Birth Spacing—The Long and Short of It

Worldwide, more than 4 million children die in the first 4 weeks of life, and 28% of these deaths are thought to be due directly to preterm birth.^{1 - 2} Reducing child mortality by 2015 is one of 8 goals in the United Nations Millennium Declaration, adopted by more than 180 nations.³ Birth spacing is mentioned in the strategies set out to achieve the goals but its potential seems underemphasized.

In this issue of JAMA, Conde-Agudelo and colleagues⁴ report the results of a systematic review and meta-analysis of observational studies investigating the association between interpregnancy interval and untoward perinatal health events that are entwined with neonatal mortality.

The authors amassed an impressive amount of cross-cultural evidence to firmly establish the J-shaped relationship between risk of adverse reproductive outcomes and interpregnancy intervals that are either short or long. A total of 67 studies including more than 11 million pregnancies met the strict inclusion criteria for the review. Twenty studies were from the United States, and the remaining 47 were from 61 countries, including research from Latin America, Asia, Africa, Europe, and Australia. As shown in the Figure in the article, the meta-regression curves for preterm birth, low birth weight, and small for gestational age showed lowest risks between approximately 20 and 60 months after delivery of the previous child and those for fetal death and early neonatal death showed lowest risks between approximately 20 and 40 months after delivery.

Twenty-six of the studies fulfilled more rigorous criteria that were required for the meta-analyses, with the number of studies providing data for the individual outcomes ranging from 4 for early neonatal death to 16 for preterm birth. These criteria included using interpregnancy interval instead of birth interval, using at least 4 cut points for intervals, and reporting odds ratios or relative risks or the data to calculate them. Smaller groups of studies were used to compute pooled unadjusted and adjusted estimates of the associations between interpregnancy interval and risks for preterm birth, low birth weight, and small for gestational age. For intervals less than 6 months, the pooled unadjusted odds ratios for these outcomes were 1.77 (1.40 adjusted) for preterm birth, 2.12 (1.61 adjusted) for low birth weight, and 1.39 (1.26 adjusted) for small for gestational age. The results of the meta-analyses provide estimates of risk for the entire range of intervals of birth spacing.

Reservations about the validity of these pooled estimates arise from the troubling amount of heterogeneity that was present. Although the authors attempted to account for some of the potential sources of heterogeneity, one important source was not adequately explored—the country where each study was conducted. The authors looked at economic development status but not at individual countries because several of the studies were multinational. However, only 4 of the 26 included studies involved multiple countries in the same study. For these studies, an indicator for multiple countries could have been used along with country indicators.

Breastfeeding is an important example of a potential confounder that differs greatly by country. Exclusive breastfeeding for the first 6 months of life delays the mother's resumption of menstruation.⁵ Estimates of exclusive breastfeeding in the first 6 months vary dramatically by ethnic group and by country. For example, in the United States, by 6 months of age only 8% of infants are breastfed exclusively, and that practice is about half as common in non-Hispanic blacks.⁶ In Africa, Asia, and Latin America and the Caribbean, Lauer et al⁷ estimated proportions of exclusive breastfeeding in infants younger than 6 months of age at 24.9%, 44.9%, and 30.8%, respectively. Even within regions of continents, much variation exists. For example, the proportion of women in Western and Eastern Africa who exclusively breastfeed for 6 months varies from 6.1% to 41.4%.⁷

Although listed as an important confounder, breastfeeding does not appear to have been controlled for in the studies used in the meta-analyses, probably because most of them were based on record review rather than interviews. These records do not contain enough or accurate information on breastfeeding. Another detail missing from the review is whether the studies' measurement of small for gestational age used appropriate race- sex-, and parity-specific growth curves.⁸

Several causal mechanisms have emerged to explain the risk associated with shorter intervals, including postpartum hormone imbalances, maternal stress, and, most plausibly, maternal nutritional depletion.⁹ Smits and Essed⁹ postulated that depletion of maternal stores of folate during and after pregnancy is responsible for adverse pregnancy outcomes, especially for women who breastfeed or begin their reproductive years with a deficit of folate due to inadequate nutrition. As support for their theory, they cite the finding that neural tube defects are twice as common among children conceived within 6 months following a previous live birth compared with those conceived after 12 to 24 months. From this proposed mechanism flow several hypotheses testable in observational studies, as well as several possible interventions. For example, women who breastfeed and do not take folic acid supplements should be at increased risk. Folate supplementation during the interpregnancy interval should reduce the risk associated with short intervals. It might be beneficial for women who present for prenatal care after a short interpregnancy interval to take folate supplements.

What determines reduced birth spacing? Short interpregnancy intervals are for the most part probably unintended, as are half of all US pregnancies.¹⁰ Short interpregnancy intervals belong to a constellation of interrelated behaviors that ultimately are tied to a complex web of health and economic development. Extremes of the age distribution, marital status, ethnicity, menstrual irregularities, higher parity, and markers of lower socioeconomic status have been shown to be associated with short interpregnancy intervals.^{11 - 12}

Interventions that increase birth spacing could positively affect both maternal and child health. For example, improving the proportion of women who breastfeed exclusively for the first 6 months of their infant's life might lengthen the birth interval while providing better nutrition for the first infant and better health outcomes for both infants.¹³ The mother would reduce her risk for maternal death, puerperal endometritis, and anemia,¹⁴ especially if she replenished her folate stores. Also, women should have access to family planning methods to control the interval length.

The risks associated with long interpregnancy intervals are more difficult to explain. The association with untoward reproductive outcomes might be mediated through preeclampsia and eclampsia.^{14 - 15} Risk of preeclampsia has been observed to increase significantly with interpregnancy interval, with an estimated odds ratio of 1.16 per additional year after pregnancy.¹⁵ Preeclampsia is a primary and significant cause of preterm delivery. Long intervals most likely are not chosen but may result from the end of a partnership, infertility, reproductive losses in the interval, health problems in mother or infant, or economic issues.

The implications of optimizing interpregnancy intervals are not limited to regions that account for a high proportion of neonatal deaths (Southeast Asia, 36%; Africa, 28%; Eastern Mediterranean, 15%; Western Pacific, 13%).² Conde-Agudelo et al¹⁶ estimated that 47 289 perinatal deaths (12.1% of all perinatal deaths) might be avoided in Latin America if the interpregnancy interval were shifted to 12 to 59 months. In high-income countries, increased birth spacing would also be beneficial. In North Carolina, for example, 11% of small-for-gestational-age births to white women and 21% of such births to black women could have been prevented by lengthening all 0- to 3-month interpregnancy intervals to at least 4 months.⁸ Clinicians should counsel all women to space pregnancies at least 12 months apart, if at all possible.

In 1999 Klebanoff¹⁷ noted the remarkable consistency in reporting the association of short interpregnancy intervals with increased risk of poor perinatal outcomes. Since then, additional findings have accrued on this topic. The meta-analyses by Conde-Agudelo et al in this issue of JAMA summarize these data and provide pooled effect measures. What is needed now is the translation of these solid research findings into practice and future studies to test hypotheses related to the effect of short and long intervals while adequately controlling for breastfeeding. More research is needed to determine whether folate supplementation should be recommended for women after delivery. Regarding the elevated risk associated with long interpregnancy intervals, more thought needs to go into generating plausible testable explanatory hypotheses. While longer intervals are likely to be beyond personal control, pregnancies occurring after an interval longer than 60 months may require more careful monitoring to avoid untoward outcomes.