Abstract
Objective
Cigarette smoking increases the risk of seropositive adult rheumatoid arthritis. The relationship of smoking with juvenile idiopathic arthritis (JIA), a heterogeneous group of 7 mutually exclusive categories of chronic childhood inflammatory arthritides, is unknown. Our objective was to evaluate the association between JIA and its categories with maternal prenatal smoking.
Methods
This case–control study used International Classification of Diseases, Ninth Revision codes from hospital records to identify 1,196 JIA cases born in Washington state and diagnosed at a quaternary pediatric center from 1997–2010. Controls (n = 5,618) were randomly selected from birth records of children without JIA, frequency matched on birth year. Prenatal smoking exposure was assessed from subjects’ birth certificates. Chart review categorized JIA into International League of Associations for Rheumatology categories. Adjusted odds ratios (OR) and 95% confidence intervals (95% CIs) were calculated using logistic regression.
Results
We did not observe an increased risk of JIA in relation to maternal prenatal smoking. Prenatal smoking was reported less often among mothers of JIA cases (11%), than among control mothers (17%; OR 0.71 [95% CI 0.58–0.87]), a relationship somewhat more marked for oligoarticular/extended oligoarticular JIA. Although this relationship persisted after adjustment, we cannot rule out that the effect may have been due to residual confounding by socioeconomic status.
Conclusion
We did not observe an increased risk of JIA or its individual categories with maternal prenatal smoking.
Introduction
Juvenile idiopathic arthritis (JIA) is a heterogeneous group of chronic inflammatory arthritides in children ages <16 years and is believed to be multifactorial in origin, secondary to environmental influences in a genetically susceptible host. JIA has an estimated incidence of 13.9 per 100,000 children and prevalence of 113.4 per 100,000 (1). JIA is categorized into 7 mutually exclusive categories by International League of Associations for Rheumatology (ILAR) criteria based on clinical presentation and the number of joints involved at onset (2).
Aside from genetic predisposition in certain categories of JIA, little is known about potential environmental etiologies. Studies of potentially predisposing environmental factors for JIA are few, due to rarity of disease, lack of registries to facilitate uniform surveillance and case identification, and heterogeneity of JIA.
Smoking is a well-established risk factor for development of adult seropositive rheumatoid arthritis (RA) (3); however, the relationship of maternal prenatal smoking to JIA is not known. Prenatal smoke exposure is associated with various childhood diseases and conditions and exposes the fetus to toxins potentially resulting in epigenetic changes (4). We explored the associations of JIA and its categories in relation to maternal prenatal smoking in a case–control study using linked clinic/hospital birth certificate records for JIA children treated at a major regional children’s hospital.
Subjects and methods
This study was approved by the Washington State Institutional Review Board and Seattle Children’s Hospital (SCH) Institutional Review Board. Potential cases included children with International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes indicating JIA (720.0, 696.0, 714.32, 714.31, 714.30, 720.89, 714.33, and 714.3), during 1997–2010 at SCH and an outpatient facility (n = 2,274). Cases included those with Washington state birth certificates (n = 1,518). Control children were randomly selected (control:case ratio 4:1) without any exclusions from remaining birth records, frequency matched on birth year (n = 6,072). Exposure data consisted of self-reported maternal prenatal smoking recorded on the birth certificate (yes/no) as determined by check box, smoking dose (average number of cigarettes/day), and trimester of exposure (prepregnancy, first, second, or third trimester; available for the years after 2002). Birth year and maternal and infant variables from the birth certificate were assessed for their possible confounding effects. Infant factors examined included gestational age (<37, 37–40, >40 weeks), birth weight (<2,500, 2,500–3,999, ≥4,000 grams), and sex. Maternal factors evaluated included marital status, age (<20, 20–34, ≥35 years), number of prior pregnancies and deliveries (none, 1, 2, ≥3), cesarean section delivery, and race/ethnicity (white, African American, Hispanic, Asian, American Indian, Hawaiian/Pacific Islander, other). Maternal education (<12, 12, 13–15, ≥16 years), prenatal care inception (no care, first, second, or third trimester), and health insurance billed at delivery (private or Medicaid/Medicare) were evaluated as proxies of socioeconomic status.
Medical records of all 1,518 cases of JIA initially identified by ICD-9 codes were reviewed, and 228 were subsequently excluded because JIA diagnosis was not supported; 22% had other rheumatologic disease (lupus, juvenile dermatomyositis, vasculitis, sarcoidosis, or arthritis onset at age ≥17 years); 3% had inflammatory bowel or celiac disease–related arthritis; 11% had reactive or postinfectious/infection-related arthritis; and the majority (64%) had no arthritis/nonrheumatic diagnoses, including mechanical arthralgias, central pain sensitization, isolated enthesitis, psoriasis, anterior knee pain, and osteochondral defect. The 1,290 verified JIA cases were categorized by ILAR classification as rheumatoid factor (RF)–positive polyarticular JIA, RF-negative polyarticular JIA, systemic JIA, oligoarticular JIA, extended oligoarticular JIA, psoriatic JIA, enthesitis-related arthritis, and undifferentiated JIA (2). Twenty-eight cases without RF titer available were categorized as RF-unknown polyarticular JIA. RF, antinuclear antibody (ANA), anti–cyclic citrullinated peptide (anti-CCP), and HLA–B27 were determined by chart review. After excluding subjects with missing data for maternal prenatal smoking, there were 1,196 JIA cases and 5,618 controls.
Levels of missing data were similar between cases and controls and were generally <1%. For maternal parity, gravidity, race/ethnicity, and gestational length, the proportion of subjects with missing data was <4%. Missing values were more common for trimester of prenatal care inception (7%), maternal education (38%), and health insurance (14% cases, 18% controls), the latter 2 only available since 1992 and 1987, respectively. The 8% of subjects without smoking data had characteristics more similar to smokers than nonsmokers; this did not vary by case/control status.
Unconditional logistic regression was used to calculate odds ratio (OR) and 95% confidence intervals (95% CIs) for the associations of JIA overall and for individual categories in relation to maternal prenatal smoking and smoking characteristics. Maternal age was adjusted for in the analysis, as prior studies have shown that it often confounds relationships with smoking and reproductive outcomes. Factors that altered the risk estimates by >10% were adjusted for in the analyses. Unless otherwise indicated, all risk estimates were adjusted for maternal age and marital status. Statistical analysis was conducted using Stata software, version 12.
Results
Mothers of subjects with JIA were more likely to be older, married, of higher education level, privately insured, and white (Table 1). The majority of cases were female. Other characteristics were similar among cases and controls.
Table 1.
Parental and infant characteristics for children with and without JIA, born in Washington State, 1997–2010*
| Characteristic | JIA cases (n = 1,196) |
Non-JIA controls (n = 5,618) |
|---|---|---|
| Maternal age, years | ||
| <20 | 98 (8.2) | 641 (11.4) |
| 20–34 | 919 (76.8) | 4,289 (76.4) |
| ≥35 | 179 (15.0) | 686 (12.2) |
| Marital status | ||
| Married | 964 (80.7) | 4,124 (73.6) |
| Unmarried | 231 (19.3) | 1,479 (26.4) |
| Number of prior live births | ||
| None | 507 (43.3) | 2,333 (42.1) |
| 1 | 404 (34.5) | 1,781 (32.2) |
| 2 | 160 (13.7) | 890 (16.1) |
| ≥3 | 100 (8.5) | 536 (9.6) |
| Number of prior pregnancies | ||
| None | 373 (31.7) | 1,777 (32.1) |
| 1 | 347 (29.5) | 1,545 (27.9) |
| 2 | 217 (18.5) | 1,040 (18.8) |
| ≥3 | 238 (20.3) | 1,174 (21.2) |
| Maternal education, years | ||
| <12 | 98 (12.7) | 630 (18.1) |
| 12 | 223 (28.9) | 1,073 (30.7) |
| 13–15 | 187 (24.2) | 914 (26.2) |
| ≥16 | 264 (34.2) | 871 (25.0) |
| Number of prior children living | ||
| 0 | 507 (43.3) | 2,354 (42.5) |
| 1 | 408 (34.8) | 1,803 (32.5) |
| 2 | 156 (13.3) | 862 (15.6) |
| ≥3 | 101 (8.6) | 523 (9.4) |
| Trimester prenatal care began | ||
| First | 949 (85.6) | 4,264 (80.6) |
| Second | 128 (11.5) | 841 (15.8) |
| Third or no prenatal care | 32 (2.9) | 188 (3.6) |
| Health insurance† | ||
| Medicaid/Medicare | 263 (25.4) | 1,717 (37.3) |
| Private | 773 (74.6) | 2,892 (62.7) |
| Maternal race/ethnicity | ||
| White | 974 (84.5) | 4,295 (78.1) |
| African American | 19 (1.5) | 216 (3.9) |
| Hispanic | 76 (6.6) | 505 (9.2) |
| Asian | 34 (3.0) | 282 (5.1) |
| American Indian | 29 (2.5) | 116 (2.1) |
| Hawaiian/Pacific Islander | 4 (0.4) | 15 (0.3) |
| Other | 17 (1.5) | 72 (1.3) |
| Cesarean section | ||
| Yes | 262 (21.9) | 1,091 (19.4) |
| No | 933 (78.1) | 4,523 (80.6) |
| Infant birth weight, grams | ||
| <2,500 | 67 (5.6) | 301 (5.3) |
| 2,500–3,999 | 993 (83.2) | 4,545 (81.1) |
| ≥4,000 | 134 (11.2) | 760 (13.6) |
| Infant sex | ||
| Female | 810 (67.7) | 2,684 (47.8) |
| Male | 386 (32.3) | 2,934 (52.2) |
| Infant gestational age, weeks | ||
| <37 | 77 (6.6) | 354 (6.4) |
| 37–42 | 994 (85.1) | 4,600 (83.6) |
| >42 | 97 (8.3) | 550 (10.0) |
Values are the number (percentage). Some variables may not add up to totals due to missing data. JIA = juvenile idiopathic arthritis.
Billed at delivery hospitalization.
Overall JIA
Prenatal smoking was reported less often among mothers of cases (11%) than among controls (17%; OR 0.71 [95% CI 0.58–0.87]) (Table 2). There was a suggestion that this effect may be more marked with higher levels of smoking (OR 0.62 [95% CI 0.38–0.99 for ≥20 cigarettes/day], P<0.05 for trend), although when children with nonsmoking mothers were excluded, no dose effect was observed (relative to 1–9 cigarettes/day, OR for 10–19 cigarettes/day 0.89 [95% CI 0.53–1.48], OR for ≥20 cigarettes/day 0.90 [95% CI 0.50–1.62], P = 0.68 for trend). Among the few subjects (15 exposed cases and 65 exposed controls) with available information, no significantly increased or decreased ORs were observed in relation to duration/timing of maternal smoking during pregnancy (relative to no smoking, OR for smoking only during prepregnancy 2.05 [95% CI 0.77–5.50], OR for smoking in any trimester 0.84 [95% CI 0.39–1.80]). Results were unchanged when subjects with mothers with arthritic or autoimmune diseases were excluded.
Table 2.
Maternal prenatal smoking in relation to JIA among cases and controls born in Washington state, 1997–2010*
| Characteristic | Cases (n = 1,196) |
Controls (n = 5,168) |
OR† | 95% CI |
|---|---|---|---|---|
| Smoked | ||||
| Yes | 134 (11.2) | 946 (16.8) | 0.71 | 0.58–0.87 |
| Cigarettes/day, no. | ||||
| 0 | 900 (89.8) | 3,971 (85.1) | 1.0 | (ref) |
| 1–9 | 35 (3.5) | 238 (5.1) | 0.77 | 0.53–1.11 |
| 10–19 | 32 (3.2) | 245 (5.3) | 0.64 | 0.44–0.94 |
| ≥20 | 20 (2.0) | 151 (3.2) | 0.62 | 0.38–0.99 |
| Smoking duration‡ | ||||
| None | 135 (90.0) | 582 (90.0) | 1.0 | (ref) |
| Prepregnancy only | 6 (4.0) | 14 (2.2) | 2.05 | 0.77–5.50 |
| Any time during pregnancy | 9 (6.0) | 51 (7.9) | 0.84 | 0.39–1.80 |
| Child | ||||
| Male | 386 (11.4) | 2,934 (16.4) | 0.74 | 0.53–1.03 |
| Female | 810 (11.1) | 2,684 (17.4) | 0.69 | 0.54–0.89 |
| Maternal | ||||
| Age at delivery, years | ||||
| <20 | 98 (24.5) | 641 (28.1) | 0.83 | 0.51–1.36 |
| 20–34 | 919 (10.9) | 4,289 (16.3) | 0.68 | 0.54–0.85 |
| ≥35 | 179 (5.6) | 686 (9.6) | 0.66 | 0.33–1.33 |
| Race/ethnicity | ||||
| White | 974 (11.7) | 4,295 (19.1) | 0.67 | 0.54–0.83 |
| African American | 19 (15.8) | 216 (16.7) | 1.02 | 0.27–3.85 |
| Hispanic | 76 (2.6) | 505 (3.4) | 0.78 | 0.17–3.47 |
| Asian | 34 (0.0) | 282 (5.7) | NA | NA |
| American Indian | 29 (31.0) | 116 (32.8) | 0.99 | 0.40–2.46 |
| Hawaiian/Pacific Islander | 4 (25.0) | 15 (46.7) | 0.38 | 0.03–4.55 |
| Other | 17 (0.0) | 72 (1.4) | NA | NA |
| Marital status | ||||
| Unmarried | 231 (28.1) | 1,479 (31.1) | 0.87 | 0.64–1.18 |
| Married | 964 (7.2) | 4,124 (11.6) | 0.64 | 0.49–0.83 |
| Trimester of prenatal care inception | ||||
| First | 949 (9.6) | 4,264 (15.1) | 0.67 | 0.53–0.85 |
| Second | 128 (21.9) | 841 (24.7) | 0.91 | 0.57–1.44 |
| Third or no prenatal care | 32 (25.0) | 188 (27.1) | 0.90 | 0.38–2.14 |
| Health insurance§ | ||||
| Private | 773 (6.5) | 2,892 (11.2) | 0.61 | 0.44–0.83 |
| Medicaid/Medicare | 263 (26.2) | 1,717 (25.0) | 1.07 | 0.79–1.44 |
| Education, years | ||||
| <12 | 98 (24.5) | 630 (24.3) | 0.98 | 0.59–1.61 |
| 12 | 223 (14.8) | 1,073 (17.3) | 0.87 | 0.57–1.31 |
| ≥13 | 451 (3.8) | 1,785 (6.1) | 0.67 | 0.40–1.15 |
Values are the number (percentage) unless otherwise indicated. Numbers may not add to total due to missing data. JIA = juvenile idiopathic arthritis; OR = odds ratio; 95% CI = 95% confidence interval; NA = not applicable.
Adjusted for maternal age and marital status, except maternal age-specific ORs were adjusted for marital status only; marital status-specific ORs were adjusted for maternal age only, and crude OR is shown for Hawaiian/Pacific Islanders.
Trend test P < 0.05 only when nonsmokers are included as 0; P = 0.68 when calculated only among smokers.
Billed at delivery hospitalization.
The inverse JIA–smoking association was slightly more marked in females than males (Table 2). ORs of <1 for the JIA–smoking association were noted in nearly all subgroups examined, except among children of African American women (OR 1.02 [95% CI 0.27–3.85]), or whose birth hospitalizations were billed to Medicaid (OR 1.07 [95% CI 0.79–1.44]). Among subjects without Medicaid (773 cases, 2,892 controls), results stratified by these characteristics were essentially unchanged, whereas among those with Medicaid (263 cases, 1,717 controls), the magnitudes of the ORs varied, with most slightly >1 and all CIs containing 1, the greatest being for the JIA–smoking association among Hispanics (OR 1.39 [95% CI 0.29–6.64], data not shown).
JIA categories
JIA categories consisted of 380 (32%) oligoarticular JIA, 73 (6%) extended oligoarticular JIA, 259 (22%) RF-negative polyarticular JIA, 55 (5%) RF-positive polyarticular JIA, 28 (2%) RF-unknown polyarticular JIA, 239 (20%) enthesitis-related arthritis, 65 (5%) systemic JIA, 61 (5%) psoriatic JIA, and 36 (3%) undifferentiated JIA (Table 3). With few exceptions, the ORs for associations of specific JIA categories with maternal prenatal smoking were <1. Decreased associations were more marked for those diagnosed or with disease onset at ages <10 years, and lowest in the youngest age group (<5 years). Statistically significant decreased ORs were noted only for oligoarticular and extended oligoarticular JIA, but decreased ORs were also noted for most other categories, with the exception of RF-positive polyarticular (OR 1.06 [95% CI 0.53–2.10]) and undifferentiated JIA (OR 1.75 [95% CI 0.78–3.96]). No significantly increased OR was observed in any JIA category in relation to smoking. Neither RF-positive nor anti-CCP–positive status was associated with increased OR. ANA-positive and HLA–B27–positive status were both associated with ORs <1. These results were similar among subjects without Medicaid, but among those with Medicaid (and cell sizes ≥5), ORs ranged from 0.77 (95% CI 0.42–1.41) for diagnosis at ages 5–9 years to 1.62 (95% CI 0.98–2.68) for diagnoses at ages 10–14 (data not shown). JIA category–specific ORs among children with Medicaid ranged from 0.94 (95% CI 0.56–1.58) for oligoarticular JIA (83 cases) to 1.41 (95% CI 0.82–1.42) for RF-negative polyarticular JIA (66 cases).
Table 3.
Maternal prenatal smoking in relation to JIA disease characteristics among cases and controls born in Washington state, 1997–2010*
| Characteristic | No. JIA cases (n = 1,196) |
Prenatal smoke exposure, % |
OR† | 95% CI |
|---|---|---|---|---|
| Age at JIA diagnosis, years | ||||
| <5 | 395 | 8.1 | 0.48 | 0.33–0.69 |
| 5–9 | 271 | 9.2 | 0.57 | 0.37–0.86 |
| 10–14 | 371 | 15.6 | 1.01 | 0.76–1.36 |
| ≥15 | 155 | 12.3 | 0.84 | 0.52–1.34 |
| Age at JIA onset, years | ||||
| <5 | 435 | 8.7 | 0.54 | 0.38–0.76 |
| 5–9 | 323 | 10.5 | 0.63 | 0.43–0.90 |
| 10–14 | 349 | 14.9 | 1.00 | 0.74–1.37 |
| 15–16 | 83 | 12.1 | 0.75 | 0.38–1.47 |
| JIA category | ||||
| Oligoarticular | 380 | 9.5 | 0.59 | 0.41–0.85 |
| Extended oligoarticular | 73 | 2.7 | 0.19 | 0.04–0.76 |
| RF-positive polyarticular | 55 | 20.0 | 1.06 | 0.53–2.10 |
| RF-negative polyarticular | 259 | 12.4 | 0.76 | 0.52–1.12 |
| RF-unknown polyarticular | 28 | 7.1 | 0.37 | 0.09–1.59 |
| Psoriatic arthritis | 61 | 11.5 | 0.76 | 0.34–1.70 |
| Enthesitis-related arthritis | 239 | 12.6 | 0.89 | 0.59–1.32 |
| Systemic arthritis | 65 | 9.2 | 0.52 | 0.22–1.24 |
| Undifferentiated arthritis | 36 | 22.2 | 1.75 | 0.78–3.96 |
| Antinuclear antibody | ||||
| Negative | 458 | 11.6 | 0.73 | 0.54–0.98 |
| Positive | 537 | 11.2 | 0.70 | 0.53–0.93 |
| Unknown | 201 | 10.5 | 0.71 | 0.45–1.11 |
| Rheumatoid factor | ||||
| Negative | 658 | 10.5 | 0.66 | 0.50–0.85 |
| Positive | 77 | 15.6 | 0.81 | 0.43–1.53 |
| Unknown | 458 | 11.6 | 0.79 | 0.59–1.06 |
| Anti–cyclic citrullinated peptide | ||||
| Negative | 183 | 12.0 | 0.73 | 0.46–1.16 |
| Positive | 43 | 14.0 | 0.76 | 0.31–1.86 |
| Unknown | 969 | 10.9 | 0.71 | 0.57–0.88 |
| HLA–B27 | ||||
| Negative | 407 | 12.3 | 0.76 | 0.56–1.04 |
| Positive | 128 | 11.7 | 0.94 | 0.56–1.58 |
| Unknown | 660 | 10.5 | 0.65 | 0.50–0.84 |
Numbers may not add to total due to missing data. JIA = juvenile idiopathic arthritis; OR = odds ratio; 95% CI = 95% confidence interval; RF = rheumatoid factor.
Adjusted for maternal age and marital status.
Differences in maternal smoking levels between JIA cases and controls may have been confounded by underlying socioeconomic status-related differences in behavior patterns with respect to whether or where parents would seek care for children with JIA symptoms. We therefore repeated the analyses after restricting cases to those more likely to have received care due to disease severity or younger age (RF-positive and/or anti-CCP–positive, and diagnosed at ages <10 years), and to subjects born in counties including or immediately adjacent to SCH. Among these 18 cases and 2,966 controls, the OR for the association of JIA with prenatal smoking was 0.73 (95% CI 0.20–2.64, data not shown) and did not vary by Medicaid status, although our ability to examine this was limited due to small numbers.
Discussion
Our data do not suggest an increased risk of JIA in relation to maternal prenatal smoking. This result was unexpected, given the evidence that cigarette smoking increases the risk of RF-positive or anti-CCP–positive RA among adults (3), and suggests that JIA may have different etiologies than adult-onset disease. That we observed no significantly increased associations in any JIA category argues against an increased risk of JIA associated with maternal prenatal smoking. However, the negative associations we observed are not necessarily strong evidence of an inverse relationship, given the presence of other possible explanations. Residual confounding due to socioeconomic status may have resulted in the negative associations we observed overall and for some JIA categories. The lack of an inverse relationship among children whose mothers had Medicaid insurance at the time of delivery supports this premise. Although we attempted to identify all children with JIA at the sole facility with board-certified pediatric rheumatologists in Washington, our JIA cases may not represent all Washington state children with JIA, as some may be treated by adult rheumatologists/practitioners outside SCH, whereas our controls were randomly selected from all children born in the same years.
Consistent with national data (5,6), pregnant women who smoked in our study were more likely to be unmarried, receive late prenatal care, or be Medicaid recipients. These characteristics may alter access to care and/or visits to a regional quaternary care hospital. We explored the possibility of an inverse effect due to the “hygiene hypothesis” (7), but our data demonstrated that lower maternal education, Medicaid status, late prenatal care, and unmarried status were less common among cases than controls, and cases and controls had similar histories of cesarean section delivery, these latter characteristics often assumed to be important in this hypothesis. The effects of maternal smoking have been attributed to mechanical alterations of the placenta and direct placental transfer of potential toxins, including nicotine, carbon monoxide, benzene, tar, and heavy metals (lead and cadmium) (8). Transplacental effects of maternal smoking on the fetus may plausibly differ from direct exposure effects among adult smokers. Furthermore the inverse effect of smoking noted for certain inflammatory diseases (ulcerative colitis and sarcoidosis) might involve nicotine-related immunosuppressive and antiinflammatory properties (9). Similar biologic pathways may have a role in JIA if an inverse relationship truly exists for some categories.
Prior studies of prenatal smoking and JIA using various methods have had mixed results. An Australian case–control study also demonstrated an inverse relationship of JIA with maternal (OR 0.56 [95% CI 0.37–0.85]) and paternal (OR 0.61 [95% CI 0.41–0.91]) smoking during the mother’s pregnancy, as self-reported by parents after the child’s diagnosis. Adjusting for maternal socioeconomic status and age at child’s birth, or for child’s age, sex, and white ancestry mitigated this effect for maternal but not paternal smoking (10), suggesting that the relationship of JIA with transplacental smoking exposure may be complex. Results of studies based on birth record–reported prenatal smoking prior to JIA diagnosis are conflicting. A Finnish cohort study of 58,841 singletons followed until age 7 years reported increased risk of JIA with maternal prenatal smoking (n = 31 JIA cases) restricted to girls (OR 6.97 [95% CI 2.00–22.9]) (11). Our study included a larger number and broader representation of JIA cases (including evaluation of JIA categories), and we did not observe increased ORs for either sex. A Swedish nationwide register–based case–control study with 3,334 JIA cases reported no association of JIA with maternal prenatal smoking (OR 1.0 [95% CI 0.8–1.1]) (12), although sex-specific ORs were not reported.
Using cases identified from outpatient/inpatient medical records maintained by pediatric rheumatologists with linkage to a state birth-registry database facilitated identification of a large number of cases for a disease with relatively low incidence. Other strengths include exposure information recorded prior to, and independently of, the child’s JIA diagnosis, reducing recall bias. Medical chart review allowed us to refine ICD-9 codes to exclude non-JIA cases that were originally misclassified and to examine associations for specific JIA categories.
Prenatal smoking, a socially undesirable trait, is likely underreported, but if accuracy of reporting does not vary by case/control status, this underreporting would have no effect on our results. Pregnancy smoking prevalence in Washington state assessed by the Pregnancy Risk Assessment Monitoring System and birth certificate data from 2000–2005 did not differ greatly and ranged from 9–12% (5,13). A study comparing cotinine levels in specimens from newborn infants of smokers and nonsmokers identified in Washington birth certificates suggests this variable is fairly accurate (14). However, if underreporting of smoking is more prevalent in lower income women, and if these women are also underrepresented among our JIA cases, the underreporting could have biased our results.
Mothers who smoked during pregnancy probably continued to smoke after delivery. A limitation of the study is that we could not assess smoke exposure following delivery, paternal smoking, passive exposure to environmental tobacco smoke during childhood, or direct smoking exposure among adolescents. Our control group may contain some JIA cases, as children may have moved out of state and been diagnosed with JIA elsewhere; this change would result in attenuation of our risk estimates. Approximately 38% of the data are missing for maternal education because this variable is available on birth certificates for children born from 1992 on, but the absence of data does not differ by case/control status and is unlikely to have affected our results.
Environmental risk factors are more amenable to interventions, so understanding how smoking impacts disease onset is important. In contrast to what would be expected based on adult studies, we did not observe an increased risk of maternal prenatal smoking with JIA or its categories. The inverse risk we observed was not apparent for all categories and is likely due to residual confounding by socioeconomic status, given the lack of association among those with Medicaid health insurance. If such an inverse risk does indeed exist, it may be restricted to children with oligoarticular and extended oligoarticular JIA. Future population-based studies with the ability to categorize specific JIA categories, and using more sensitive exposure metrics (like cotinine levels assayed from cord blood or newborn screening archives) (15), would help elucidate potential biologic mechanisms.
Significance & Innovations.
The use of linked birth certificate/medical records data for juvenile idiopathic arthritis (JIA) cases treated by pediatric rheumatologists at a major facility is unique and enabled us to obtain maternal prenatal smoking exposure information for a large number of JIA cases.
In contrast to what would be expected based on adult studies, we did not observe an increased risk of JIA in relation to maternal prenatal smoking.
Acknowledgments
The authors thank Bill O’Brien of the University of Washington for his programming assistance, and the Washington State Department of Health for data access.
Dr. Shenoi’s work was supported by the NIH (grant T32-AR-007108).
Dr. Wallace has received consulting fees, speaking fees, and/or honoraria (less than $10,000 each) from Amgen and Novartis, and research grants from Amgen and Pfizer.
Footnotes
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Shenoi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Shenoi, Wallace, Mueller.
Acquisition of data. Shenoi, Bell, Mueller.
Analysis and interpretation of data. Shenoi, Bell, Wallace, Mueller.
References
- 1.Towner SR, Michet CJ, Jr, O’Fallon WM, Nelson AM. The epidemiology of juvenile arthritis in Rochester, Minnesota 1960–1979. Arthritis Rheum. 1983;26:1208–13. doi: 10.1002/art.1780261006. [DOI] [PubMed] [Google Scholar]
- 2.Petty RE, Southwood TR, Manners P, Baum J, Glass DN, Goldenberg J, et al. International League of Associations for Rheumatology classification of juvenile idiopathic arthritis: second revision, Edmonton, 2001. J Rheumatol. 2004;31:390–2. [PubMed] [Google Scholar]
- 3.Sugiyama D, Nishimura K, Tamaki K, Tsuji G, Nakazawa T, Morinobu A, et al. Impact of smoking as a risk factor for developing rheumatoid arthritis: a meta-analysis of observational studies. Ann Rheum Dis. 2010;69:70–81. doi: 10.1136/ard.2008.096487. [DOI] [PubMed] [Google Scholar]
- 4.Knopik VS, Maccani MA, Francazio S, McGeary JE. The epigenetics of maternal cigarette smoking during pregnancy and effects on child development. Dev Psychopathol. 2012;24:1377–90. doi: 10.1017/S0954579412000776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tong VT, Jones JR, Dietz PM, D’Angelo D, Bombard JM Centers for Disease Control and Prevention. Trends in smoking before, during, and after pregnancy: Pregnancy Risk Assessment Monitoring System (PRAMS), United States, 31 sites, 2000–2005. MMWR Surveill Summ. 2009;58:1–29. [PubMed] [Google Scholar]
- 6.Tong VT, Dietz PM, Morrow B, D’Angelo DV, Farr SL, Rockhill KM, et al. Trends in smoking before, during, and after pregnancy: Pregnancy Risk Assessment Monitoring System, United States, 40 sites, 2000–2010. MMWR Surveill Summ. 2013;62:1–19. [PubMed] [Google Scholar]
- 7.Rook GA. Hygiene hypothesis and autoimmune diseases. Clin Rev Allergy Immunol. 2012;42:5–15. doi: 10.1007/s12016-011-8285-8. [DOI] [PubMed] [Google Scholar]
- 8.Shea AK, Steiner M. Cigarette smoking during pregnancy. Nicotine Tob Res. 2008;10:267–78. doi: 10.1080/14622200701825908. [DOI] [PubMed] [Google Scholar]
- 9.Sopori M. Effects of cigarette smoke on the immune system. Nat Rev Immunol. 2002;2:372–7. doi: 10.1038/nri803. [DOI] [PubMed] [Google Scholar]
- 10.Ellis JA, Ponsonby AL, Pezic A, Chavez RA, Allen RC, Akikusa JD, et al. CLARITY: childhood arthritis risk factor identification study. Pediatr Rheumatol Online J. 2012;10:37. doi: 10.1186/1546-0096-10-37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jaakkola JJ, Gissler M. Maternal smoking in pregnancy as a determinant of rheumatoid arthritis and other inflammatory polyarthropathies during the first 7 years of life. Int J Epidemiol. 2005;34:664–71. doi: 10.1093/ije/dyi006. [DOI] [PubMed] [Google Scholar]
- 12.Carlens C, Jacobsson L, Brandt L, Cnattingius S, Stephansson O, Askling J. Perinatal characteristics, early life infections and later risk of rheumatoid arthritis and juvenile idiopathic arthritis. Ann Rheum Dis. 2009;68:1159–64. doi: 10.1136/ard.2008.089342. [DOI] [PubMed] [Google Scholar]
- 13.Allen AM, Dietz PM, Tong VT, England L, Prince CB. Prenatal smoking prevalence ascertained from two population-based data sources: birth certificates and PRAMS questionnaires, 2004. Public Health Rep. 2008;123:586–92. doi: 10.1177/003335490812300508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Searles Nielsen S, Dills RL, Glass M, Mueller BA. Accuracy of prenatal smoking data from Washington State birth certificates in a population-based sample with cotinine measurements. Ann Epidemiol. 2014;24:236–9. doi: 10.1016/j.annepidem.2013.12.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yang J, Pearl M, Jacob P, III, Delorenze GN, Benowitz NL, Yu L, et al. Levels of cotinine in dried blood specimens from newborns as a biomarker of maternal smoking close to the time of delivery. Am J Epidemiol. 2013;178:1648–54. doi: 10.1093/aje/kwt182. [DOI] [PMC free article] [PubMed] [Google Scholar]