Our findings indicate that the duration of residence of immigrants in urban Ontario was independently associated with increases in PTB, but not in SGA. Recent immigrants were at lower risk of PTB compared with a mostly Canadian-born population, but immigrants became at higher risk after 10 years of stay in Canada. There was no strong evidence indicating that the influence of duration of residence on birth outcomes varied with the maternal region of birth or with preterm subgroups.
Our study had some limitations. Non-matched immigrants who stayed in Ontario after arrival were counted as non-immigrants, thus diluting the differences in outcomes between these two groups. Because data collection on immigrants started in 1985, we could not identify those immigrants who obtained their permanent residency before 1985. This misclassification, however, would bias our results against significance, because false non-immigrants were immigrants with at least 17 years of residence, and so were at a higher risk of PTB. Because immigration data were linked until 31 December 2000, we had to exclude all newcomers to Ontario after that date to avoid misclassification by immigrant status. This resulted in some under-representation of recent immigrants in our study population. Although the immigration data were of good quality, some variables were measured on arrival and not at delivery, which is a concern for time-dependent variables. Marital status and maternal education may have changed for some women, especially for young women on arrival, who may have got married and become more educated with a longer duration of residence. As educational attainment is not reversible, the bias resulting from adjustment for an underestimate of the true educational attainment of long-term immigrants would have reduced the significance of our results, because higher education protects against adverse birth outcomes. To reduce this bias, we repeated analyses restricted to mothers aged 20 years or older on arrival, but the results did not substantially change.
The use of a Canadian standard for SGA34 may not be appropriate for some ethnic groups (e.g. Asians), for which lower birthweight for gestational age is likely to reflect physiological rather than pathological differences.38,39 However, even if the use of a non ethnic-specific standard overestimates the association between immigrant status and SGA, it is less likely to affect the association between duration of residence and SGA, as the measurement of SGA was independent of duration of residence.
Finally, we had no data on some important predictors of the outcomes, such as tobacco smoking, alcohol consumption, and maternal height, weight, or body mass index (BMI). However, these are mediators rather than confounders of the association between duration of residence and birth outcomes.
Despite these limitations, the association between duration of residence and PTB was quite robust, remaining unaltered across a range of sensitivity analyses. Our main finding regarding a linear increase in PTB with time spent in Canada is consistent with previous studies on Mexicans showing increases in PTB after 2 years of residence in Washington,18 and after 5 years of residence in California.19 Unlike the study that found that Finns lowered their risk of low birthweight and PTB after 3 years of stay in Sweden, but that Sub-Saharan Africans did not,20 we did not find evidence that the association between duration of residence and PTB varied by the maternal world region of origin. Although the increase in PTB with time spent in Canada affected all women, it is noteworthy that a significant number of immigrant women were children or adolescents on arrival. This suggests that adolescent immigrants may constitute a vulnerable cohort, because this is the group that accounts for most deliveries 10–20 years later.
The increase in PTB rates observed in the USA and Canada during the last two decades has been mainly attributed to increased obstetric interventions near term to prevent neonatal complications.3,40 We did not find evidence that this trend observed in the whole population during the last decade was mirrored among immigrants after arrival, suggesting that a ‘population health’ perspective41 may be more appropriate than medical care patterns to explain changes in PTB with time spent in Canada.
Our findings do not support the ‘convergence hypothesis’, which seems to hold for increased BMI/obesity13,14 and behavioural risk factors,12,14,15 and is not consistent for mortality.21 Recent immigrants were at lower risk of PTB, but lost their advantage after approximately 10 years. Instead of remaining at that level, as predicted by the hypothesis, immigrants experienced a continuous deterioration that placed them at a disadvantage after 10 years of stay, compared with non-immigrants. Moreover, because SGA was higher among recent immigrants than among non-immigrants, convergence would have predicted a decrease over time, but duration of residence had no visible impact on this outcome. The definition of the comparison groups and the length of the observation period may impact on the study conclusions. Our study is unique in its ability to measure durations of residence of over 20 years. If our data had been limited to immigrants with less than 10 years of stay we would have erroneously concluded there was convergence in PTB, and missed the observed ‘overshoot’ after 10 years of stay.
Regarding the frequently debated epidemiologic paradox of low birthweight and PTB and the healthy migrant effect, it is noteworthy that these two hypotheses have been largely discussed in the literature in the absence of information on duration of residence. Our findings on PTB suggest that these two phenomena may apply only to recent immigrants. Although studies reporting the epidemiologic paradox in PTB could not distinguish between recent and long-term immigrants, it is likely that the results were driven by the healthier recent immigrants, as most immigrant women have their babies within 10 years of migration. Recent immigrants may enjoy better health than the native-born women soon after arrival, but lose the short-term benefits of selective migration with cumulative exposure to the new physical and social environment, experiencing a sort of ‘regression to the mean’.42
The lack of an association between duration and SGA births in our study suggests that the influence of duration of residence on perinatal health may be outcome-specific. Although our data were not detailed enough to elucidate mechanisms explaining the deterioration of PTB among immigrants, we can advance some hypotheses. One candidate explanation is acculturation, which has already been linked with adverse birth outcomes.7,8,10,43 This suggests that the effect of duration of residence might be explained by changes in health behaviours and related risk factors. Indeed, duration of residence has been used as a proxy for acculturation, and has been associated with increases in BMI/obesity, smoking, alcohol consumption, and physical inactivity,12–15,44,45 factors that may negatively affect gestational age. The prevalence of these factors among immigrants to Canada has been reported to increase with longer stays in Canada.46–49 Whereas high pre-pregnancy BMI has been associated with increases in at least one PTB subtype in some studies,50–53 it has been more consistently found to be protective for SGA.50,51,54,55 It is reasonable to speculate that in our immigrant population increases in BMI might have compensated for the detrimental effect of changes in other risk factors affecting gestational age, thus rendering a null association between duration of residence and SGA.
Another potential pathway leading to an association between duration of residence and PTB, but not SGA, albeit unexplored among immigrant women, may involve psychosocial factors. A few studies found that different measures of maternal stress were associated with PTB, but not with intrauterine growth restriction.56–58 Working conditions, such as long working hours, prolonged standing, and physically demanding work has also been associated with PTB.59–61 Psychosocial exposures such as job strain or low job control and satisfaction may be concomitant causes of PTB among immigrant women,59,60,62–64 who are more likely to be employed in manual, clerical, and shift jobs, and in jobs requiring less than their education level.65
Previous studies have described increases in PTB rates among Mexicans with prolonged residency in the USA. This study was conducted on an ethnically diverse immigrant population, and the association between duration and PTB did not vary according to the world region of maternal origin. This suggests that not only Hispanics but all immigrants were equally affected by prolonged residency in urban Ontario. However, our findings cannot be generalised to rural areas, urban settings with distinctive immigrant groups, or to other health outcomes, without additional empirical research.