Sudden infant death syndrome (SIDS) is the leading cause of postneonatal mortality in the United States; the 50% rate of reduction has remained steady since the initiation of the national “Back to Sleep” (now called “Safe to Sleep”) campaign in 1994.1 Most of this decline is attributed to changes in infant sleep position (from prone to back); however, in recent years, the rate of decline in SIDS deaths has plateaued, despite ongoing educational efforts that promote safe sleep and other risk reduction measures.1 Although most of these efforts have addressed the sleep environment, there are a number of other recommended infant care and maternal behaviors that may potentially further reduce the risk of SIDS. In this review, we identify some of these other factors and the magnitude of their association with SIDS and describe interventions to improve maternal behaviors that may lead to greater progress in reducing the incidence of SIDS.
Epidemiologic Studies Investigating the Association Between SIDS and Prenatal Care, Smoke Exposure, Substance Use, and Vaccinations
Several case-control studies have reported an association between prenatal care and SIDS. A national study of linked birth-death files (1995–1998) showed a protective effect of initiation of prenatal care in the first and second trimesters compared with third trimester (adjusted odds ratio [aOR] 0.63, 95% confidence interval [CI] 0.57–0.70, and 0.87, 0.78–0.97, respectively).2 No prenatal care compared with third trimester care almost doubled the risk of SIDS (aOR 1.70, 95% CI 1.44–2.00).2 A population-based case-control study in Washington State found an aOR of 2.5 (1.6–3.9) of third trimester care versus first trimester care.3 The New Zealand Cot Death Study found that infants whose mothers attended less than six antenatal visits had an aOR for SIDS of 1.84 (1.19–2.84) and those who missed more than one visit also had an increased risk (aOR 1.59, 1.18–2.19).4 The Chicago Infant Mortality Study calculated a Kessner index for each mother based on when prenatal care was initiated and the total number of prenatal visits. An intermediate level of prenatal care was associated with an almost 2 times higher risk of SIDS (aOR 1.8, 1.1–2.8) compared with adequate care, whereas inadequate level of care was associated with a 6.5 times higher risk (3.5–12).5
Smoke Exposure Before and After Pregnancy
Before the Back to Sleep campaign, there was a 2- to 3-fold increased risk of SIDS when mothers smoked prenatally and postnatally, with risk increasing with the number of cigarettes smoked.6 A meta-analysis of 39 studies that included publications until April 1997 (after Back to Sleep began) found an unadjusted OR of 2.77 (2.45–3.13) for maternal prenatal smoking and an aOR of 2.08 (1.83–2.38).6 In a comprehensive review of maternal smoking during pregnancy and risk for SIDS, the pooled relative risk (RR) of studies conducted before the campaign (which included recommendations to avoid smoking in pregnancy) was 2.86 (2.77–2.95); the RR of studies conducted after the Back to Sleep campaign began was 3.93 (3.78–4.08).7 It is estimated that up to one-third of SIDS deaths could be prevented if mothers did not smoke during pregnancy.7 A retrospective study in Colorado using birth and death registries (1989–1998) evaluated the contributions of maternal smoking during pregnancy in the pre- and post-Back to Sleep periods. The percent of SIDS attributed to smoking increased from 50% to 80%, as the rate of prone sleeping decreased.8
Because both maternal and paternal smoking prenatally and postnatally are highly correlated, it is difficult to determine the separate influence of each on SIDS. Analyses from the Infant and Maternal Health Survey found that maternal smoking during and after pregnancy was associated with a 3-fold increased risk of SIDS for white and black infants (aOR 3.06, 2.19–4.29 and 3.10, 2.27–4.24, respectively).9 When mothers smoked only after the birth of their infants, the respective odds ratios were lower (aOR, 1.75 [1.04–2.95] and 2.33 [1.48–3.67], respectively). Likewise, a meta-analysis of four studies in which maternal postnatal smoking was controlled for prenatal smoking found that the summary odds ratio was 1.94 (1.55–2.43).6 In another meta-analysis, the summary odds ratio for seven studies in which fathers smoked but mothers didn't was 1.49 (1.25–1.77).7 In the New Zealand Cot Death Study, maternal postnatal smoking alone was associated with an almost 3-fold increased risk for <1 pack per day and a slightly higher risk for ≥1 pack per day (OR 2.56, 1.73–3.75 and 3.43, 2.04–5.77, respectively).10 There was an additional increased risk for SIDS when fathers smoked, but only if the mothers also smoked (OR 4.40, 3.26–5.95 and 7.40, 4.92–11.13, respectively). There was a dose response for environmental tobacco smoke from household members who smoked, but only if the parents smoked also. The risk of SIDS is particularly high when an adult who smoke shares a bed with an infant (OR 2.3–21.6), even when the adult does not smoke in bed.11
Although infants who are low birth weight have a higher risk of SIDS,11 the increased risk associated with prenatal smoking is not believed to be caused by lower birth weight, which is frequently found in infants of mothers who smoke.12 A number of studies in animals examining the effects of in-utero exposure to tobacco smoke or nicotine have shown alterations in neurotransmitter receptors in areas of the brain important for autonomic function.11 In other animal studies, prenatal exposure to nicotine caused increased apnea episodes, hypoventilation, and reduced arousal in response to hypoxia.11 In human infants, prenatal tobacco exposure is associated with changes in autonomic function, arousal and cardiovascular reflexes, all of which may make infants more susceptible to SIDS.11
Alcohol and Illicit Substance Use
It is difficult to separate the effects of individual substances on the risk of SIDS because multiple substances are often used. In addition, few studies have examined the association between substance use and SIDS. A study of Northern Plains American Indians found that periconceptional maternal alcohol use was associated with a 6-fold increased risk of SIDS (aOR 6.2, 1.6–23.3).11 Maternal first-trimester binge drinking was associated with an 8-fold increased risk (aOR 8.2, 1.9–35.3). Both were independent of prenatal cigarette smoking exposure. A retrospective study from Western Australia found that a maternal alcoholism diagnosis during pregnancy or within 1 year postpregnancy was associated with an increased risk of SIDS (adjusted hazard ratio [aHR] 6.92, 4.02–11.90 and 8.61, 5.04–14.69, respectively).13 The authors estimated that at least 16.41% of SIDS deaths were attributable to maternal alcoholism.13 A prospective study from Denmark found that binge drinking (≥4 drinks per week) was associated with an almost 4-fold increased risk of SIDS (aHR 3.56, 1.15–8.43), and ≥3 binge episodes (≥5 drinks per episode) with a 3-fold increased risk (aHR 2.69, 1.27–5.69).14 In a Dutch study, maternal alcohol use in the 24 hours before death was associated with an 8-fold increased risk of SIDS (aOR 8.09, 2.25–29.11).15 However, the New Zealand Cot Death Study did not find an association between having 1 to 7 or ≥8 drinks on any one occasion and SIDS in the month prior to death.16 Siblings of infants with fetal alcohol syndrome have a 10-fold increased risk of SIDS.17
Results from studies of illicit substance use and SIDS are conflicting. A meta-analysis of eight studies found that infants exposed to cocaine had a 4-fold increased risk of SIDS compared with drug-free infants (OR 4.10, 3.17–5.30).18 One of the larger studies included in the meta-analysis was conducted in New York City and found the highest risk was associated with methadone, followed by heroin and cocaine (adjusted risk ratio 3.6, 2.5–5.1; 2.3, 1.3–4.0; and 1.6, 1.2–2.3, respectively).19 Alternatively, a large study from Los Angeles, CA, found a 15-fold increased risk of SIDS for opiate users and a 7-fold increased risk for cocaine users (RR 15.10, 6.3–36.2 and 6.87, 4.04–11.68, respectively).20
Only one study has been published on the association between maternal cannabis use and SIDS. Infants of mothers who used cannabis postpartum had an increased risk of SIDS only when used at night (aOR 2.35, 1.6–4.05), but not during the day.16 Parental alcohol and/or illicit substance use in combination with bed sharing places the infant at particularly high risk for SIDS.11
There is no evidence of a causal relationship between infant vaccinations and SIDS. Evidence suggests that vaccinations may confer a protective effect against SIDS. In a meta-analysis of nine case-control studies that reported vaccination rates with diphtheria-pertussis-tetanus, the reductions in risk were about 50% for both unadjusted (OR 0.50, 0.46–0.73) and adjusted summary odds ratios (aOR 0.54, 0.39–0.75).21
Interventions to Promote Healthy Behaviors
The national Safe to Sleep campaign provides extensive education about reducing the risk of SIDS. Although including advice to avoid smoking during pregnancy and around the infant, as well as recommending routine infant vaccinations, the campaign focuses most heavily on the sleep environment (ie, infant sleep position, bedding materials, surfaces, and location).22 To further reduce the incidence of SIDS to as close to zero as possible, implementing the other recommendations to reduce the risk is essential. In the subsequent sections, we discuss published interventions that have sought to increase rates of prenatal care, smoking and substance use cessation, and infant vaccination adherence.
A 1995 Cochrane review identified five studies in the Cochrane Pregnancy and Childbirth Group's Trial Register that used direct incentives to pregnant women explicitly linked to initiation and frequency of prenatal care.23 Three of the studies were in the US and two in Central America. Incentives included cash, store gift cards, taxi vouchers for transportation to and from clinic, baby blankets, and baby carriers. Outcomes included adequacy and frequency of prenatal care, and timing of the first prenatal care visit. Pregnant women receiving incentives were more likely to obtain adequate quality prenatal care defined by the number of “procedures” they received (history-taking, diagnostic tests, physical examination, lactation and family planning counseling, iron supplementation, and immunization); the mean difference was 5.84, 1.88–9.80. Women who received incentives were more likely to obtain frequent prenatal care, defined as greater than five visits in one trial for which data were available (risk ratio 1.18, 1.01–1.38). There was no difference in early initiation of prenatal care between the intervention and control groups (RR 1.04, 0.78–1.38). Of note, recipients of incentives were more likely to deliver their infants by cesarean delivery (RR 1.97, 1.18–3.30). These findings should be interpreted with caution, given the limited number of studies, the relatively small total sample size, and the limited generalizability of the studies, as many participants were drawn from impoverished communities in Central America.
Approximately 10% of women in the US smoked during the last 3 months of pregnancy, according to the 2011 Pregnancy Risk Assessment and Monitoring System.24 Smoking during pregnancy increases the chance of complications in pregnancy, spontaneous miscarriages, stillbirth, low birth weight, and preterm birth.25,26 It also increases the chances for growth and developmental delays of the fetus as well as future health problems for the baby and mother.27 However, it has been shown that quitting smoking at any time during the pregnancy has benefits and can improve fetal growth.27
The World Health Organization recommends that all pregnant women who are either current smokers or who have recently quit, receive advice and psychosocial interventions by health care providers.27 A large meta-analysis of 77 randomized controlled trials was conducted to assess the effects of psychosocial smoking cessation interventions during pregnancy. Psychosocial interventions are nonpharmacological strategies to help motivate women to quit smoking, including counseling, health education, feedback, financial incentives, and social support from peers or partners.26 The pooled results from this study found an 18% reduction in both preterm births and low birth weight in women who received psychosocial interventions to stop smoking during pregnancy.26 Incentive-based interventions appeared to have the greatest effect in having women stop smoking, but only if provided intensively. However, all of the interventions were positively received by women and did not appear to have any negative physical or psychological effects. Although some studies have suggested that the effectiveness of individual interventions is lower among women of lower socioeconomic status (SES), the meta-analysis found that the effectiveness was the same across SES.26
A study in Scotland found that 24% of pregnant women self-reported as being smokers in 2009.25 This led to the National Institute for Health and Care Excellence to request evidence for the effectiveness of financial incentives for smoking cessation during pregnancy. This study of 612 participants found that financial incentives helped to motivate smokers to quit, with 22.5% in the intervention group quitting smoking versus 8.6% in the control group.25
Pharmacotherapies (nicotine replacement therapy [NRT], bupropion, and varenicline) have been found to be effective for smoking cessation among nonpregnant women who were smokers.28 However, the safety and efficacy among pregnant women who are smokers is unknown. A recent review of nine trials that enrolled pregnant women who are smokers found that NRT was no more effective than a placebo and there was no evidence that NRT affects birth outcomes.28
Alcohol is used by approximately 15% of pregnant women, with 3% to 4% reporting binge drinking during pregnancy.29 “Risky drinking” (defined as >7 drinks per week and >3 drinks on a single occasion)29 often precedes alcohol use in pregnancy and accounts for women's difficulty in obtaining abstinence from alcohol during pregnancy, even when they are concerned about its harmful effects on the fetus. The American Colleges of Obstetrics and Gynecology recommends that risky drinkers be advised to cut back on alcohol use prior to conception as many women don't know they are pregnant during the early stages when embryogenesis is taking place. Prior drinking is also predictive of drinking during pregnancy.29
A randomized controlled trial in Boston, MA, of 304 pregnant women and their partners evaluated the effectiveness of a brief, one-time intervention consisting of a knowledge assessment and feedback, contracting and goal setting, and behavioral modification.30 Monetary incentives were provided for completion of initial and postpartum interviews. The brief intervention was more effective in reducing alcohol consumption during pregnancy among women who were drinking more heavily at the time of intervention. In addition, the effectiveness of the intervention was significantly enhanced when the woman's support partner also participated in the intervention. Reductions in alcohol consumption during pregnancy were found in both intervention and control groups.
A Cochrane Database Systematic Review was conducted to assess the effectiveness of psychological and educational interventions to reduce alcohol consumption among pregnant women or women planning pregnancy.31 Four randomized controlled trials met the inclusion criteria. For most outcomes, there were no significant differences between groups, and results related to reducing or abstaining from alcohol use were mixed. The small number of studies, the high risk of bias in some of the studies, and complexity of interventions make it difficult to identify the most effective strategy to limit alcohol use during pregnancy.
A randomized controlled trial compared the efficacy of motivational enhancement therapy coupled with cognitive-behavioral therapy (MET-CBT) brief advice for treatment of alcohol and substance (nonopiate) use in pregnancy.32 The MET-CBT was provided by trained nurses for six 30-minute sessions, while the brief advice was provided by the participant's obstetrical provider, typically lasting about 1 minute. Substance use decreased in both groups between intake and delivery (according to self-report and urine testing), but increased again after delivery. There were no significant differences between groups.
A Dutch study randomly assigned pregnant women to 1 of 3 intervention arms: computer-tailored feedback, health counseling by their midwives, or routine care by their midwives.33 The computer-tailored intervention, which was anonymous, involved the participant responding to questions via a computer program, with advice then generated by the program to respond to the participant's specific circumstances and needs. The computer-tailored group stopped using alcohol more often than the counseling and routine care groups.
The US is currently experiencing an epidemic of drug overdose deaths and between 2000 and 2009, there was a 5-fold increase in opiate use in pregnancy.34 For pregnant women who use opioids, methadone maintenance has been successful at reducing relapse, reducing risk-taking behavior, and increasing compliance with prenatal care. However, the process of reducing the dosage of methadone over time can lead to higher relapse and higher rates of fetal morbidity and mortality. Another potential therapy, buprenorphine, appeared to have a positive effect on the infants of mothers who received this treatment.34 Interventions that provided monetary incentives also seemed to be effective in treating opioid use in pregnancy and could be used in conjunction with the other therapies.34
A randomized controlled trial found that contingency management therapy (positive reinforcement with monetary vouchers) was successful in reducing cocaine use as compared to a community reinforcement approach. Another study looking at the use of motivational interviewing among people who use substances did not find that it was more effective than usual care.35 However, the addition of incentives improved compliance with outpatient treatment attendance and retention.36
To date, there have been no studies on the use of pharmacological treatments for cocaine use during pregnancy.34 However, preliminary findings from a randomized, placebo-controlled trial showed that women who were given oral micronized progesterone self-reported less use of cocaine compared to women taking placebo. This needs to be confirmed in larger clinical trials.
Two studies found reinforcement-based therapy (individual counseling supplemented by abstinence-contingent support for housing, food, recreational activities, and skills training) to have the potential to reduce the use of methamphetamine by pregnant women, but more research is required.34
Interventions that target cannabis use in pregnant women have not been evaluated. For women who are not pregnant but use marijuana, motivational interviewing, CBT, and contingency management appear to have had some success in reducing use.34 With the current movement for legalization of marijuana in the US, interventions to identify users and reduce use during pregnancy should be promoted.
Infant and childhood vaccinations are important for the health and well-being of children. However, few studies have evaluated “vaccine hesitancy” or have quantified any effect of interventions. A recent systematic review found that the most effective interventions involved multiple strategies, including dialogue-based interventions, reminder-recall approaches, and incentive-based interventions.37 The outcomes measured were vaccine uptake and knowledge, awareness, or attitudes regarding vaccinations. The interventions that were tailored to people or populations and addressed specific concerns appeared to be the most effective. Passive interventions, such as posters and websites, were associated with <10% vaccine uptake.37 A meta-analysis of studies from low- and middle-income countries found that interventions that incentivize demand for vaccine, or “demand-side interventions” (monetary incentives, knowledge transfer, or communication campaigns) lead to increased vaccine uptake.38
Another review assessed the effects of face-to-face interventions to inform parents and caregivers about childhood vaccinations. However, due to the limited evidence available and the quality of the studies, they concluded that these interventions have little or no effect on vaccine uptake or knowledge.39 A meta-analysis by Harvey et al.40 found that postal reminders were more effective than phone reminders on childhood vaccine uptake, but the use of both reminders (postal and phone) together resulted in even greater vaccine uptake. They also found that discussions with a health care provider increased the likelihood of vaccine uptake, as opposed to parents receiving written educational information. They concluded that postal and telephone reminders may be sufficient for most cases and that discussion-based interventions may be more effective with parents with high levels of vaccine hesitancy and for families of children with high rates of noncompliance.40
One systematic review found 19 studies that applied new media (smartphone applications, text messaging, social media, websites, and portals) to increase vaccine uptake and coverage.41 They found some evidence that text messaging, accessing immunization campaign websites, web-based portals, and computerized reminders increased vaccine uptake. There were little to no data available on the use of social media, email, and smartphone applications on vaccine coverage rates.
Although the Internet has tremendous ability to connect people with information, there is also a fair amount of negative and incorrect messaging regarding vaccinations. Anti-vaccination blogs and websites abound and may discourage people from getting vaccinated.41 However, this strategy of using the Internet and new media has great potential to increase vaccine uptake and coverage. More research will be needed to assess the effectiveness and cost-effectiveness of such interventions.