Introduction

Smoking in pregnancy causes much morbidity and mortality [1–3]. A systematic review of randomised control trials (RCTs) shows that stopping smoking in pregnancy improves birth outcomes [4]. Many countries, including the UK, recommend nicotine replacement therapy (NRT) for smoking cessation during pregnancy [5, 6] and 11% of UK pregnant smokers receive NRT prescriptions [7]. Although NRT provides nicotine without other toxic elements present in tobacco smoke, the ability for nicotine to cross the placenta and concentrate in fetal blood and amniotic fluid [8] leads to concerns that nicotine within NRT may cause fetal harm [9].

Preclinical literature could inform us of potential harms that might be due to nicotine, and many preclinical studies have investigated relationships between nicotine administered to pregnant animals and adverse outcomes [10]. However, there is no agreed core outcome set for measuring potential offspring harm from nicotine and there appears to be great diversity in outcomes used. For many preclinical study outcomes, there is no direct evidence that these are indicative of harm to humans; hence, such outcomes are probably best considered theoretically indicative of harm to infants born to pregnant women who use NRT. For example, in one study fetal brain stem injury was observed in rat offspring after gestational nicotine administration in maternal drinking water, and it was extrapolated that maternal smoking in pregnancy may cause similar issues, potentially leading to sudden infant death syndrome [11]. Other preclinical study issues which make finding’s relevance to humans difficult to ascertain include outcomes measured ‘in vitro’, in isolated parts of animals or in species which have little in common with humans [12].

Systematic reviews involve collating all studies relevant to research questions together, and through evidence synthesis can comprehensively answer important research questions[13]. For example, a recent systematic review of studies investigating NRT use by pregnant women found no evidence that NRT use during pregnancy causes fetal harm [14]. However, despite a thorough search, we have been unable to find any systematic reviews of preclinical studies investigating the impacts of nicotine exposure in pregnancy on mammalian offspring outcomes. We found a 20-year old review, but within this, search strategies were not described, so it was impossible to determine how comprehensive this was [10]. Other reviews have been narrative with selective choices of included literature [8, 15–18], and this lack of systematic preclinical literature synthesis has resulted in calls for rigorous, regularly-updated systematic reviews of preclinical studies to focus research efforts [19].

Scoping reviews allow rapid identification of unanswered research questions, and may indicate both the feasibility and potential utility of conducting full systematic reviews [20]. We have conducted a systematic scoping review of preclinical studies investigating potentially adverse impacts of gestational nicotine administration on offspring outcomes. We identify, quantify and categorise the full range of offspring outcomes measured in studies involved, and then use the taxonomy developed to hypothesise about the utility of preclinical literature for identifying potential adverse effects of nicotine on human offspring following administration of NRT in pregnancy.

Methods

We adapted the framework for conducting scoping reviews described by Arskey and O’Malley [20]. This involved: posing a research question; identifying potentially relevant studies; selecting those appropriate for inclusion; extracting and charting relevant data; and then collating, summarizing and reporting findings. We used established methods [21] to comprehensively search for preclinical studies in which non-tobacco nicotine has been administered to mammals in pregnancy, and we identified those which reported fetal or offspring outcomes.

Screening strategy (Study selection)

Initially, LP screened citations (i.e. titles +/- abstracts) and those that were obviously not relevant to the review were excluded. Next, using pre-determined inclusion and exclusion criteria, three reviewers (LP, RT and TC) each piloted study selection procedures on 100 citation records. Individual assessors noted uncertainties and discussed these with the other two researchers. Following resolution of issues arising from the pilot screening, and based on increasing familiarity with the literature, inclusion and exclusion criteria for the scoping review were finalised (Table 1). LP and RT then applied the inclusion and exclusion criteria to all the citations. Citations were grouped into those for inclusion, exclusion or where a decision could not be made. Next, all reviewers (LP, RT & TC) discussed citations where there was uncertainty about whether to include or exclude them.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0280805.t001

Results

Searches produced 6354 citation records, 1798 of which were duplicates. Following title and abstract screening of the remaining 4556 records, 476 records were assessed as relevant for inclusion in the scoping review (Fig 1). We were unable to obtain abstracts for seven of the 476 citation records, meaning outcomes for these citations were determined by the title alone.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA diagram.

Prisma diagram of citation selection process.

https://doi.org/10.1371/journal.pone.0280805.g001

We identified 729 fetal and offspring outcomes within the 476 abstracts or titles. The majority of papers (440) reported outcomes relating to rodents, while 16 related to sheep, 14 to primates and 3 to rabbits. The draft taxonomy for categorising outcomes contains three classification groups: outcomes measured in i) complete animals, ii) body systems and, iii) ‘other’ (see Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Draft taxonomy for categorising outcomes in preclinical literature.

https://doi.org/10.1371/journal.pone.0280805.t002

Discussion

We have provided a thorough, quantitative description of outcomes used in preclinical studies investigating mammalian prenatal nicotine exposure. We identified 729 different fetal or offspring outcomes reported from 476 studies. The most frequently reported outcomes measured cellular-level brain changes; respiratory system outcomes measuring either cellular, macroscopic or functional changes, and physical outcomes measured in whole animals, such as birth weight and litter size.

Many studies reported multiple outcomes and, as we only extracted data from titles and abstracts; some outcomes reported only in manuscripts’ full texts may have been missed. However, authors usually include the most important study outcomes in abstracts, so we should have identified those outcomes which researchers most strongly believed to be related to maternal nicotine administration. Also, as authors are likely to be biased towards including statistically significant findings in abstracts, we have probably also identified most outcomes for which statistical associations with nicotine administration were demonstrated. If we had retrieved and extracted information from manuscripts’ text, we may have found extra methodological details to prompt either exclusion from or inclusion in the review. For example, some studies might have involved nicotine administration in both pregnancy and the weaning period, without study abstracts mentioning this. Hence, some included studies may have assessed impacts other than those of nicotine administration in pregnancy. It is unclear how much these factors might have affected the number and breadth of outcomes identified, but there is no reason to suspect effects would be substantial.

A strength of this work is its novelty. It is the first to offer a systematic and comprehensive overview of principal offspring outcomes in animal studies investigating the impacts of nicotine in pregnancy, and the draft taxonomy is the first attempt to categorise these. Another study strength lies with its rigour. We followed established methods for conducting scoping reviews and carefully developed search strategies for identifying relevant work. This included, for one of the bibliographic databases searched, applying a filter designed to find all preclinical studies. Hence, we believe that we will have missed few relevant papers. Similarly, we used a robust, team-orientated approach to extract data; two researchers independently conducted this with arbitration from a third, where necessary. Consequently, we believe that most outcomes have been identified from citations and that these will have been categorised consistently. To ensure that the draft taxonomy was appropriately descriptive, we adopted an iterative developmental approach, revising categories and descriptors to help the final version have the greatest possible utility. However, this draft taxonomy can be revised if necessary.

This work helps move the literature beyond narrative reviews [15–19]. Although the manuscript makes no quality assessments, by demonstrating which outcomes are most regularly reported, it highlights where further investigation might be most fruitful. Detailed systematic reviews would most likely provide useful information if these answered focussed questions about impacts on the brain, respiratory system or on whole animals’ physical characteristics. Such reviews could investigate the effects of varying nicotine doses and administration regimens on the same outcomes across all preclinical studies, and by including robust evaluation included studies’ methodological quality, would enable effective assessment of their findings’ validity. Such reviews would very likely help us better understand any relationships between gestational nicotine exposures and offspring harm in preclinical studies.

Whilst a systematic approach to synthesising the preclinical literature could indicate ‘signals’ of potential harms from nicotine, the question of how findings relate to humans and in particular, the impacts of NRT used in pregnancy for smoking cessation remain difficult to answer. Whether or not nicotine administered in pregnancy is harmful to human infants can only be fully answered by studies conducted in women.

Conclusions

We provide an overview of preclinical, mammalian study outcomes used to investigate impacts on offspring from prenatal nicotine exposure, and a draft taxonomy that categorises outcomes in relation to offspring’s development, affected body systems and their functions. This highlights those areas where most preclinical research has focused, and therefore could benefit most from full systematic reviews to fully understand any relationships between gestational nicotine administration and offspring harms.

Supporting information