an enteric coating to control drug release (12). Users of phthalate-containing medications represent a highly exposed population, having metabolite burdens up to 50-fold higher than people with only environmental phthalate exposure (13).
Studying phthalate health effects is difficult for several reasons: 1) the short half-life of phthalate metabolites requires longitudinal measurements to characterize exposure history (6); 2) phthalate metabolite assays are prohibitively expensive at the scale of adequately sized epidemiologic studies; and 3) accurate measurement of phthalate metabolites requires urine specimens (14), which are rarely collected in epidemiologic cohorts. To overcome these obstacles, we capitalized on the high relative exposure among users of phthalate-containing medications to examine associations between gestational and childhood phthalate exposure and childhood cancer incidence in a Danish nationwide cohort.
Methods
Ethics Approval
This study was approved by the institutional review board at the University of Vermont and by the Danish Children’s Cancer Group. This study was also registered with the Danish Data Protection Board and adhered to the European Union’s General Data Protection Regulations.
Danish Population-Based Registries
Denmark has a tax-supported universal health-care system with free access to primary care, inpatient and outpatient clinics, and emergency care (15). As part of this system, Denmark maintains a network of population-based registries to record information on health, employment, vital status, and emigration (15). We enumerated our study population from several Danish registries with linkage via the unique Central Personal Register code assigned to all legal residents (16).
Phthalates in Prescription Medications
Since 1995, the Danish Medicines Agency has maintained a database of all pharmaceutical products marketed in Denmark. Records include the following data fields: the VNR code (a unique identifier assigned to medicinal agents distinguished by manufacturer, active ingredient, formulation, dose, and package size), the Anatomical Therapeutic Chemical code, dates of market entry and removal, and a listing of active and inactive ingredients. Ingredient data include the mass of each constituent per unit of medication (eg, milligrams per capsule). We searched drug ingredient fields for Danish and English text strings encompassing a comprehensive set of phthalate compounds. We then prepared a dataset of oral medications formulated with at least 1 phthalate (17,18) and merged this with the Danish National Prescription Registry (DNPR) (19). The DNPR records information about every community pharmacy transaction in Denmark since 1995, including transaction date, the Anatomical Therapeutic Chemical and VNR codes of the prescribed drug, and the quantity dispensed.
Study Population and Data Collection
Our source population included all live births in Denmark between 1997 and 2017, ascertained from the Danish Medical Birth Registry (20), which records data from all live births in Denmark since 1973, including home deliveries. Each birth record contains the delivery date, mode of delivery (vaginal vs cesarean section), type of pregnancy (singleton vs multiple), mother and infant Central Personal Register numbers, maternal data (eg, age, height, weight, smoking status, and parity), and infant data (eg, gestational age, birth weight, and presence of congenital malformations). We included both mothers and infants in the live birth cohort and ascertained their exposure to phthalate-containing medications by linkage with the ingredient-augmented DNPR. We identified cases of incident childhood cancer through age 19 years in the Danish Cancer Registry (International Classification of Diseases–10 codes appear in Supplementary Table 1, available online) (21). We retrieved vital status from the Danish Civil Registration System (16).
Definitions of Analytic Variables
We defined separate phthalate exposure groups for gestational and childhood time periods. Gestational exposure was characterized by phthalate-containing drug formulations dispensed to mothers. Childhood exposure was characterized by phthalate-containing drug formulations dispensed to children. Gestational exposures comprised the 9-month window preceding delivery dates and were initially divided into estimated trimesters. Childhood phthalate exposure was initially defined as the cumulative milligrams of phthalates ingested via medications. Data sparsity required simplifying these exposures as dichotomous variables. Phthalate exposures were defined both overall (exposure to any phthalate) and by specific compounds (ie, dibutyl phthalate [DBP], diethyl phthalate [DEP], cellulose acetate phthalate [CAP], polyvinyl acetate phthalate [PVAP], and hydroxypropyl methylcellulose phthalate [HPMCP]). We also defined exposure to phthalate polymers (CAP, HPMCP, and PVAP, considered biologically inactive) and to low molecular weight phthalates (DBP and DEP, suspected endocrine disruptors). Phthalate exposure was considered positive if at least 1 prescription was filled with a phthalate-containing drug product. We conducted a sensitivity analysis in which exposure required at least 2 phthalate-containing drug fills.
Follow-up began at birth and continued until attainment of age 19 years, first cancer diagnosis, death from any cause, emigration from Denmark, or January 1, 2018, whichever came first. We studied incidence of any childhood cancer as well as incidence of 16 specific cancer types or groups (leukemia [overall, acute lymphocytic, and acute myeloid], lymphoma [overall, Hodgkin, non-Hodgkin, and Burkitt], brain or central nervous system, neuroblastoma, retinoblastoma, kidney, liver, osteosarcoma, soft tissue sarcoma, reproductive organ cancers, and other or unclassified cancers); see Supplemental Table 1 (available online) for International Classification of Diseases–10 codes.
We defined child and maternal covariables with known or suspected associations with either childhood cancer or medication-associated phthalate exposure. Child-level covariables included natal sex (22), birth year, gestational age (23), birth weight (24,25), small for gestational age, and presence of congenital malformations or chromosomal abnormalities (26). Maternal-level covariables included age at delivery (27-29), prepregnancy body mass index (available since 2004) (30), mode of delivery (31-33), type of pregnancy (34), parity (35), and smoking status during pregnancy (36). Detailed definitions of categorical variables appear in Table 1.
Characteristics . | Any phthalate exposure No. (%) (n = 113 529) . | No phthalate exposure No. (%) (n = 1 165 156) . |
---|---|---|
Child characteristics | ||
Natal sex | ||
Female | 53 841 (47) | 568 638 (49) |
Male | 59 688 (53) | 596 518 (51) |
Birth year | ||
1997-2002 | 36 694 (32) | 342 801 (29) |
2003-2007 | 30 659 (27 | 284 379 (24) |
2008-2012 | 32 500 (29) | 266 379 (23) |
2013-2017 | 13 676 (12) | 271 597 (23) |
Gestational age, wk | ||
<34 | 2420 (2.1) | 23 607 (2.0) |
34-36 | 5889 (5.2) | 53 824 (4.6) |
37-41 | 98 974 (87) | 1 024 467(88) |
≥42 | 6246 (5.5) | 63 258 (5.4) |
Birth weight category, g | ||
Extremely low (<1000) | 318 (0.3) | 4186 (0.4) |
Very low (1000-1499) | 672 (0.6) | 6 173 (0.5) |
Low (1500-2499) | 5337 (4.7) | 49 629 (4.3) |
Normal (2500-3999) | 86 520 (76.5) | 889 878 (76.7) |
High (≥4000) | 20 190 (17.9) | 209 804 (18.1) |
(Missing) | 492 | 5486 |
Small for gestational age | ||
Yes | 11 317 (10.0) | 114 720 (9.9) |
No | 101 720 (90.0) | 1 044 950 (90.1) |
(Missing) | 492 | 5486 |
Congenital malformations | ||
Nervous system | 228 (0.2) | 1587 (0.1) |
Maxillofacial | 282 (0.2) | 2215 (0.2) |
Circulatory system | 1543 (1.4) | 12 306 (1.1) |
Respiratory system | 305 (0.3) | 2434 (0.2) |
Cleft lip/palate | 300 (0.3) | 2014 (0.2) |
Digestive system | 839 (0.7) | 7843 (0.7) |
Reproductive organs | 687 (0.6) | 6199 (0.5) |
Urinary tract | 366 (0.3) | 3513 (0.3) |
Musculoskeletal | 2385 (2.1) | 21 554 (1.8) |
Chromosomal abnormalities | 272 (0.2) | 1289 (0.1) |
Medication exposures | ||
Bisacodyl | 723 (0.6) | 1720 (0.1) |
Budesonide | 67 (0.1) | 399 (0.03) |
Clarithromycin | 81 732 (72.0) | 23 248 (2.0) |
Colestipol | <5 | <5 |
Diclofenac | 2088 (1.8) | 12 014 (1.0) |
Dipyridamole | <5 | <5 |
Duloxetine | 190 (0.2) | 461 (0.04) |
Erythromycin | 36 900 (32.5) | 143 846 (12.3) |
Esomeprazole | 2091 (1.8) | 12 491 (1.1) |
Fluoxetine | 449 (0.4) | 2840 (0.2) |
Ibuprofen | 11 557 (10.2) | 81 271 (7.0) |
Lithium | 32 (0.03) | 80 (0.01) |
Mesalazine | 166 (0.1) | 920 (0.1) |
Mianserin | 22 (0.02) | 148 (0.01) |
Multienzymes | 189 (0.2) | 96 (0.01) |
Mycophenolic acid | 7 (<0.01) | 47 (<0.01) |
Naproxen | 8742 (7.7) | 42 361 (3.6) |
Pentoxyverine | 76 (0.1) | 565 (0.1) |
Propantheline | 127 (0.1) | 863 (0.1) |
Rabeprazole | 14 (0.01) | 8 (<0.01) |
Sulfasalazine | 101 (0.1) | 93 (<0.01) |
Theophylline | 76 (0.1) | 132 (0.01) |
Valproic acid | 1046 (0.9) | 4029 (0.3) |
Verapamil | 17 (0.01) | 82 (<0.01) |
Maternal characteristics | ||
Age at delivery, y | ||
<20 | 1661 (1.5) | 15 595 (1.3) |
20-24 | 13 394 (12) | 131 073 (11) |
25-29 | 37 381 (33) | 388 657 (33) |
30-34 | 40 248 (35) | 414 131 (36) |
35-39 | 17 735 (16) | 182 337 (16) |
≥40 | 3110 (2.7) | 33 363 (2.9) |
Prepregnancy BMI, kg/m2 | ||
Underweight (<18.5) | 2967 (4.4) | 32 837 (4.4) |
Normal (18.5-24.9) | 41 183 (60.7) | 458 351 (62.0) |
Overweight (25-29.9) | 14 839 (21.9) | 156 011 (21.1) |
Obese (≥30) | 8880 (13.1) | 92 511 (12.5) |
(Missing)a | 45 660 | 425 446 |
Mode of delivery | ||
Vaginal | 88 442 (77.9) | 930 350 (79.8) |
Cesarean section | 25 087 (22.1) | 234 806 (20.2) |
Type of pregnancy | ||
Singleton | 108 997 (96.0) | 1 117 295 (95.9) |
Twins | 4440 (3.9) | 46 830 (4.0) |
Triplets or greater | 92 (0.1) | 1031 (0.1) |
Parity | ||
1 | 47 622 (41.9) | 496 409 (42.6) |
2 or 3 | 60 331 (53.1) | 608 788 (52.2) |
≥4 | 5576 (4.9) | 59 959 (5.1) |
Smoking during pregnancy | ||
Nonsmoker | 84 636 (82.0) | 907 165 (84.6) |
Quit during pregnancy | 2803 (2.7) | 27 689 (2.6) |
Smoked during pregnancy | 15 772 (15.3) | 137 553 (12.8) |
(Missing) | 10 318 | 92 749 |
Medication exposures | ||
Bisacodyl | 2655 (2.3) | 18 408 (1.6) |
Budesonide | 950 (0.8) | 5786 (0.5) |
Clarithromycin | 21 260 (18.7) | 130 624 (11.2) |
Colestipol | 107 (0.1) | 717 (0.1) |
Diclofenac | 48 146 (42.4) | 422 126 (36.2) |
Dipyridamole | 243 (0.2) | 1221 (0.1) |
Duloxetine | 4527 (4.0) | 33 070 (2.8) |
Erythromycin | 44 897 (39.5) | 301 785 (25.9) |
Esomeprazole | 9299 (8.2) | 66 885 (5.7) |
Fluoxetine | 4291 (3.8) | 33 164 (2.8) |
Ibuprofen | 89 112 (78.5) | 834 784 (71.6) |
Lithium | 753 (0.7) | 5022 (0.4) |
Mesalazine | 2370 (2.1) | 14 032 (1.2) |
Mianserin | 3149 (2.8) | 24 382 (2.1) |
Multienzymes | 155 (0.1) | 938 (0.1) |
Mycophenolic acid | 20 (0.02) | 126 (0.01) |
Naproxen | 20 556 (18.1) | 176 297 (15.1) |
Pentoxyverine | 208 (0.2) | 1220 (0.1) |
Propantheline | 470 (0.4) | 3741 (0.3) |
Rabeprazole | 455 (0.4) | 3124 (0.3) |
Sulfasalazine | 2277 (2.0) | 11 017 (0.9) |
Theophylline | 566 (0.5) | 2624 (0.2) |
Valproic acid | 1233 (1.1) | 8765 (0.8) |
Verapamil | 870 (0.8) | 5247 (0.5) |
Characteristics . | Any phthalate exposure No. (%) (n = 113 529) . | No phthalate exposure No. (%) (n = 1 165 156) . |
---|---|---|
Child characteristics | ||
Natal sex | ||
Female | 53 841 (47) | 568 638 (49) |
Male | 59 688 (53) | 596 518 (51) |
Birth year | ||
1997-2002 | 36 694 (32) | 342 801 (29) |
2003-2007 | 30 659 (27 | 284 379 (24) |
2008-2012 | 32 500 (29) | 266 379 (23) |
2013-2017 | 13 676 (12) | 271 597 (23) |
Gestational age, wk | ||
<34 | 2420 (2.1) | 23 607 (2.0) |
34-36 | 5889 (5.2) | 53 824 (4.6) |
37-41 | 98 974 (87) | 1 024 467(88) |
≥42 | 6246 (5.5) | 63 258 (5.4) |
Birth weight category, g | ||
Extremely low (<1000) | 318 (0.3) | 4186 (0.4) |
Very low (1000-1499) | 672 (0.6) | 6 173 (0.5) |
Low (1500-2499) | 5337 (4.7) | 49 629 (4.3) |
Normal (2500-3999) | 86 520 (76.5) | 889 878 (76.7) |
High (≥4000) | 20 190 (17.9) | 209 804 (18.1) |
(Missing) | 492 | 5486 |
Small for gestational age | ||
Yes | 11 317 (10.0) | 114 720 (9.9) |
No | 101 720 (90.0) | 1 044 950 (90.1) |
(Missing) | 492 | 5486 |
Congenital malformations | ||
Nervous system | 228 (0.2) | 1587 (0.1) |
Maxillofacial | 282 (0.2) | 2215 (0.2) |
Circulatory system | 1543 (1.4) | 12 306 (1.1) |
Respiratory system | 305 (0.3) | 2434 (0.2) |
Cleft lip/palate | 300 (0.3) | 2014 (0.2) |
Digestive system | 839 (0.7) | 7843 (0.7) |
Reproductive organs | 687 (0.6) | 6199 (0.5) |
Urinary tract | 366 (0.3) | 3513 (0.3) |
Musculoskeletal | 2385 (2.1) | 21 554 (1.8) |
Chromosomal abnormalities | 272 (0.2) | 1289 (0.1) |
Medication exposures | ||
Bisacodyl | 723 (0.6) | 1720 (0.1) |
Budesonide | 67 (0.1) | 399 (0.03) |
Clarithromycin | 81 732 (72.0) | 23 248 (2.0) |
Colestipol | <5 | <5 |
Diclofenac | 2088 (1.8) | 12 014 (1.0) |
Dipyridamole | <5 | <5 |
Duloxetine | 190 (0.2) | 461 (0.04) |
Erythromycin | 36 900 (32.5) | 143 846 (12.3) |
Esomeprazole | 2091 (1.8) | 12 491 (1.1) |
Fluoxetine | 449 (0.4) | 2840 (0.2) |
Ibuprofen | 11 557 (10.2) | 81 271 (7.0) |
Lithium | 32 (0.03) | 80 (0.01) |
Mesalazine | 166 (0.1) | 920 (0.1) |
Mianserin | 22 (0.02) | 148 (0.01) |
Multienzymes | 189 (0.2) | 96 (0.01) |
Mycophenolic acid | 7 (<0.01) | 47 (<0.01) |
Naproxen | 8742 (7.7) | 42 361 (3.6) |
Pentoxyverine | 76 (0.1) | 565 (0.1) |
Propantheline | 127 (0.1) | 863 (0.1) |
Rabeprazole | 14 (0.01) | 8 (<0.01) |
Sulfasalazine | 101 (0.1) | 93 (<0.01) |
Theophylline | 76 (0.1) | 132 (0.01) |
Valproic acid | 1046 (0.9) | 4029 (0.3) |
Verapamil | 17 (0.01) | 82 (<0.01) |
Maternal characteristics | ||
Age at delivery, y | ||
<20 | 1661 (1.5) | 15 595 (1.3) |
20-24 | 13 394 (12) | 131 073 (11) |
25-29 | 37 381 (33) | 388 657 (33) |
30-34 | 40 248 (35) | 414 131 (36) |
35-39 | 17 735 (16) | 182 337 (16) |
≥40 | 3110 (2.7) | 33 363 (2.9) |
Prepregnancy BMI, kg/m2 | ||
Underweight (<18.5) | 2967 (4.4) | 32 837 (4.4) |
Normal (18.5-24.9) | 41 183 (60.7) | 458 351 (62.0) |
Overweight (25-29.9) | 14 839 (21.9) | 156 011 (21.1) |
Obese (≥30) | 8880 (13.1) | 92 511 (12.5) |
(Missing)a | 45 660 | 425 446 |
Mode of delivery | ||
Vaginal | 88 442 (77.9) | 930 350 (79.8) |
Cesarean section | 25 087 (22.1) | 234 806 (20.2) |
Type of pregnancy | ||
Singleton | 108 997 (96.0) | 1 117 295 (95.9) |
Twins | 4440 (3.9) | 46 830 (4.0) |
Triplets or greater | 92 (0.1) | 1031 (0.1) |
Parity | ||
1 | 47 622 (41.9) | 496 409 (42.6) |
2 or 3 | 60 331 (53.1) | 608 788 (52.2) |
≥4 | 5576 (4.9) | 59 959 (5.1) |
Smoking during pregnancy | ||
Nonsmoker | 84 636 (82.0) | 907 165 (84.6) |
Quit during pregnancy | 2803 (2.7) | 27 689 (2.6) |
Smoked during pregnancy | 15 772 (15.3) | 137 553 (12.8) |
(Missing) | 10 318 | 92 749 |
Medication exposures | ||
Bisacodyl | 2655 (2.3) | 18 408 (1.6) |
Budesonide | 950 (0.8) | 5786 (0.5) |
Clarithromycin | 21 260 (18.7) | 130 624 (11.2) |
Colestipol | 107 (0.1) | 717 (0.1) |
Diclofenac | 48 146 (42.4) | 422 126 (36.2) |
Dipyridamole | 243 (0.2) | 1221 (0.1) |
Duloxetine | 4527 (4.0) | 33 070 (2.8) |
Erythromycin | 44 897 (39.5) | 301 785 (25.9) |
Esomeprazole | 9299 (8.2) | 66 885 (5.7) |
Fluoxetine | 4291 (3.8) | 33 164 (2.8) |
Ibuprofen | 89 112 (78.5) | 834 784 (71.6) |
Lithium | 753 (0.7) | 5022 (0.4) |
Mesalazine | 2370 (2.1) | 14 032 (1.2) |
Mianserin | 3149 (2.8) | 24 382 (2.1) |
Multienzymes | 155 (0.1) | 938 (0.1) |
Mycophenolic acid | 20 (0.02) | 126 (0.01) |
Naproxen | 20 556 (18.1) | 176 297 (15.1) |
Pentoxyverine | 208 (0.2) | 1220 (0.1) |
Propantheline | 470 (0.4) | 3741 (0.3) |
Rabeprazole | 455 (0.4) | 3124 (0.3) |
Sulfasalazine | 2277 (2.0) | 11 017 (0.9) |
Theophylline | 566 (0.5) | 2624 (0.2) |
Valproic acid | 1233 (1.1) | 8765 (0.8) |
Verapamil | 870 (0.8) | 5247 (0.5) |
Body mass index (BMI) data were available only from 2004 onward.
Characteristics . | Any phthalate exposure No. (%) (n = 113 529) . | No phthalate exposure No. (%) (n = 1 165 156) . |
---|---|---|
Child characteristics | ||
Natal sex | ||
Female | 53 841 (47) | 568 638 (49) |
Male | 59 688 (53) | 596 518 (51) |
Birth year | ||
1997-2002 | 36 694 (32) | 342 801 (29) |
2003-2007 | 30 659 (27 | 284 379 (24) |
2008-2012 | 32 500 (29) | 266 379 (23) |
2013-2017 | 13 676 (12) | 271 597 (23) |
Gestational age, wk | ||
<34 | 2420 (2.1) | 23 607 (2.0) |
34-36 | 5889 (5.2) | 53 824 (4.6) |
37-41 | 98 974 (87) | 1 024 467(88) |
≥42 | 6246 (5.5) | 63 258 (5.4) |
Birth weight category, g | ||
Extremely low (<1000) | 318 (0.3) | 4186 (0.4) |
Very low (1000-1499) | 672 (0.6) | 6 173 (0.5) |
Low (1500-2499) | 5337 (4.7) | 49 629 (4.3) |
Normal (2500-3999) | 86 520 (76.5) | 889 878 (76.7) |
High (≥4000) | 20 190 (17.9) | 209 804 (18.1) |
(Missing) | 492 | 5486 |
Small for gestational age | ||
Yes | 11 317 (10.0) | 114 720 (9.9) |
No | 101 720 (90.0) | 1 044 950 (90.1) |
(Missing) | 492 | 5486 |
Congenital malformations | ||
Nervous system | 228 (0.2) | 1587 (0.1) |
Maxillofacial | 282 (0.2) | 2215 (0.2) |
Circulatory system | 1543 (1.4) | 12 306 (1.1) |
Respiratory system | 305 (0.3) | 2434 (0.2) |
Cleft lip/palate | 300 (0.3) | 2014 (0.2) |
Digestive system | 839 (0.7) | 7843 (0.7) |
Reproductive organs | 687 (0.6) | 6199 (0.5) |
Urinary tract | 366 (0.3) | 3513 (0.3) |
Musculoskeletal | 2385 (2.1) | 21 554 (1.8) |
Chromosomal abnormalities | 272 (0.2) | 1289 (0.1) |
Medication exposures | ||
Bisacodyl | 723 (0.6) | 1720 (0.1) |
Budesonide | 67 (0.1) | 399 (0.03) |
Clarithromycin | 81 732 (72.0) | 23 248 (2.0) |
Colestipol | <5 | <5 |
Diclofenac | 2088 (1.8) | 12 014 (1.0) |
Dipyridamole | <5 | <5 |
Duloxetine | 190 (0.2) | 461 (0.04) |
Erythromycin | 36 900 (32.5) | 143 846 (12.3) |
Esomeprazole | 2091 (1.8) | 12 491 (1.1) |
Fluoxetine | 449 (0.4) | 2840 (0.2) |
Ibuprofen | 11 557 (10.2) | 81 271 (7.0) |
Lithium | 32 (0.03) | 80 (0.01) |
Mesalazine | 166 (0.1) | 920 (0.1) |
Mianserin | 22 (0.02) | 148 (0.01) |
Multienzymes | 189 (0.2) | 96 (0.01) |
Mycophenolic acid | 7 (<0.01) | 47 (<0.01) |
Naproxen | 8742 (7.7) | 42 361 (3.6) |
Pentoxyverine | 76 (0.1) | 565 (0.1) |
Propantheline | 127 (0.1) | 863 (0.1) |
Rabeprazole | 14 (0.01) | 8 (<0.01) |
Sulfasalazine | 101 (0.1) | 93 (<0.01) |
Theophylline | 76 (0.1) | 132 (0.01) |
Valproic acid | 1046 (0.9) | 4029 (0.3) |
Verapamil | 17 (0.01) | 82 (<0.01) |
Maternal characteristics | ||
Age at delivery, y | ||
<20 | 1661 (1.5) | 15 595 (1.3) |
20-24 | 13 394 (12) | 131 073 (11) |
25-29 | 37 381 (33) | 388 657 (33) |
30-34 | 40 248 (35) | 414 131 (36) |
35-39 | 17 735 (16) | 182 337 (16) |
≥40 | 3110 (2.7) | 33 363 (2.9) |
Prepregnancy BMI, kg/m2 | ||
Underweight (<18.5) | 2967 (4.4) | 32 837 (4.4) |
Normal (18.5-24.9) | 41 183 (60.7) | 458 351 (62.0) |
Overweight (25-29.9) | 14 839 (21.9) | 156 011 (21.1) |
Obese (≥30) | 8880 (13.1) | 92 511 (12.5) |
(Missing)a | 45 660 | 425 446 |
Mode of delivery | ||
Vaginal | 88 442 (77.9) | 930 350 (79.8) |
Cesarean section | 25 087 (22.1) | 234 806 (20.2) |
Type of pregnancy | ||
Singleton | 108 997 (96.0) | 1 117 295 (95.9) |
Twins | 4440 (3.9) | 46 830 (4.0) |
Triplets or greater | 92 (0.1) | 1031 (0.1) |
Parity | ||
1 | 47 622 (41.9) | 496 409 (42.6) |
2 or 3 | 60 331 (53.1) | 608 788 (52.2) |
≥4 | 5576 (4.9) | 59 959 (5.1) |
Smoking during pregnancy | ||
Nonsmoker | 84 636 (82.0) | 907 165 (84.6) |
Quit during pregnancy | 2803 (2.7) | 27 689 (2.6) |
Smoked during pregnancy | 15 772 (15.3) | 137 553 (12.8) |
(Missing) | 10 318 | 92 749 |
Medication exposures | ||
Bisacodyl | 2655 (2.3) | 18 408 (1.6) |
Budesonide | 950 (0.8) | 5786 (0.5) |
Clarithromycin | 21 260 (18.7) | 130 624 (11.2) |
Colestipol | 107 (0.1) | 717 (0.1) |
Diclofenac | 48 146 (42.4) | 422 126 (36.2) |
Dipyridamole | 243 (0.2) | 1221 (0.1) |
Duloxetine | 4527 (4.0) | 33 070 (2.8) |
Erythromycin | 44 897 (39.5) | 301 785 (25.9) |
Esomeprazole | 9299 (8.2) | 66 885 (5.7) |
Fluoxetine | 4291 (3.8) | 33 164 (2.8) |
Ibuprofen | 89 112 (78.5) | 834 784 (71.6) |
Lithium | 753 (0.7) | 5022 (0.4) |
Mesalazine | 2370 (2.1) | 14 032 (1.2) |
Mianserin | 3149 (2.8) | 24 382 (2.1) |
Multienzymes | 155 (0.1) | 938 (0.1) |
Mycophenolic acid | 20 (0.02) | 126 (0.01) |
Naproxen | 20 556 (18.1) | 176 297 (15.1) |
Pentoxyverine | 208 (0.2) | 1220 (0.1) |
Propantheline | 470 (0.4) | 3741 (0.3) |
Rabeprazole | 455 (0.4) | 3124 (0.3) |
Sulfasalazine | 2277 (2.0) | 11 017 (0.9) |
Theophylline | 566 (0.5) | 2624 (0.2) |
Valproic acid | 1233 (1.1) | 8765 (0.8) |
Verapamil | 870 (0.8) | 5247 (0.5) |
Characteristics . | Any phthalate exposure No. (%) (n = 113 529) . | No phthalate exposure No. (%) (n = 1 165 156) . |
---|---|---|
Child characteristics | ||
Natal sex | ||
Female | 53 841 (47) | 568 638 (49) |
Male | 59 688 (53) | 596 518 (51) |
Birth year | ||
1997-2002 | 36 694 (32) | 342 801 (29) |
2003-2007 | 30 659 (27 | 284 379 (24) |
2008-2012 | 32 500 (29) | 266 379 (23) |
2013-2017 | 13 676 (12) | 271 597 (23) |
Gestational age, wk | ||
<34 | 2420 (2.1) | 23 607 (2.0) |
34-36 | 5889 (5.2) | 53 824 (4.6) |
37-41 | 98 974 (87) | 1 024 467(88) |
≥42 | 6246 (5.5) | 63 258 (5.4) |
Birth weight category, g | ||
Extremely low (<1000) | 318 (0.3) | 4186 (0.4) |
Very low (1000-1499) | 672 (0.6) | 6 173 (0.5) |
Low (1500-2499) | 5337 (4.7) | 49 629 (4.3) |
Normal (2500-3999) | 86 520 (76.5) | 889 878 (76.7) |
High (≥4000) | 20 190 (17.9) | 209 804 (18.1) |
(Missing) | 492 | 5486 |
Small for gestational age | ||
Yes | 11 317 (10.0) | 114 720 (9.9) |
No | 101 720 (90.0) | 1 044 950 (90.1) |
(Missing) | 492 | 5486 |
Congenital malformations | ||
Nervous system | 228 (0.2) | 1587 (0.1) |
Maxillofacial | 282 (0.2) | 2215 (0.2) |
Circulatory system | 1543 (1.4) | 12 306 (1.1) |
Respiratory system | 305 (0.3) | 2434 (0.2) |
Cleft lip/palate | 300 (0.3) | 2014 (0.2) |
Digestive system | 839 (0.7) | 7843 (0.7) |
Reproductive organs | 687 (0.6) | 6199 (0.5) |
Urinary tract | 366 (0.3) | 3513 (0.3) |
Musculoskeletal | 2385 (2.1) | 21 554 (1.8) |
Chromosomal abnormalities | 272 (0.2) | 1289 (0.1) |
Medication exposures | ||
Bisacodyl | 723 (0.6) | 1720 (0.1) |
Budesonide | 67 (0.1) | 399 (0.03) |
Clarithromycin | 81 732 (72.0) | 23 248 (2.0) |
Colestipol | <5 | <5 |
Diclofenac | 2088 (1.8) | 12 014 (1.0) |
Dipyridamole | <5 | <5 |
Duloxetine | 190 (0.2) | 461 (0.04) |
Erythromycin | 36 900 (32.5) | 143 846 (12.3) |
Esomeprazole | 2091 (1.8) | 12 491 (1.1) |
Fluoxetine | 449 (0.4) | 2840 (0.2) |
Ibuprofen | 11 557 (10.2) | 81 271 (7.0) |
Lithium | 32 (0.03) | 80 (0.01) |
Mesalazine | 166 (0.1) | 920 (0.1) |
Mianserin | 22 (0.02) | 148 (0.01) |
Multienzymes | 189 (0.2) | 96 (0.01) |
Mycophenolic acid | 7 (<0.01) | 47 (<0.01) |
Naproxen | 8742 (7.7) | 42 361 (3.6) |
Pentoxyverine | 76 (0.1) | 565 (0.1) |
Propantheline | 127 (0.1) | 863 (0.1) |
Rabeprazole | 14 (0.01) | 8 (<0.01) |
Sulfasalazine | 101 (0.1) | 93 (<0.01) |
Theophylline | 76 (0.1) | 132 (0.01) |
Valproic acid | 1046 (0.9) | 4029 (0.3) |
Verapamil | 17 (0.01) | 82 (<0.01) |
Maternal characteristics | ||
Age at delivery, y | ||
<20 | 1661 (1.5) | 15 595 (1.3) |
20-24 | 13 394 (12) | 131 073 (11) |
25-29 | 37 381 (33) | 388 657 (33) |
30-34 | 40 248 (35) | 414 131 (36) |
35-39 | 17 735 (16) | 182 337 (16) |
≥40 | 3110 (2.7) | 33 363 (2.9) |
Prepregnancy BMI, kg/m2 | ||
Underweight (<18.5) | 2967 (4.4) | 32 837 (4.4) |
Normal (18.5-24.9) | 41 183 (60.7) | 458 351 (62.0) |
Overweight (25-29.9) | 14 839 (21.9) | 156 011 (21.1) |
Obese (≥30) | 8880 (13.1) | 92 511 (12.5) |
(Missing)a | 45 660 | 425 446 |
Mode of delivery | ||
Vaginal | 88 442 (77.9) | 930 350 (79.8) |
Cesarean section | 25 087 (22.1) | 234 806 (20.2) |
Type of pregnancy | ||
Singleton | 108 997 (96.0) | 1 117 295 (95.9) |
Twins | 4440 (3.9) | 46 830 (4.0) |
Triplets or greater | 92 (0.1) | 1031 (0.1) |
Parity | ||
1 | 47 622 (41.9) | 496 409 (42.6) |
2 or 3 | 60 331 (53.1) | 608 788 (52.2) |
≥4 | 5576 (4.9) | 59 959 (5.1) |
Smoking during pregnancy | ||
Nonsmoker | 84 636 (82.0) | 907 165 (84.6) |
Quit during pregnancy | 2803 (2.7) | 27 689 (2.6) |
Smoked during pregnancy | 15 772 (15.3) | 137 553 (12.8) |
(Missing) | 10 318 | 92 749 |
Medication exposures | ||
Bisacodyl | 2655 (2.3) | 18 408 (1.6) |
Budesonide | 950 (0.8) | 5786 (0.5) |
Clarithromycin | 21 260 (18.7) | 130 624 (11.2) |
Colestipol | 107 (0.1) | 717 (0.1) |
Diclofenac | 48 146 (42.4) | 422 126 (36.2) |
Dipyridamole | 243 (0.2) | 1221 (0.1) |
Duloxetine | 4527 (4.0) | 33 070 (2.8) |
Erythromycin | 44 897 (39.5) | 301 785 (25.9) |
Esomeprazole | 9299 (8.2) | 66 885 (5.7) |
Fluoxetine | 4291 (3.8) | 33 164 (2.8) |
Ibuprofen | 89 112 (78.5) | 834 784 (71.6) |
Lithium | 753 (0.7) | 5022 (0.4) |
Mesalazine | 2370 (2.1) | 14 032 (1.2) |
Mianserin | 3149 (2.8) | 24 382 (2.1) |
Multienzymes | 155 (0.1) | 938 (0.1) |
Mycophenolic acid | 20 (0.02) | 126 (0.01) |
Naproxen | 20 556 (18.1) | 176 297 (15.1) |
Pentoxyverine | 208 (0.2) | 1220 (0.1) |
Propantheline | 470 (0.4) | 3741 (0.3) |
Rabeprazole | 455 (0.4) | 3124 (0.3) |
Sulfasalazine | 2277 (2.0) | 11 017 (0.9) |
Theophylline | 566 (0.5) | 2624 (0.2) |
Valproic acid | 1233 (1.1) | 8765 (0.8) |
Verapamil | 870 (0.8) | 5247 (0.5) |
Body mass index (BMI) data were available only from 2004 onward.
Statistical Analysis
We tabulated the frequency and proportion of child and maternal characteristics according to overall phthalate exposure status. We fit crude and multivariable Cox regression models to estimate associations between gestational and childhood phthalate exposures and incidence of childhood cancer (overall and type specific). Gestational exposure and childhood exposure were modeled as separate, mutually adjusted exposures. Childhood exposure was modeled as a time-varying factor and was characterized as “unexposed” until 6 months after meeting the exposure definition. The 6-month lag allowed a reasonable amount of time for the exposure to exert an effect on the outcome. We increased the lag to 12 months in a sensitivity analysis to better protect against reverse causation bias. We also carried out sensitivity analyses in which users of azathioprine were excluded, given this drug’s association with cancer (37). Models of type-specific cancers censored follow-up upon diagnosis with another cancer. We carried out preplanned stratified analyses according to natal sex and follow-up periods based on child’s age. Table cells with fewer than 5 individuals are reported as less than 5 in accordance with Danish privacy law. Associations were estimated as hazard ratios (HRs) with accompanying 95% confidence limits. To guard against potentially spurious findings across the 16 evaluated cancer sites, we subjected our associations to semi-Bayes shrinkage (38) using a conservative variance of 0.125 (consistent with a fourfold hypothetical range across all phthalate and cancer associations). Analyses were carried out with SAS version 9.4 (SAS Institute, Cary, NC); plots were created with the ggplot2 package for R (39). No hypothesis testing was performed (40-43).
Results
Searching the Danish Medicines Agency database revealed 430 marketed drug products (representing 29 medications) formulated with at least 1 phthalate. None of the phthalate-containing drug products appeared on the Danish Medicines Agency’s listing of over-the-counter medications (44). Cohort members filled prescriptions for 25 of the 29 medications, all of which were also represented by phthalate-free formulations (Supplementary Table 2, available online). We identified 1 278 685 children of whom 113 529 (8.9%) were exposed to at least 1 medication-associated phthalate during gestation and/or childhood. Table 1 reports the distribution of child and maternal characteristics according to phthalate exposure status. Children with any phthalate exposure were more likely to be born earlier in the study period [consistent with the time trend in the prevalence of phthalates in medication formulations (17)] and were more often exposed to clarithromycin, erythromycin, ibuprofen, and naproxen. Mothers of children with any phthalate exposure were more likely than mothers of unexposed children to have filled prescriptions for clarithromycin and erythromycin during pregnancy.
Tables 2 and 3 report results from our main analyses. Over 13.1 million person-years of observation, 2027 cases of childhood cancer were diagnosed. The combined rarity of phthalate exposure and type-specific cancer incidence precluded fitting complex multivariable models. We were able to fit simpler multivariable models with mutual adjustment of gestational and childhood phthalate exposures and adjustment for birth year and specific medication exposures (ie, the imbalanced factors in Table 1). Cancers diagnosed among children exposed to phthalates were somewhat more likely to be metastatic at diagnosis (43% of cases among phthalate exposed and 33% of cases among phthalate unexposed; prevalence ratio = 1.29, 95% confidence interval [CI] = 0.93 to 1.67). Table 2 reports associations between specific phthalate exposures and incidence of any childhood cancer. Exposure to any phthalate during childhood was associated with an approximately 20% higher hazard of childhood cancer (HR = 1.19, 95% CI = 0.97 to 1.47), with a median time between first phthalate exposure and cancer diagnosis of 3.3 years. Exposure to any phthalate during gestation was not associated with overall childhood cancer incidence (HR = 0.87, 95% CI = 0.60 to 1.24). Associations between childhood phthalate exposure and incidence of any cancer varied by the specific type of phthalate. We saw evidence for positive associations with exposure to DEP (HR = 1.67, 95% CI = 1.16 to 2.41), CAP (HR = 2.31, 95% CI = 0.74 to 7.20), and PVAP (HR = 4.70, 95% CI = 0.66 to 33.30). We observed a near-null association with exposure to HPMCP (HR = 1.16, 95% CI = 0.94 to 1.44). The association between childhood DBP exposure and childhood cancer incidence was not estimable because of very low exposure frequency. Childhood exposure to phthalate polymers (CAP, HPMCP, and PVAP) was weakly associated with childhood cancer (HR = 1.18, 95% CI = 0.96 to 1.46). Childhood exposure to low molecular weight phthalates (DBP and DEP) was associated with a 66% higher hazard of childhood cancer (HR = 1.66, 95% CI = 1.15 to 2.39). No type of gestational phthalate exposure was associated with incidence of any childhood cancer, with the possible exception of CAP (HR = 1.59, 95% CI = 0.51 to 4.91). Phthalate-specific associations were essentially unchanged after further adjustment for birth year (data not shown).
Phthalate . | Exposure period . | No. of cases . | Person-years . | Hazard ratio (95% CI)a . |
---|---|---|---|---|
Any phthalate | Gestation | |||
Unexposed | 1997 | 12 878 437 | 1.00 (Referent) | |
Exposed | 30 | 232 311 | 0.87 (0.60 to 1.24) | |
Childhood | ||||
Unexposed | 1923 | 12 516 243 | 1.00 (Referent) | |
Exposed | 95 | 594 504 | 1.19 (0.97 to 1.47) | |
Cellulose acetate phthalate (CAP) | Gestation | |||
Unexposed | <5 | 13 098 472 | 1.00 (Referent) | |
Exposed | <5 | 12 276 | 1.59 (0.51 to 4.91) | |
Childhood | ||||
Unexposed | <5 | 13 103 506 | 1.00 (Referent) | |
Exposed | <5 | 7241 | 2.31 (0.74 to 7.20) | |
Dibutyl phthalate (DBP) | Gestation | |||
Unexposed | <5 | 13 095 380 | 1.00 (Referent) | |
Exposed | <5 | 15 367 | 0.89 (0.22 to 3.55) | |
Childhood | ||||
Unexposed | 2018 | 13 109 090 | 1.00 (Referent) | |
Exposed | <5 | 1658 | Not estimable | |
Diethyl phthalate (DEP) | Gestation | |||
Unexposed | 2002 | 12 904 083 | 1.00 (Referent) | |
Exposed | 25 | 206 665 | 0.81 (0.55 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 982 931 | 1.00 (Referent) | |
Exposed | 30 | 127 816 | 1.67 (1.16 to 2.41) | |
Hydroxypropyl methylcellulose phthalate (HPMCP) | Gestation | |||
Unexposed | 2006 | 12 937 462 | 1.00 (Referent) | |
Exposed | 21 | 173 285 | 0.81 (0.53 to 1.25) | |
Childhood | ||||
Unexposed | 1927 | 12 524 945 | 1.00 (Referent) | |
Exposed | 91 | 585 802 | 1.16 (0.94 to 1.44) | |
Polyvinyl acetate phthalate (PVAP) | Gestation | |||
Unexposed | 2027 | 13 110 318 | 1.00 (Referent) | |
Exposed | <5 | 429 | Not estimable | |
Childhood | ||||
Unexposed | <5 | 13 109 241 | 1.00 (Referent) | |
Exposed | <5 | 1507 | 4.70 (0.66 to 33.30) | |
Phthalate polymers (CAP, HPMCP, and PVAP) | Gestation | |||
Unexposed | 2003 | 12 925 172 | 1.00 (Referent) | |
Exposed | 24 | 185 575 | 0.86 (0.58 to 1.29) | |
Childhood | ||||
Unexposed | 1924 | 12 517 554 | 1.00 (Referent) | |
Exposed | 94 | 593 193 | 1.18 (0.96 to 1.46) | |
Low molecular weight phthalates (DBP and DEP) | Gestation | |||
Unexposed | 2000 | 12 889 001 | 1.00 (Referent) | |
Exposed | 27 | 221 746 | 0.82 (0.56 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 981 592 | 1.00 (Referent) | |
Exposed | 30 | 129 156 | 1.66 (1.15 to 2.39) |
Phthalate . | Exposure period . | No. of cases . | Person-years . | Hazard ratio (95% CI)a . |
---|---|---|---|---|
Any phthalate | Gestation | |||
Unexposed | 1997 | 12 878 437 | 1.00 (Referent) | |
Exposed | 30 | 232 311 | 0.87 (0.60 to 1.24) | |
Childhood | ||||
Unexposed | 1923 | 12 516 243 | 1.00 (Referent) | |
Exposed | 95 | 594 504 | 1.19 (0.97 to 1.47) | |
Cellulose acetate phthalate (CAP) | Gestation | |||
Unexposed | <5 | 13 098 472 | 1.00 (Referent) | |
Exposed | <5 | 12 276 | 1.59 (0.51 to 4.91) | |
Childhood | ||||
Unexposed | <5 | 13 103 506 | 1.00 (Referent) | |
Exposed | <5 | 7241 | 2.31 (0.74 to 7.20) | |
Dibutyl phthalate (DBP) | Gestation | |||
Unexposed | <5 | 13 095 380 | 1.00 (Referent) | |
Exposed | <5 | 15 367 | 0.89 (0.22 to 3.55) | |
Childhood | ||||
Unexposed | 2018 | 13 109 090 | 1.00 (Referent) | |
Exposed | <5 | 1658 | Not estimable | |
Diethyl phthalate (DEP) | Gestation | |||
Unexposed | 2002 | 12 904 083 | 1.00 (Referent) | |
Exposed | 25 | 206 665 | 0.81 (0.55 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 982 931 | 1.00 (Referent) | |
Exposed | 30 | 127 816 | 1.67 (1.16 to 2.41) | |
Hydroxypropyl methylcellulose phthalate (HPMCP) | Gestation | |||
Unexposed | 2006 | 12 937 462 | 1.00 (Referent) | |
Exposed | 21 | 173 285 | 0.81 (0.53 to 1.25) | |
Childhood | ||||
Unexposed | 1927 | 12 524 945 | 1.00 (Referent) | |
Exposed | 91 | 585 802 | 1.16 (0.94 to 1.44) | |
Polyvinyl acetate phthalate (PVAP) | Gestation | |||
Unexposed | 2027 | 13 110 318 | 1.00 (Referent) | |
Exposed | <5 | 429 | Not estimable | |
Childhood | ||||
Unexposed | <5 | 13 109 241 | 1.00 (Referent) | |
Exposed | <5 | 1507 | 4.70 (0.66 to 33.30) | |
Phthalate polymers (CAP, HPMCP, and PVAP) | Gestation | |||
Unexposed | 2003 | 12 925 172 | 1.00 (Referent) | |
Exposed | 24 | 185 575 | 0.86 (0.58 to 1.29) | |
Childhood | ||||
Unexposed | 1924 | 12 517 554 | 1.00 (Referent) | |
Exposed | 94 | 593 193 | 1.18 (0.96 to 1.46) | |
Low molecular weight phthalates (DBP and DEP) | Gestation | |||
Unexposed | 2000 | 12 889 001 | 1.00 (Referent) | |
Exposed | 27 | 221 746 | 0.82 (0.56 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 981 592 | 1.00 (Referent) | |
Exposed | 30 | 129 156 | 1.66 (1.15 to 2.39) |
Gestational and childhood phthalate exposures were modeled as independent terms (ie, mutually adjusted) in Cox regression models. CI = confidence interval.
Phthalate . | Exposure period . | No. of cases . | Person-years . | Hazard ratio (95% CI)a . |
---|---|---|---|---|
Any phthalate | Gestation | |||
Unexposed | 1997 | 12 878 437 | 1.00 (Referent) | |
Exposed | 30 | 232 311 | 0.87 (0.60 to 1.24) | |
Childhood | ||||
Unexposed | 1923 | 12 516 243 | 1.00 (Referent) | |
Exposed | 95 | 594 504 | 1.19 (0.97 to 1.47) | |
Cellulose acetate phthalate (CAP) | Gestation | |||
Unexposed | <5 | 13 098 472 | 1.00 (Referent) | |
Exposed | <5 | 12 276 | 1.59 (0.51 to 4.91) | |
Childhood | ||||
Unexposed | <5 | 13 103 506 | 1.00 (Referent) | |
Exposed | <5 | 7241 | 2.31 (0.74 to 7.20) | |
Dibutyl phthalate (DBP) | Gestation | |||
Unexposed | <5 | 13 095 380 | 1.00 (Referent) | |
Exposed | <5 | 15 367 | 0.89 (0.22 to 3.55) | |
Childhood | ||||
Unexposed | 2018 | 13 109 090 | 1.00 (Referent) | |
Exposed | <5 | 1658 | Not estimable | |
Diethyl phthalate (DEP) | Gestation | |||
Unexposed | 2002 | 12 904 083 | 1.00 (Referent) | |
Exposed | 25 | 206 665 | 0.81 (0.55 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 982 931 | 1.00 (Referent) | |
Exposed | 30 | 127 816 | 1.67 (1.16 to 2.41) | |
Hydroxypropyl methylcellulose phthalate (HPMCP) | Gestation | |||
Unexposed | 2006 | 12 937 462 | 1.00 (Referent) | |
Exposed | 21 | 173 285 | 0.81 (0.53 to 1.25) | |
Childhood | ||||
Unexposed | 1927 | 12 524 945 | 1.00 (Referent) | |
Exposed | 91 | 585 802 | 1.16 (0.94 to 1.44) | |
Polyvinyl acetate phthalate (PVAP) | Gestation | |||
Unexposed | 2027 | 13 110 318 | 1.00 (Referent) | |
Exposed | <5 | 429 | Not estimable | |
Childhood | ||||
Unexposed | <5 | 13 109 241 | 1.00 (Referent) | |
Exposed | <5 | 1507 | 4.70 (0.66 to 33.30) | |
Phthalate polymers (CAP, HPMCP, and PVAP) | Gestation | |||
Unexposed | 2003 | 12 925 172 | 1.00 (Referent) | |
Exposed | 24 | 185 575 | 0.86 (0.58 to 1.29) | |
Childhood | ||||
Unexposed | 1924 | 12 517 554 | 1.00 (Referent) | |
Exposed | 94 | 593 193 | 1.18 (0.96 to 1.46) | |
Low molecular weight phthalates (DBP and DEP) | Gestation | |||
Unexposed | 2000 | 12 889 001 | 1.00 (Referent) | |
Exposed | 27 | 221 746 | 0.82 (0.56 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 981 592 | 1.00 (Referent) | |
Exposed | 30 | 129 156 | 1.66 (1.15 to 2.39) |
Phthalate . | Exposure period . | No. of cases . | Person-years . | Hazard ratio (95% CI)a . |
---|---|---|---|---|
Any phthalate | Gestation | |||
Unexposed | 1997 | 12 878 437 | 1.00 (Referent) | |
Exposed | 30 | 232 311 | 0.87 (0.60 to 1.24) | |
Childhood | ||||
Unexposed | 1923 | 12 516 243 | 1.00 (Referent) | |
Exposed | 95 | 594 504 | 1.19 (0.97 to 1.47) | |
Cellulose acetate phthalate (CAP) | Gestation | |||
Unexposed | <5 | 13 098 472 | 1.00 (Referent) | |
Exposed | <5 | 12 276 | 1.59 (0.51 to 4.91) | |
Childhood | ||||
Unexposed | <5 | 13 103 506 | 1.00 (Referent) | |
Exposed | <5 | 7241 | 2.31 (0.74 to 7.20) | |
Dibutyl phthalate (DBP) | Gestation | |||
Unexposed | <5 | 13 095 380 | 1.00 (Referent) | |
Exposed | <5 | 15 367 | 0.89 (0.22 to 3.55) | |
Childhood | ||||
Unexposed | 2018 | 13 109 090 | 1.00 (Referent) | |
Exposed | <5 | 1658 | Not estimable | |
Diethyl phthalate (DEP) | Gestation | |||
Unexposed | 2002 | 12 904 083 | 1.00 (Referent) | |
Exposed | 25 | 206 665 | 0.81 (0.55 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 982 931 | 1.00 (Referent) | |
Exposed | 30 | 127 816 | 1.67 (1.16 to 2.41) | |
Hydroxypropyl methylcellulose phthalate (HPMCP) | Gestation | |||
Unexposed | 2006 | 12 937 462 | 1.00 (Referent) | |
Exposed | 21 | 173 285 | 0.81 (0.53 to 1.25) | |
Childhood | ||||
Unexposed | 1927 | 12 524 945 | 1.00 (Referent) | |
Exposed | 91 | 585 802 | 1.16 (0.94 to 1.44) | |
Polyvinyl acetate phthalate (PVAP) | Gestation | |||
Unexposed | 2027 | 13 110 318 | 1.00 (Referent) | |
Exposed | <5 | 429 | Not estimable | |
Childhood | ||||
Unexposed | <5 | 13 109 241 | 1.00 (Referent) | |
Exposed | <5 | 1507 | 4.70 (0.66 to 33.30) | |
Phthalate polymers (CAP, HPMCP, and PVAP) | Gestation | |||
Unexposed | 2003 | 12 925 172 | 1.00 (Referent) | |
Exposed | 24 | 185 575 | 0.86 (0.58 to 1.29) | |
Childhood | ||||
Unexposed | 1924 | 12 517 554 | 1.00 (Referent) | |
Exposed | 94 | 593 193 | 1.18 (0.96 to 1.46) | |
Low molecular weight phthalates (DBP and DEP) | Gestation | |||
Unexposed | 2000 | 12 889 001 | 1.00 (Referent) | |
Exposed | 27 | 221 746 | 0.82 (0.56 to 1.20) | |
Childhood | ||||
Unexposed | 1988 | 12 981 592 | 1.00 (Referent) | |
Exposed | 30 | 129 156 | 1.66 (1.15 to 2.39) |
Gestational and childhood phthalate exposures were modeled as independent terms (ie, mutually adjusted) in Cox regression models. CI = confidence interval.
Cancer type . | Exposure period . | No. of cases . | Median duration of prediagnosis exposure, ya . | Hazard ratio (95% CI) b . |
---|---|---|---|---|
Leukemia, combined | Gestation | |||
Unexposed | 673 | 1.00 (Referent) | ||
Exposed | 12 | 1.07 (0.60 to 1.89) | ||
Childhood | ||||
Unexposed | 663 | 1.00 (Referent) | ||
Exposed | 22 | 2.5 | 0.85 (0.55 to 1.30) | |
Leukemia, acute lymphocytic | Gestation | |||
Unexposed | 518 | 1.00 (Referent) | ||
Exposed | 11 | 1.27 (0.70 to 2.31) | ||
Childhood | ||||
Unexposed | 512 | 1.00 (Referent) | ||
Exposed | 17 | 2.5 | 0.81 (0.50 to 1.32) | |
Leukemia, acute myeloid | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.66 (0.09 to 4.70) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.4 | 1.22 (0.44 to 3.37) | |
Lymphoma, combined | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.65 (0.21 to 2.03) | ||
Childhood | ||||
Unexposed | 218 | 1.00 (Referent) | ||
Exposed | 25 | 4.9 | 2.07 (1.36 to 3.14) | |
Lymphoma, non-Hodgkin | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.10 (0.35 to 3.45) | ||
Childhood | ||||
Unexposed | 138 | 1.00 (Referent) | ||
Exposed | 15 | 2.6 | 2.29 (1.33 to 3.92) | |
Lymphoma, Hodgkin | Gestation | |||
Unexposed | 69 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | 60 | 1.00 (Referent) | ||
Exposed | 9 | 7.7 | 2.04 (1.01 to 4.12) | |
Lymphoma, Burkitt | Gestation | |||
Unexposed | 21 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 0.90 (0.12 to 6.75) | |
Brain or Central nervous system | Gestation | |||
Unexposed | 263 | 1.00 (Referent) | ||
Exposed | 5 | 1.10 (0.45 to 2.67) | ||
Childhood | ||||
Unexposed | 256 | 1.00 (Referent) | ||
Exposed | 9 | 1.2 | 0.84 (0.43 to 1.65) | |
Neuroblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.59 (0.08 to 4.25) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.2 | 1.57 (0.49 to 5.04) | |
Retinoblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.48 (0.37 to 6.03) | ||
Childhood | ||||
Unexposed | 93 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | Not estimable | |
Kidney | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.61 (0.08 to 4.36) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.17 (0.37 to 3.72) | |
Liver | Gestation | |||
Unexposed | 37 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.18 (0.16 to 8.86) | |
Osteosarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.46 (0.06 to 3.27) | ||
Childhood | ||||
Unexposed | 90 | 1.00 (Referent) | ||
Exposed | 16 | 6.5 | 2.78 (1.63 to 4.75) | |
Soft tissue sarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.17 (0.29 to 4.76) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.8 | 0.98 (0.36 to 2.69) | |
Reproductive organ | Gestation | |||
Unexposed | 57 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 2.1 | 0.38 (0.05 to 2.78) | |
Other or unclassified | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.73 (0.24 to 2.30) | ||
Childhood | ||||
Unexposed | 203 | 1.00 (Referent) | ||
Exposed | 11 | 5.6 | 0.95 (0.52 to 1.75) |
Cancer type . | Exposure period . | No. of cases . | Median duration of prediagnosis exposure, ya . | Hazard ratio (95% CI) b . |
---|---|---|---|---|
Leukemia, combined | Gestation | |||
Unexposed | 673 | 1.00 (Referent) | ||
Exposed | 12 | 1.07 (0.60 to 1.89) | ||
Childhood | ||||
Unexposed | 663 | 1.00 (Referent) | ||
Exposed | 22 | 2.5 | 0.85 (0.55 to 1.30) | |
Leukemia, acute lymphocytic | Gestation | |||
Unexposed | 518 | 1.00 (Referent) | ||
Exposed | 11 | 1.27 (0.70 to 2.31) | ||
Childhood | ||||
Unexposed | 512 | 1.00 (Referent) | ||
Exposed | 17 | 2.5 | 0.81 (0.50 to 1.32) | |
Leukemia, acute myeloid | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.66 (0.09 to 4.70) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.4 | 1.22 (0.44 to 3.37) | |
Lymphoma, combined | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.65 (0.21 to 2.03) | ||
Childhood | ||||
Unexposed | 218 | 1.00 (Referent) | ||
Exposed | 25 | 4.9 | 2.07 (1.36 to 3.14) | |
Lymphoma, non-Hodgkin | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.10 (0.35 to 3.45) | ||
Childhood | ||||
Unexposed | 138 | 1.00 (Referent) | ||
Exposed | 15 | 2.6 | 2.29 (1.33 to 3.92) | |
Lymphoma, Hodgkin | Gestation | |||
Unexposed | 69 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | 60 | 1.00 (Referent) | ||
Exposed | 9 | 7.7 | 2.04 (1.01 to 4.12) | |
Lymphoma, Burkitt | Gestation | |||
Unexposed | 21 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 0.90 (0.12 to 6.75) | |
Brain or Central nervous system | Gestation | |||
Unexposed | 263 | 1.00 (Referent) | ||
Exposed | 5 | 1.10 (0.45 to 2.67) | ||
Childhood | ||||
Unexposed | 256 | 1.00 (Referent) | ||
Exposed | 9 | 1.2 | 0.84 (0.43 to 1.65) | |
Neuroblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.59 (0.08 to 4.25) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.2 | 1.57 (0.49 to 5.04) | |
Retinoblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.48 (0.37 to 6.03) | ||
Childhood | ||||
Unexposed | 93 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | Not estimable | |
Kidney | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.61 (0.08 to 4.36) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.17 (0.37 to 3.72) | |
Liver | Gestation | |||
Unexposed | 37 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.18 (0.16 to 8.86) | |
Osteosarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.46 (0.06 to 3.27) | ||
Childhood | ||||
Unexposed | 90 | 1.00 (Referent) | ||
Exposed | 16 | 6.5 | 2.78 (1.63 to 4.75) | |
Soft tissue sarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.17 (0.29 to 4.76) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.8 | 0.98 (0.36 to 2.69) | |
Reproductive organ | Gestation | |||
Unexposed | 57 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 2.1 | 0.38 (0.05 to 2.78) | |
Other or unclassified | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.73 (0.24 to 2.30) | ||
Childhood | ||||
Unexposed | 203 | 1.00 (Referent) | ||
Exposed | 11 | 5.6 | 0.95 (0.52 to 1.75) |
Median time elapsed from receipt of first phthalate-containing prescription drug product until cancer diagnosis (applicable only to time-varying childhood exposure). CI = confidence interval.
Gestational and childhood phthalate exposures were modeled as independent terms (ie, mutually adjusted) in Cox regression models.
Cancer type . | Exposure period . | No. of cases . | Median duration of prediagnosis exposure, ya . | Hazard ratio (95% CI) b . |
---|---|---|---|---|
Leukemia, combined | Gestation | |||
Unexposed | 673 | 1.00 (Referent) | ||
Exposed | 12 | 1.07 (0.60 to 1.89) | ||
Childhood | ||||
Unexposed | 663 | 1.00 (Referent) | ||
Exposed | 22 | 2.5 | 0.85 (0.55 to 1.30) | |
Leukemia, acute lymphocytic | Gestation | |||
Unexposed | 518 | 1.00 (Referent) | ||
Exposed | 11 | 1.27 (0.70 to 2.31) | ||
Childhood | ||||
Unexposed | 512 | 1.00 (Referent) | ||
Exposed | 17 | 2.5 | 0.81 (0.50 to 1.32) | |
Leukemia, acute myeloid | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.66 (0.09 to 4.70) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.4 | 1.22 (0.44 to 3.37) | |
Lymphoma, combined | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.65 (0.21 to 2.03) | ||
Childhood | ||||
Unexposed | 218 | 1.00 (Referent) | ||
Exposed | 25 | 4.9 | 2.07 (1.36 to 3.14) | |
Lymphoma, non-Hodgkin | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.10 (0.35 to 3.45) | ||
Childhood | ||||
Unexposed | 138 | 1.00 (Referent) | ||
Exposed | 15 | 2.6 | 2.29 (1.33 to 3.92) | |
Lymphoma, Hodgkin | Gestation | |||
Unexposed | 69 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | 60 | 1.00 (Referent) | ||
Exposed | 9 | 7.7 | 2.04 (1.01 to 4.12) | |
Lymphoma, Burkitt | Gestation | |||
Unexposed | 21 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 0.90 (0.12 to 6.75) | |
Brain or Central nervous system | Gestation | |||
Unexposed | 263 | 1.00 (Referent) | ||
Exposed | 5 | 1.10 (0.45 to 2.67) | ||
Childhood | ||||
Unexposed | 256 | 1.00 (Referent) | ||
Exposed | 9 | 1.2 | 0.84 (0.43 to 1.65) | |
Neuroblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.59 (0.08 to 4.25) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.2 | 1.57 (0.49 to 5.04) | |
Retinoblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.48 (0.37 to 6.03) | ||
Childhood | ||||
Unexposed | 93 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | Not estimable | |
Kidney | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.61 (0.08 to 4.36) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.17 (0.37 to 3.72) | |
Liver | Gestation | |||
Unexposed | 37 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.18 (0.16 to 8.86) | |
Osteosarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.46 (0.06 to 3.27) | ||
Childhood | ||||
Unexposed | 90 | 1.00 (Referent) | ||
Exposed | 16 | 6.5 | 2.78 (1.63 to 4.75) | |
Soft tissue sarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.17 (0.29 to 4.76) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.8 | 0.98 (0.36 to 2.69) | |
Reproductive organ | Gestation | |||
Unexposed | 57 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 2.1 | 0.38 (0.05 to 2.78) | |
Other or unclassified | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.73 (0.24 to 2.30) | ||
Childhood | ||||
Unexposed | 203 | 1.00 (Referent) | ||
Exposed | 11 | 5.6 | 0.95 (0.52 to 1.75) |
Cancer type . | Exposure period . | No. of cases . | Median duration of prediagnosis exposure, ya . | Hazard ratio (95% CI) b . |
---|---|---|---|---|
Leukemia, combined | Gestation | |||
Unexposed | 673 | 1.00 (Referent) | ||
Exposed | 12 | 1.07 (0.60 to 1.89) | ||
Childhood | ||||
Unexposed | 663 | 1.00 (Referent) | ||
Exposed | 22 | 2.5 | 0.85 (0.55 to 1.30) | |
Leukemia, acute lymphocytic | Gestation | |||
Unexposed | 518 | 1.00 (Referent) | ||
Exposed | 11 | 1.27 (0.70 to 2.31) | ||
Childhood | ||||
Unexposed | 512 | 1.00 (Referent) | ||
Exposed | 17 | 2.5 | 0.81 (0.50 to 1.32) | |
Leukemia, acute myeloid | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.66 (0.09 to 4.70) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.4 | 1.22 (0.44 to 3.37) | |
Lymphoma, combined | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.65 (0.21 to 2.03) | ||
Childhood | ||||
Unexposed | 218 | 1.00 (Referent) | ||
Exposed | 25 | 4.9 | 2.07 (1.36 to 3.14) | |
Lymphoma, non-Hodgkin | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.10 (0.35 to 3.45) | ||
Childhood | ||||
Unexposed | 138 | 1.00 (Referent) | ||
Exposed | 15 | 2.6 | 2.29 (1.33 to 3.92) | |
Lymphoma, Hodgkin | Gestation | |||
Unexposed | 69 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | 60 | 1.00 (Referent) | ||
Exposed | 9 | 7.7 | 2.04 (1.01 to 4.12) | |
Lymphoma, Burkitt | Gestation | |||
Unexposed | 21 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 0.90 (0.12 to 6.75) | |
Brain or Central nervous system | Gestation | |||
Unexposed | 263 | 1.00 (Referent) | ||
Exposed | 5 | 1.10 (0.45 to 2.67) | ||
Childhood | ||||
Unexposed | 256 | 1.00 (Referent) | ||
Exposed | 9 | 1.2 | 0.84 (0.43 to 1.65) | |
Neuroblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.59 (0.08 to 4.25) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.2 | 1.57 (0.49 to 5.04) | |
Retinoblastoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.48 (0.37 to 6.03) | ||
Childhood | ||||
Unexposed | 93 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | Not estimable | |
Kidney | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.61 (0.08 to 4.36) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.17 (0.37 to 3.72) | |
Liver | Gestation | |||
Unexposed | 37 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | Not reportable | 1.18 (0.16 to 8.86) | |
Osteosarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.46 (0.06 to 3.27) | ||
Childhood | ||||
Unexposed | 90 | 1.00 (Referent) | ||
Exposed | 16 | 6.5 | 2.78 (1.63 to 4.75) | |
Soft tissue sarcoma | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 1.17 (0.29 to 4.76) | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 3.8 | 0.98 (0.36 to 2.69) | |
Reproductive organ | Gestation | |||
Unexposed | 57 | 1.00 (Referent) | ||
Exposed | <5 | Not estimable | ||
Childhood | ||||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 2.1 | 0.38 (0.05 to 2.78) | |
Other or unclassified | Gestation | |||
Unexposed | <5 | 1.00 (Referent) | ||
Exposed | <5 | 0.73 (0.24 to 2.30) | ||
Childhood | ||||
Unexposed | 203 | 1.00 (Referent) | ||
Exposed | 11 | 5.6 | 0.95 (0.52 to 1.75) |
Median time elapsed from receipt of first phthalate-containing prescription drug product until cancer diagnosis (applicable only to time-varying childhood exposure). CI = confidence interval.
Gestational and childhood phthalate exposures were modeled as independent terms (ie, mutually adjusted) in Cox regression models.
Table 3 reports associations between gestational and childhood phthalate exposures and incidence of specific cancers or cancer groups from models with a 6-month exposure lag. Several associations were not estimable because of sparse data. Whereas most estimable associations were near null or null centered, we observed a positive association between childhood phthalate exposure and osteosarcoma (HR = 2.78, 95% CI = 1.63 to 4.75). Childhood phthalate exposure was also associated with combined lymphoma (HR = 2.07, 95% CI = 1.36 to 3.14), driven by associations with non-Hodgkin lymphoma (HR = 2.29, 95% CI = 1.33 to 3.92) and Hodgkin lymphoma (HR = 2.04, 95% CI = 1.01 to 4.12). The osteosarcoma association persisted after further adjustment for birth year and for childhood exposure to clarithromycin, erythromycin, ibuprofen, and naproxen (HRadj = 2.78, 95% CI = 1.63 to 4.76). It was also similar under the 12-month exposure lag (HR = 2.91, 95% CI = 1.70 to 4.98) and after excluding azathioprine users from the analysis (HR = 2.65, 95% CI = 1.53 to 4.60). Lymphoma associations were similarly robust to multivariable adjustment, increasing the exposure lag to 12 months and excluding azathioprine users. Associations between childhood phthalate exposure and type-specific cancers were not substantially changed after semi-Bayes shrinkage (Supplementary Table 3, available online) or by requiring 2 filled prescriptions to define exposure (Supplementary Figure 1, available online). Associations between gestational phthalate exposure and type-specific cancer incidence were either not estimable or near null.
Our data could not support sex-stratified analyses of type-specific cancer associations or of phthalate-specific associations. However, we observed no modification by natal sex of the hazard ratios associating gestational or childhood phthalate exposure and incidence of any childhood cancer (Figure 1). The osteosarcoma association was slightly stronger in natal female children than in natal male children (Figure 1). Type-specific models for leukemia, combined lymphoma, and osteosarcoma were generated for age-specific follow-up periods (age 0-5 years vs 6-19 years for leukemia, and age 0-9 years vs 10-19 years for lymphoma and osteosarcoma; Figure 2). The near-null leukemia association was not substantially modified by age at diagnosis. The positive association between childhood phthalate exposure and incidence of combined lymphoma (Table 3; HR = 2.07, 95% CI = 1.36 to 3.14) was apparent in the 0- to 9-year and 10- to 19-year follow-up periods (Figure 2). The positive association between childhood phthalate exposure and osteosarcoma was also apparent in both follow-up periods.
Discussion
Childhood phthalate exposure was strongly associated with incidence of osteosarcoma. This association was slightly stronger for natal females, did not vary according to follow-up period, and was robust to adjustment for birth year and exposure to clarithromycin, erythromycin, ibuprofen, and naproxen—the only measured factors that differed between phthalate-exposed and -unexposed children. Exclusion of azathioprine users did not materially affect the osteosarcoma association nor did semi-Bayes shrinkage or requiring 2 phthalate-containing prescriptions to define exposure. The association also persisted when the exposure lag period increased from 6 to 12 months, providing assurance that the association cannot be explained by use of analgesics or antibiotics to manage symptoms of subclinical disease such as bone and/or joint pain or infection. Osteosarcoma may have a hormonal etiology because its sex-specific incidence patterns track closely with puberty onset and because of its link with childhood growth hormone exposure (22,45). It is therefore plausible that endocrine disruption by phthalates could increase osteosarcoma risk. This association is also consistent with evidence linking phthalate exposure with early onset of puberty (46-48), which has in turn been linked with osteosarcoma (49).
We also observed a positive association between childhood phthalate exposure and incidence of lymphoma, driven by associations with Hodgkin and non-Hodgkin lymphoma. These associations were robust to multivariable adjustment, increase to a 12-month exposure lag, exclusion of azathioprine users, semi-Bayes shrinkage, and requiring 2 phthalate-containing prescriptions for exposure. The potential mechanisms underlying the lymphoma association are unclear and warrant further study.
Exposure to any phthalate during childhood was associated with an approximately 20% higher incidence hazard for any childhood cancer. Combined childhood exposure to DBP and DEP-low molecular weight phthalates with purportedly greater biological activity (6) was more strongly associated with cancer incidence than childhood exposure to high molecular weight phthalate polymers.
Most associations between gestational phthalate exposure and site-specific childhood cancers were near null, and all were measured with poor precision. Although some of these associations appeared protective and were therefore of oppositive direction to childhood exposure associations (ie, for combined lymphoma and osteosarcoma), their imprecision precluded meaningful comparison.
Ours is the first study to measure associations between gestational and childhood phthalate exposures and childhood cancer incidence. Others have studied phthalate exposure in relation to adult hormone-related malignancies such as thyroid, prostate, and breast cancer (50-56). Two studies reported positive associations between urinary DEHP metabolites and thyroid cancer incidence (55,56). DEHP, butyl-benzyl phthalate, and di-isobutyl phthalate were all positively associated with prostate cancer incidence in abdominally obese men (54). We could not assess these specific phthalates, as they were not used in any pharmaceuticals marketed in Denmark. Breast cancer findings are inconsistent, with only modest agreement across studies (50-53). Our group recently studied the impact of medication-associated phthalate exposure on breast cancer incidence in a Danish cohort of more than 1 million women (57). We observed a twofold higher hazard for estrogen receptor–positive breast cancer among women with high exposure to DBP, which mimics estrogen in vitro (57,58). Another Danish study reported associations between high cumulative exposure to DBP and DEP and incident colorectal cancer (59).
Key strengths of our study are its large size, virtual immunity from selection bias, completeness of follow-up, and high-validity cancer registry data (15,60). Most limitations of our study are a consequence of sparse data resulting from the combination of rare exposures with rare outcomes. We initially divided gestational exposures into trimesters; however, the categories were too sparse to support estimation and were ultimately combined. This may have masked true associations by mixing etiologically relevant and etiologically inert exposure windows. We also sought to cross-classify gestational and childhood exposures so we could assess their interdependence, but this approach was abandoned because of model-fitting issues. Sparse data also limited our ability to specify complex multivariable models for all cancer outcomes to control for the comprehensive set of measured maternal and child factors that could potentially confound associations between phthalate exposure and childhood cancer incidence. Most outcome models could not accommodate any covariables beyond the 2 terms for gestational and childhood phthalate exposures. Nonetheless, we noted a high degree of balance in most measured candidate confounders between phthalate exposure groups. Such balance is expected because receipt of a phthalate-containing drug formulation instead of a phthalate-free formulation is essentially random, a function of which specific drug product was available at the pharmacy when the patient presented. Adjusting regression models for the few factors that showed imbalances between exposure groups did not change our estimates. We are, therefore, confident that the associations we observed were not generated by uncontrolled confounding.
Two important sources of measurement error warrant consideration. First, our medication-based classification scheme could not capture phthalate exposure through other environmental sources. Our reference group of phthalate-unexposed participants is therefore certain to contain individuals with exposure from nonmedication sources, which could have biased association estimates toward the null. Because phthalate exposure is near ubiquitous in modern society, it would be impossible to enumerate a truly unexposed reference group. Nonetheless, the considerably higher exposure observed among people who ingest phthalate-containing medications compared with those exposed only through background sources (13) provides reassurance that nondifferential exposure misclassification would not be sufficient to mask true associations. Second, our classification scheme inherently assumed that filled prescriptions were ingested by mothers and children, which we were not able to confirm. However, a prescription record in the DNPR indicates that a patient presented to a pharmacy and tendered a copay for the filled prescription, implying intent to adhere to the medication order. Finally, we did not have information on the specific indication(s) for which phthalate-containing medications were prescribed. However, all phthalate-containing drug classes were also well represented by phthalate-free formulations, and presence or absence of phthalates in a relevant prescription fill was expected to be random. This is also expected to be true for extended-release drug formulations, as phthalates are not the only excipient used to impart this property. Together, these features provide reassurance that our associations are not likely attributable to confounding by indication or by active drug ingredients.
In summary, we observed positive associations between childhood phthalate exposure and the incidence of osteosarcoma and lymphoma in a Danish nationwide cohort. Lingering questions include which specific phthalate(s) are responsible for these associations, by what mechanism(s) they occur, the impact of dose and timing, and to what extent childhood cancer cases could be avoided by reducing or eliminating the phthalate content of medications and other consumer goods.
Funding
This work was supported by an award from the Saint Baldrick’s Foundation (grant number 584644 to TPA).
Notes
Role of the funder: The funder had no role in study design, data collection, data analysis, interpretation of results, and manuscript preparation or decision to submit the manuscript for publication.
Disclosures: No author reported a conflict of interest.
Author contributions: Conceptualization (TPA, PD, LGS); Funding acquisition (TPA); Data curation (PD, DPC-F, HTS, BOE, SPU, KE); Formal analysis & Software (BÖE, SPU, KE); Methodology (TPA, DPC-F, TLL, LGS); Supervision (TPA, DPC-F); Visualization (TPA); Writing—original draft (TPA); Writing—review and editing (TPA, LGS, PD, BÖE, TLL, HTS, DPC-F).
Data Availability
Danish registry data, including those used for this study, are protected by Danish privacy law and are accessible only through application to appropriate Danish authorities.
References
Danish Medicines Agency. Over-the-counter medicines. https://laegemiddelstyrelsen.dk/en/pharmacies/over-the-counter-medicines/. Accessed December 8, 2021.