Outcomes in Pregnancies Complicated by Methamphetamine Use

This retrospective cohort studyCohort StudyDefinition: A cohort study follows groups of people over time to compare outcomes between those exposed to a factor and those who are not.Evidence quality: Cohort studies occupy the middle of the evidence pyramid and provide strong evidence of association but not definitive causation.Court use: Courts may infer increased or decreased risk, but should not conclude that the exposure directly caused the outcome in an individual case.Glossary › used linked California maternal and infant administrative datasets for all pregnancies from 2005 through 2008. The analysis identified 8,542 pregnancies (0.4%) with a documented ICD-9 diagnosis related to methamphetamine use and compared them with 2,031,328 pregnancies without such a diagnosis; pregnancies with cocaine, opioid, and/or cannabis exposure were excluded. 

After multivariable adjustment, methamphetamine-related diagnosis was associated with higher odds of gestational hypertension (OR 1.8, 95% CI 1.6–2.0), preeclampsia (OR 2.7, 95% CI 2.4–3.0), severe preeclampsia (OR 3.9, 95% CI 3.3–4.6), abruption (OR 5.5, 95% CI 4.9–6.3), intrauterine fetal death (OR 5.1, 95% CI 3.7–7.2), preterm birth (OR 2.9, 95% CI 2.7–3.1), birthweight <2500 g (OR 3.5, 95% CI 3.3–3.8), neonatal death (OR 3.1, 95% CI 2.3–4.2), and infant death (OR 2.5, 95% CI 1.7–3.7). The authors concluded that maternal methamphetamine use during pregnancy was associated with increased maternal, fetal, and infant morbidity and mortality.

This large, population-based retrospective cohort provides estimates of several outcomes but establishes associationsAssociationWhat it is: A statistical relationship between two variables.Does not mean legally: Association alone does not establish causation.Glossary › rather than causationCausationWhat it is: A relationship in which one factor produces an outcome.Does not mean legally: Scientific causation at the population level does not automatically establish legal causation in an individual case.Glossary ›. Methamphetamine exposure was identified through ICD-9 diagnosis codes rather than standardized exposure assessment, Because the diagnosis code was the exposure measure, the study also could not determine how methamphetamine use was identified in an individual case—for example, through self-report, toxicology testing, or another clinical source. The authors also noted that failure to diagnose or code methamphetamine use likely resulted in some users being included in the comparison group, potentially biasing the results toward the null. However, ICD coding may also identify a selected subset of people whose methamphetamine use was more clinically visible or severe, so the direction and magnitude of exposure misclassificationMisclassificationWhat it is: Incorrect labeling of exposure or outcome status.Does not mean legally: Numerical precision does not imply factual accuracy if classification is flawed.Glossary › cannot be determined from this study alone. The exposed and comparison groups also differed substantially in prenatal care and other characteristics: 25.5% of the methamphetamine group had no prenatal care versus 2.0% of controls,42.1% had fewer than five prenatal visits versus 4.0% and 88.9% had public or no insurance versus 48.5% of controls. Although adjusted models included race/ethnicity, age, insurance, parity, education, prenatal care, chronic hypertension, diabetes, gestational diabetes, tobacco, and alcohol, the authors could not control for all possible substance exposures and acknowledged residual confoundingConfoundingWhat it is: When another factor influences both the exposure and outcome, creating a misleading association.Does not mean legally: An observed association cannot be assumed causal when confounding has not been adequately addressed.Glossary ›.

They further noted that methamphetamine use may serve as a marker for multiple interacting biopsychosocial factors associated with poor pregnancy outcomes. The relative measures of association should also be considered alongside the absolute frequencies: preterm birth occurred in 23.4% versus 8.9%, abruption in 5.2% versus 0.8%, and intrauterine fetal death in 1.4% versus 0.3%. The association with neonatal death was no longer present after adjustment for gestational age (OR 0.7, 95% CI 0.5–1.1), suggesting that prematurity may account for much of that association. Accordingly, the study supports an association between a recorded methamphetamine-use diagnosis and adverse pregnancy and infant outcomes in California during 2005–2008. However, it cannot establish that methamphetamine use alone caused these outcomes or determine the extent to which other factors, including social conditions, access to health care, or use of other substances, contributed to the observed associations. It also cannot be used to predict the outcome of an individual pregnancy based solely on a history or diagnosis of methamphetamine use.

Gorman et al. is frequently useful because its very large population-based sample demonstrates that a documented methamphetamine-related diagnosis identifies a population with substantially higher rates of several adverse pregnancy outcomes. But the size of the study should not be confused with greater certainty about what caused those outcomes. The pregnancies identified by a methamphetamine diagnosis differed markedly from the comparison group in prenatal care and socioeconomic circumstances, and the authors themselves recognized that methamphetamine use may function as a marker for multiple interacting biopsychosocial factors associated with poor pregnancy outcomes.

The distinction between identifying a higher-risk population and attributing an outcome to methamphetamine is critical. This study cannot by itself establish that methamphetamine caused a particular adverse outcome or that a particular pregnancy will have an adverse outcome because methamphetamine use occurred.

This study cannot by itself establish that methamphetamine caused an adverse outcomes or that a pregnancy will have an adverse outcome because methamphetamine use occurred. Many different biopsychosocial markers, in addition to methamphetamine use, could have contributed to poor outcomes.