Little is known about how prenatal maternal stress (PNMS) influences risks of asthma in humans. In this small study, we sought to determine whether disaster-related PNMS would predict asthma risk in children. In June 1998, we assessed severity of objective hardship and subjective distress in women pregnant during the January 1998 Quebec Ice Storm. Lifetime asthma symptoms, diagnoses, and corticosteroid utilization were assessed when the children were 12 years old (
Currently affecting 235 million people worldwide, asthma is the most common chronic disease among children [
The biological mechanisms by which prenatal maternal stress (PNMS) affects the fetal immune system are unclear and may differ based on the type, duration, frequency, and severity of the stressor and the sex of the child [
Strong evidence supports the notion that PNMS has different programming effects on male and female animals [
The process by which PNMS differentially affects girls and boys can be mediated by the placenta. In a review, Clifton [
To our knowledge, there are no published studies examining whether sex moderates the association between PNMS and asthma risk in humans. Yet, sex is recognized as an important factor for lung development and asthma [
Past studies on the effect of prenatal maternal anxiety/stress have several limitations. While animal studies are capable of randomly assigning stressors to pregnant dams, their results are not directly applicable to humans [
Natural disasters provide research opportunities with a number of methodological advantages. Disasters randomize the distribution of objective hardship, provide greater standardization of the nature and timing of the stressor, and allow the researcher to tease apart the effects of objective hardship and subjective distress.
In January 1998, 100 mm of freezing rain in the province of Quebec led to the most severe ice storm and one of the worst natural disasters, in Canada’s history [
Sixty-eight mothers from our Project Ice Storm cohort [
Questions used to assess the four dimensions (Threat, Loss, Scope, and Change) of objective prenatal maternal stress in the mothers after the ice storm.
Threat | Loss | Scope | Change |
---|---|---|---|
(1) Were you injured? |
(1) Did your residence suffer damage as a result of the ice storm? |
(1) How many days were you without electricity? |
(1) Did your family stay together for the duration of the ice storm? |
(2) Was anyone close to you injured? |
(2) Did you experience a loss of personal income? |
(2) How many days were you without the use of your telephone? |
(2) Did you spend any time in a temporary shelter? |
(3) Were you ever in danger due to |
(3) How much was the total financial loss including income, food, and damage to home? |
(3) How often were you required to change residence during the ice storm? | |
(3.2) …exposure to downed electrical power lines |
(4) Did you take in guests during the ice storm? | ||
(3.3) …exposure to carbon monoxide |
(5) Did you experience an increase in physical work during the ice storm? | ||
(3.5) …lack of food |
(6) Number of nights away from home: | ||
(3.6) …falling branches and ice |
|||
|
|||
8 points | 8 points | 8 points | 8 points |
Information concerning maternal life events was obtained when the children were 6 months and 11.5 years of age using a 29-item Life Experiences Survey (LES) [
The 28-item General Health Questionnaire (GHQ-28) [
The Douglas Hospital Research Ethics Board approved all phases of Project Ice Storm. This pilot phase was approved on June 2012 (Protocol 10/23). Women pregnant during the ice storm were recruited in June 1998 by postal questionnaires. The first questionnaire, assessing the hardship and enduring reactions to the storm, the GHQ, and family demographics, was mailed to 1440 women by their obstetricians. A total of 224 women responded to this initial questionnaire, with 178 giving permission for further contact. Six months after each woman’s due date, a second questionnaire assessing pregnancy outcomes was mailed. Of these, 177 women returned the second questionnaire. A number of subjects reported miscarriages or stillbirths, and many families were lost to follow-up during a gap in funding in the early years of the study. Since then, Project Ice Storm families have been assessed up to 17 times, including an assessment at age 11.5 years (
First, we conducted
Child factors at birth (i.e., weight and gestational age) were allowed to enter into the equation during Block 1, in stepwise fashion. Parental factors (i.e., SES, obstetric complications, maternal anxiety (at 6 months and 11.5 years), maternal life events (6 months and 11.5 years), parental history of asthma, and number of cigarettes the mother smoked per day during pregnancy) were allowed to enter into the equation during Block 2, also by stepwise. The children’s sex was forced into the equation during Block 3. Timing of the exposure to the stress was allowed to enter into the equation Block 4, in stepwise fashion. Objective hardship was forced into the equation during Block 5. Subjective distress was forced into the equation during Block 6. The interactions terms of objective hardship × sex of the child and subjective distress × sex of the child were entered in Block 7, in a stepwise fashion. Predictors that were not significantly related to the outcome measures were trimmed from the model due to the relatively small sample size and the analyses were reconducted. Hosmer-Lemeshow tests were conducted to provide information on the goodness of fit of the models.
A comparison between the families who completed the asthma questionnaire versus families who responded to the June 1998 questionnaire but who did not answer the asthma questionnaire indicated that participating families had higher socioeconomic status levels (see Table
Comparison of characteristics between study responders and nonresponders.
Characteristics | Responders ( |
Nonresponders ( |
|
||
---|---|---|---|---|---|
Mean | SD | Mean | SD | ||
Anxiety | 2.26 | 2.02 | 2.36 | 2.31 | 0.780 |
Total psychiatric symptoms | 5.78 | 4.71 | 6.99 | 6.51 | 0.160 |
Subjective stress | 11.02 | 12.29 | 12.01 | 12.82 | 0.620 |
Objective stress | 11.71 | 3.79 | 10.88 | 4.52 | 0.200 |
Socioeconomic status | 25.90 | 11.94 | 31.68 | 12.64 | 0.003 |
Descriptive statistics for the categorical variables (Table
Chi-squared tests providing descriptive statistics (
Characteristics | Wheezing | Doctor-diagnosed asthma | Inhaled corticosteroids consumption | |||
---|---|---|---|---|---|---|
Yes (%) |
No (%) |
Yes (%) |
No (%) |
Yes (%) |
No (%) |
|
|
||||||
Parental asthma | ||||||
Neither parent | 23 (53) | 20 (47) | 8 (19) | 35 (81) | 16 (37) | 27 (63) |
At least one parent | 14 (56) | 11 (44) | 6 (24) | 19 (76) | 9 (27) | 16 (73) |
Chi2 (1) | 0.04, |
0.28, |
0.01, |
|||
|
||||||
Sex | ||||||
Boy | 20 (63) | 12 (38) | 8 (25) | 24 (75) | 14 (44) | 18 (56) |
Girl | 17 (47) | 19 (53) | 6 (17) | 30 (83) | 11 (31) | 25 (69) |
Chi2 (1) | 1.59, |
0.72, |
1.27, |
Variables | Wheezing in the chest | Doctor-diagnosed asthma | Inhaled corticosteroids consumption | ||||||
---|---|---|---|---|---|---|---|---|---|
Yes |
No |
|
Yes |
No |
|
Yes |
No |
|
|
Socioeconomic status |
|
|
− |
26.36 |
25.78 |
−0.16 |
26.16 |
25.74 |
−0.14 |
Maternal anxiety |
2.27 |
2.26 |
−0.03 |
2.71 |
2.15 |
−0.93 |
2.52 |
2.12 |
−0.79 |
Maternal anxiety |
1.13 |
0.69 |
−1.12 |
1.36 |
0.83 |
−1.06 |
1.15 |
0.81 |
−0.81 |
Life events (6 months) |
|
|
− |
7.14 |
5.13 |
−1.64 |
|
|
− |
Life events (11.5 years) | 2.19 |
1.46 |
−1.67 |
|
|
− |
|
|
− |
Obstetric Complications | 5.00 |
4.26 |
−1.02 |
5.43 |
4.46 |
−1.08 |
|
|
− |
Child’s birth weight (grams) | 3345.95 (547.80) | 3366.52 (601.24) | 0.15 |
3313.87 |
3366.07 |
0.30 |
3362.77 |
3350.10 |
−0.08 |
Child’s gestational age (weeks) | 39.39 |
39.42 |
0.06 |
39.29 |
39.44 |
0.26 |
39.47 |
39.37 |
−0.20 |
Timing of stress (days) |
143.35 (89.45) | 134.19 (76.00) | −0.45 |
168.21 |
131.65 |
−1.48 |
152.36 |
131.51 |
−1.00 |
Objective exposure | 11.38 |
10.87 |
−0.49 |
12.00 |
10.93 |
−0.84 |
11.28 |
11.07 |
−0.20 |
Subjective distress | 13.07 |
8.57 |
−1.52 |
13.00 |
10.50 |
−0.68 |
13.46 |
9.60 |
−1.25 |
The final hierarchical logistic regression models for each outcome variable, after trimming of nonsignificant predictors, are presented in Table
Final logistic regression models predicting asthma-related outcomes.
Models |
|
S.E. | Exp( |
90% CI for Exp( | |
---|---|---|---|---|---|
|
|||||
Constant | −0.126 | 0.716 | 0.882 | ||
Life events (11.5 years) | 0.369* | 0.227 | 1.447 | 0.997 | 2.100 |
Sex (0 = male) | −1.587* | 0.918 | 0.204 | 0.045 | 0.926 |
Subjective distress | −0.030 | 0.038 | 0.971 | 0.912 | 1.033 |
Subjective distress × sex | 0.137* | 0.072 | 1.147 | 1.019 | 1.291 |
|
|||||
|
|||||
Constant | −1.655* | 0.891 | 0.191 | ||
Life events (6 months) | 0.237* | 0.121 | 1.267 | 1.039 | 1.545 |
Sex (0 = male) | −2.270* | 1.181 | 0.103 | 0.015 | 0.721 |
Subjective distress | −0.126* | 0.072 | 0.881 | 0.783 | 0.992 |
Subjective distress × sex | 0.214** | 0.088 | 1.238 | 1.071 | 1.431 |
|
|||||
|
|||||
Constant | −0.999 | 0.823 | 0.368 | ||
Life events (11.5 years) | 0.769** | 0.290 | 2.158 | 1.338 | 3.479 |
Parental asthma history | −1.661* | 0.934 | 0.190 | 0.041 | 0.883 |
Sex (0 = male) | −1.759 | 1.103 | 0.172 | 0.028 | 1.058 |
Subjective distress | −0.057 | 0.047 | 0.945 | 0.874 | 1.021 |
Subjective distress × sex | 0.173** | 0.083 | 1.188 | 1.037 | 1.362 |
Procedure: analyses were conducted on the entire sample (
The available data also allowed us to analyze whether PNMS was associated with maternal-reported asthma. Six mothers reported that their child has asthma, although they did not report a doctor’s diagnosis. In general, results for mother-reported asthma were similar to those for doctor-diagnosed asthma (results not shown).
Results of our pilot study are the first to suggest that disaster-related maternal subjective distress interacts with the child’s sex to influence asthma risk. In girls only, higher levels of subjective maternal distress in pregnancy were associated with increased lifetime risk for wheezing, doctor-diagnosed asthma, and inhaled corticosteroid usage. Perinatal and current maternal life events were also associated with increased risk for asthma-related outcomes. Objective hardship and timing of the ice storm during pregnancy did not predict asthma-related outcomes in the final models with this sample, nor did they interact significantly with other predictors.
These results suggest that fetal immune function appears to have been influenced, in girls, by the mothers’ posttraumatic stress-type symptoms from the ice storm rather than by what happened to them objectively, above and beyond the effects of other perinatal and current life events experienced by the mothers. Although further research is required, our findings support previous evidence suggesting that maternal distress during pregnancy opens a window for fetal programming of the immune system.
Although population rates of asthma during childhood naturally tend to be higher in boys than girls, results showed that PNMS was associated with the risk for asthma in girls only. Past studies have found that PNMS can masculinise some aspects of female development [
A comprehensive understanding of the etiology of immune disorders in early childhood is necessary for effective primary prevention. In a study on developmental programming, Pincus and colleagues [
Our results also support the hypothesis that maternal life events during the perinatal and childhood periods are associated with an increased risk for asthma. Similarly, a twin study found that parental stress increased asthma morbidity in 1- and 3-year-old offspring [
The present study has limitations. First, the observational design used in this study does not allow us to infer causation. Results only suggest that associations exist between PNMS exposure and asthma outcomes. Second, we did not have access to data on household pets or current parental smoking. Third, asthma symptoms may be under- or overreported by some mothers. Fourth, the sample size is small and thus can lead to an overestimation of the odds ratios. In addition, we did not have data on pubertal status of the subjects. Finally, replication of the study with a larger and more representative sample is required. The sample at baseline was not representative of the larger catchment area. Participants had higher SES scores than the regional averages and than nonresponders from the larger cohort. However, it is possible that the higher SES scores minimized the effects of PNMS rather than enhance it.
It is also worth noting that factors commonly related to wheezing and asthma in children were not found in the univariate analysis (e.g., parental asthma, gender). One explanation is that there may be different pathways by which asthma develops, as implied by the principle of equifinality, and that PNMS exposure weakens the genetic effects suggested by parental asthma; PNMS may be responsible for sporadic (nongenetic) cases of asthma. PNMS may also have altered the usual trends in sex differences, as noted above.
Despite the limitations, results of the present pilot study contribute to the PNMS literature. To our knowledge, this is the first prospective study of the effects of PNMS on asthma outcomes that assessed an independent, naturally randomized stressor. Moreover, the offspring’s age in this sample increases the likelihood of accurate lifetime diagnoses for asthma compared to younger samples examined in other studies [
We provided evidence that disaster-related PNMS may be associated with asthma outcomes in a sex-specific manner. Findings highlight the complexity of the association between PNMS and the fetal programming of chronic immune disorders. There is a need to produce evidence-based knowledge on the biological mechanisms by which a mother’s subjective distress in the face of a natural disaster affects pre- and postnatal immune functioning over time. We suggest some directions for future research as follows.
What roles do CG, sex hormones, and the placenta play in the development of immune disorders for prenatally stressed girls? Does PNMS cause fetal methylation of genes expressed in the lungs or lymphoid tissue that are involved in immune response function? Does PNMS induce asthma, does it exacerbate the disease, or does it speed up its progression? How do the biological mechanisms associated with objective hardship versus subjective distress differ? Do the effects of PNMS on immune disorders change over the lifespan? What interventions could prevent the effects of PNMS on immune disorders?
A multidisciplinary and longitudinal approach is necessary to complement our results and provide additional insights to prevent childhood immune disorders.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This study was funded by the Stairs Memorial Fund of McGill University, by the Canadian Institutes of Health Research (MOP-57849, MOP-79424, MOP-93660, and MOP-111177, (Suzanne King, the Principal Investigator, and David P. Laplante)), and NIH ES 017588 (Lester Kobzik and Robert Lim). Anne-Marie Turcotte-Tremblay is funded by the University of Montreal Hospital Research Center and the University of Montreal Public Health Research Institute. The authors are grateful for the contribution of Audrey Yelle-Béland, Yuqing Huang, and Hoi Hing Lung for their assistance in collecting the asthma data. In addition, the authors would like to thank Kelsey N. Dancause, Cathy Vaillancourt, Franz Veru, and Chunbo Yu for their comments on this paper.