Rising preconception and prenatal cannabis use in the United States is a growing public health issue.¹ Between 2002 and 2021, cannabis use in pregnancy doubled from 3.4% to 7.2% of pregnant individuals, albeit higher in the first trimester than in later trimesters.^{2, 3} During the COVID-19 pandemic, nearly 25% of pregnant individuals in some parts of the United States reported cannabis use.¹ These changing patterns of substance use in the preconception and prenatal periods likely reflect the growing legalisation of cannabis in many states, public health interventions aimed at decreasing alcohol and tobacco use in pregnancy and the perceived safety of cannabis use.⁴

Emerging data suggest that prenatal cannabis exposure is associated with increases in the risk of adverse pregnancy outcomes, including low birthweight, neonatal intensive care unit admission and preterm birth,⁴ as well as long-term adverse child neurodevelopmental outcomes.⁵ The problem is that cannabis use co-occurs with tobacco use nearly half the time, and tobacco exposure is a well-recognised risk factor for these same (and several more) adverse pregnancy outcomes.⁴

Many individuals use cannabis in the prenatal period to address symptoms of nausea, pain and stress.⁴ Cannabis use may increase appetite and promote weight gain. Whether metabolic changes due to cannabis use in the preconception period may increase the risk of gestational diabetes mellitus (GDM), the most frequent metabolic complication of pregnancy affecting one in eight pregnant individuals in the United States annually, requires further investigation.⁶

In this issue of Paediatric and Perinatal Epidemiology, Pan and colleagues⁷ combined data (meta-analysis) from eight prospective studies to evaluate associations of preconception cannabis use with GDM, and then secondarily, to assess whether this association was modified by tobacco use and body mass index (BMI). The investigators harmonised data in the Preconception Period Analysis of Risks and Exposures influencing health and Development (PrePARED) consortium which included 17,880 pregnant individuals. PrePARED is a consortium of studies incorporating information on preconception and pregnancy health. Exposure to cannabis was primarily based on self-report with only one study using urine samples. Perinatal outcome ascertainment was exclusively by self-report.

The authors reported no association between preconception cannabis use in the past year and GDM. However, when analyses were stratified by tobacco use, the authors found an association among those who never used tobacco: Those who used cannabis more than a week had a higher risk of developing GDM than those who did not use cannabis in the past year (odds ratio 2.65, 95% confidence interval 1.15, 6.09). This association was not present among other groups with tobacco use, including former or current tobacco users. These results were similar across maternal BMI. In sensitivity analyses, the authors reported that probabilistic bias analysis to correct for potential exposure misclassification and unmeasured confounding including physical activity had a minimal effect on the estimated effect.

These results naturally lead to the question of clinical and public health implications of why cannabis exposure would be associated with GDM only among individuals who reported never using tobacco. Well-established risk factors for GDM include increasing maternal age, BMI and gestational weight gain.⁶ To what extent substance use may affect GDM risk remains unclear. Some recent data from the Pregnancy Risk Assessment Monitoring System (PRAMS) suggest a link between prenatal tobacco use and an increased risk of GDM.⁸ In this analysis, the authors speculate that preconception cannabis use may increase appetite, promote fat deposition and adipogenesis and increase insulin resistance. This is based on preclinical studies that have uncovered potential underlying mechanisms for the observed adverse outcomes but do not necessarily recapitulate typical human use.

The associations between cannabis use, these mediating risk factors and GDM remain to be fully established. (Further) Why these factors would increase the risk of GDM with cannabis use only among tobacco nonusers is puzzling. Most participants who used cannabis before pregnancy did so less than weekly, possibly explaining the overall null association between preconception cannabis use and GDM. The intensity and frequency of preconception cannabis use may likely differ between tobacco users and nonusers, as well as other differences in risk behaviours and socio-demographic characteristics associated with cannabis use. Whether these differences are sufficient to affect GDM risk requires further study, which may be challenging given so many individuals who use cannabis also use tobacco.

Some important limitations of this study, and many in this field, include measurement error and unmeasured confounding. Self-report may underestimate substance use and misclassify individuals in terms of exposure intensity. Various forms of cannabis and routes of exposure may also affect the dose. Other substances (i.e. cocaine, heroin, synthetic opioids) may be unmeasured or poorly measured confounders. In addition, included studies did not assess whether maternal preconception cannabis use continued during pregnancy as well as paternal cannabis use, both of which may plausibly enhance the risk of adverse outcomes.

Strengths of this study include the inclusion of only prospective cohorts and attempts at harmonisation of data across studies. The authors also attempted to capture patterns of use beyond dichotomous classification by frequency of use, which may reveal nuances in estimates of exposure-related risk that would be otherwise missed. Within the framework of a meta-analysis, the authors conducted extensive sensitivity analyses to account for potential exposure misclassification and unmeasured confounding, multiple imputations for missing covariates to address selection bias and adjustment when possible, for the use of other drugs. The authors also conducted subset analyses restricted to nulliparous individuals as well data from the year 2000 and onwards to account for changing trends in drug use.

Currently, the vast majority of pregnant individuals who use cannabis report that they believe there is little or no harm concerning adverse perinatal outcomes from occasional marijuana use.⁹ The American College of Obstetricians and Gynecologists recommends that pregnant individuals and those considering pregnancy should be counselled about concerns regarding potential adverse health consequences of continued cannabis use during pregnancy and should be encouraged to discontinue cannabis use.¹⁰ Nevertheless, unlike tobacco and alcohol use, providers often do not consistently counsel pregnant patients or those planning to get pregnant about cannabis use because they also underestimate or lack knowledge about the health risks.

In conclusion, this meta-analysis found that self-reported cannabis use appears to have ‘little effect’ on overall GDM risk. These results do highlight the importance of investigating substance use in the preconception period, which is an important period to address risks for preventing adverse pregnancy outcomes. Researchers, providers, policymakers and patients require further evidence that highlights the consequences of cannabis use in the preconception and prenatal periods. This meta-analysis helps address this gap in evidence by highlighting the clinical and public health importance of examining the impact of cannabis, the most frequently used substance in pregnancy, on GDM, one of the most frequent medical complications of pregnancy.

在美国，孕前和产前大麻使用的增加是一个日益严重的公共卫生问题。¹ 2002 年至 2021 年间，怀孕期间吸食大麻的孕妇比例翻了一番，从 3.4% 增至 7.2%，尽管妊娠早期的比例高于妊娠后期。^{2, 3}在 COVID-19 大流行期间，美国部分地区近 25% 的孕妇报告使用过大麻。¹孕前和产前时期物质使用模式的这些变化可能反映了许多州大麻日益合法化、旨在减少怀孕期间饮酒和吸烟的公共卫生干预措施以及人们对大麻使用安全性的认识。⁴

新数据表明，产前接触大麻与不良妊娠结局的风险增加有关，包括低出生体重、新生儿重症监护室入住和早产⁴以及长期不良儿童神经发育结局。⁵问题在于，近一半的时间大麻使用与烟草使用同时发生，而烟草暴露是导致这些相同（以及其他几种）不良妊娠结局的公认风险因素。⁴

许多人在产前使用大麻来缓解恶心、疼痛和压力症状。⁴使用大麻可能会增加食欲并促进体重增加。孕前使用大麻导致的代谢变化是否会增加妊娠糖尿病（GDM）的风险，这是最常见的妊娠代谢并发症，每年影响美国八分之一的孕妇，需要进一步调查。⁶

在本期儿科和围产期流行病学中，Pan 及其同事⁷结合了八项前瞻性研究的数据（荟萃分析），以评估孕前大麻使用与 GDM 的关联，然后评估这种关联是否因烟草使用和身体状况而改变质量指数（BMI）。研究人员协调了影响健康和发展的风险和暴露的孕前期分析 (PrePARED) 联盟的数据，该联盟包括 17,880 名孕妇。PrePARED 是一个研究联盟，纳入了有关孕前和怀孕健康的信息。大麻暴露主要基于自我报告，只有一项研究使用尿液样本。围产期结局完全通过自我报告确定。

作者报告说，过去一年孕前吸食大麻与 GDM 之间没有关联。然而，当按烟草使用情况进行分层分析时，作者发现从未使用过烟草的人之间存在关联：那些使用大麻超过一周的人比那些在过去一年中不使用大麻的人患 GDM 的风险更高（赔率比率 2.65，95% 置信区间 1.15、6.09）。这种关联在其他吸烟群体中不存在，包括以前或现在的吸烟者。这些结果在母亲体重指数上是相似的。在敏感性分析中，作者报告说，用于纠正潜在暴露错误分类和包括体力活动在内的未测量混杂因素的概率偏差分析对估计效果的影响很小。

这些结果自然会引发临床和公共卫生影响的问题，即为什么仅在从未使用烟草的个体中接触大麻与 GDM 相关。GDM 的明确危险因素包括母亲年龄增加、体重指数 (BMI) 和妊娠期体重增加。⁶物质使用在多大程度上可能影响 GDM 风险仍不清楚。妊娠风险评估监测系统 (PRAMS) 的一些最新数据表明，产前吸烟与 GDM 风险增加之间存在联系。⁸在这项分析中，作者推测，孕前使用大麻可能会增加食欲，促进脂肪沉积和脂肪生成，并增加胰岛素抵抗。这是基于临床前研究，这些研究揭示了观察到的不良结果的潜在潜在机制，但不一定概括典型的人类使用。

大麻使用、这些中介风险因素和 GDM 之间的关联仍有待完全确定。（进一步）为什么这些因素会增加仅在不吸烟者中使用大麻的 GDM 风险令人费解。大多数在怀孕前吸食大麻的参与者每周吸食大麻的时间少于一周，这可能解释了孕前吸食大麻与 GDM 之间总体上没有关联的原因。烟草使用者和非使用者之间，孕前使用大麻的强度和频率可能有所不同，以及与大麻使用相关的风险行为和社会人口特征的其他差异。这些差异是否足以影响 GDM 风险需要进一步研究，考虑到许多吸食大麻的人也吸食烟草，这可能具有挑战性。

这项研究以及该领域的许多研究的一些重要局限性包括测量误差和未测量的混杂因素。自我报告可能会低估药物的使用情况，并根据暴露强度对个人进行错误分类。各种形式的大麻和接触途径也可能影响剂量。其他物质（即可卡因、海洛因、合成阿片类药物）可能是未测量或测量不当的混杂因素。此外，纳入的研究没有评估母亲在怀孕期间是否继续使用孕前大麻以及父亲使用大麻，这两者都可能增加不良后果的风险。

这项研究的优点包括仅纳入前瞻性队列并尝试协调跨研究的数据。作者还试图捕捉除按使用频率进行二分分类之外的使用模式，这可能会揭示暴露相关风险估计中的细微差别，否则可能会被忽略。在荟萃分析的框架内，作者进行了广泛的敏感性分析，以解释潜在的暴露错误分类和未测量的混杂因素，对缺失协变量进行多重插补，以解决选择偏差，并在可能的情况下对其他药物的使用进行调整。作者还进行了仅限于未生育个体的子集分析以及 2000 年及以后的数据，以解释吸毒趋势的变化。

目前，绝大多数使用大麻的孕妇报告称，他们认为偶尔使用大麻对围产期不良后果影响很小或没有危害。⁹美国妇产科医师学会建议，应向怀孕者和考虑怀孕的人咨询有关怀孕期间继续使用大麻对健康造成的潜在不良后果的担忧，并应鼓励他们停止使用大麻。¹⁰然而，与烟草和酒精的使用不同，提供者往往不会始终如一地就大麻的使用向怀孕患者或计划怀孕的患者提供咨询，因为他们也低估或缺乏对健康风险的了解。

总之，这项荟萃分析发现，自我报告的大麻使用似乎对总体 GDM 风险“影响不大”。这些结果确实凸显了在孕前阶段调查药物使用的重要性，这是解决预防不良妊娠结局风险的重要时期。研究人员、提供者、政策制定者和患者需要进一步的证据来强调在怀孕前和产前期间使用大麻的后果。这项荟萃分析强调了检查大麻（妊娠期最常用的物质）对妊娠期糖尿病（妊娠期最常见的医学并发症之一）影响的临床和公共卫生重要性，有助于解决这一证据差距。