The Growth and Energy Balance Consequences of Adolescent Inhalant Abuse

Home , Rebound effect

The growth and energy balance consequences of adolescent inhalant abuse

Rose Crossin BSc (Hons)

ORCID: https://orcid.org/0000-0003-1814-1330

Submitted in total fulfilment of the requirements for the degree

Doctor of Philosophy

October 2018

The Florey Department of Neuroscience and Mental Health

The University of Melbourne

Rose Crossin (737900)

Abstract Inhalant abuse is the intentional inhalation of products containing toluene (e.g. glue or petrol) in order to induce an altered mental state. The primary populations who misuse inhalants are adolescents, particularly those in vulnerable or disadvantaged groups, including indigenous peoples. Inhalant abuse has been associated with symptoms such as appetite suppression and impaired weight gain, suggesting that it may have an effect on energy balance. Disruptions to energy balance may have an effect on growth, particularly during periods of rapid growth such as adolescence. However, the energy balance and growth consequences arising from adolescent inhalant abuse have never been fully characterised, and the persistence of these during abstinence remains unknown.

Thus, the aims of this PhD thesis were to:

1. characterise the energy balance and growth consequences of adolescent inhalant abuse; 2. understand how these effects may persist into abstinence, and; 3. elucidate the mechanisms underlying these changes.

Initial exploration of the growth consequences of adolescent inhalant abuse, and the persistence of effects into abstinence, was conducted using a retrospective analysis of data collected from a cohort of Australian Indigenous males who chronically sniffed petrol during adolescence. This investigation was further expanded by conducting a systematic review and meta-analysis on the growth consequences of inhalant abuse (in clinical studies) and toluene exposure (in pre-clinical studies). This included sub-group analysis to identify potential moderators of the effects. Characterisation of the effects of adolescent inhalant abuse, within the context of the energy balance equation, and the testing of potential mechanistic hypotheses, was conducted using an established rodent model of chronic intermittent toluene (CIT) exposure. In the CIT model, male adolescent rats were exposed to toluene with exposure patterns replicating human inhalant abuse, including a period of sustained abstinence.

Adolescent inhalant abuse significantly impaired both body weight and height, which was consistent across the epidemiological study, meta-analysis, and rodent model. These changes persisted into sustained abstinence. Furthermore, using the rodent model, it was determined that adolescent inhalant abuse induces a negative energy balance phenotype, with decreased food intake occurring concurrently with increased energy expenditure at the end of the exposure period. Potential causal hypotheses for the observed growth impairments were tested. Using a pair-fed group, the hypothesis that toluene induced under-nutrition underlies the observed growth impairments was tested. This hypothesis was rejected by the data, with differences between the pair-fed and CIT group related to homeostatic recovery of body weight, skeletal responses to under-nutrition, and glycaemic status indicating different mechanisms are involved in CIT-induced changes to body weight. A further causal hypothesis was identified, positing that adrenal insufficiency may underlie the observed negative energy balance phenotype and growth impairments. This hypothesis was supported by the data, with adrenal histology and endocrine testing revealing results consistent with primary adrenal

1 Rose Crossin (737900) insufficiency, including adrenal hypertrophy, and increased basal adrenocorticotropic hormone (ACTH) concurrent with no increase in basal corticosterone.

In conclusion, these studies revealed that adolescent inhalant abuse causes significant and persistent growth impairments. These have the potential to lead to further adverse health consequences. Body weight suppression has the potential to result in cognitive impairment, skeletal disorders, and increase the risk of an individual developing a metabolic disorder if weight is regained. Additionally, height impairment is associated with psychological issues such as poor self-image and social anxiety. Thus, these effects may contribute to long-term health consequences for individuals even if inhalant abuse ceases, and further research of these consequences relative to inhalant abuse would be beneficial. The adrenal insufficiency hypothesis requires testing in a clinical setting, but if confirmed, presents a significant health risk to those with a history of inhalant abuse, as the disorder can be life-threatening when undiagnosed and untreated. However, this may provide a target for future clinical research, which may improve the health outcomes of highly vulnerable individuals. Collectively, this thesis has utilised multiple methods both clinical and pre-clinical, in order to better understand the health consequences of adolescent inhalant abuse, and to provide translational outputs that will directly improve the detection and treatment of inhalant abuse in adolescents.

2 Rose Crossin (737900)

Declaration

I, Rose Crossin, declare that:

- This thesis comprises my original work towards the Doctor of Philosophy - Due acknowledgment has been made in the text to all other material used - The thesis is fewer than the maximum word limit in length, as approved by the Research Higher Degree Committee

Signed,

Rose Crossin

13 October 2018

3 Rose Crossin (737900)

Preface

Publication status of chapters References for Chapter 1, the Supplementary Section of Chapter 4, and Chapter 5 are reported in the Bibliography. All other references are contained within the Articles tabulated below.

Chapter Article title Status 1 General introduction and literature review Unpublished material not submitted for publication 2 Adolescent inhalant abuse leads to other drug Published by Australian and use and impaired growth; implications for New Zealand Journal of Public diagnosis Health on 23 October 2016 2 The persistence of growth impairments Published by Addiction associated with adolescent inhalant abuse Research and Theory on 15 following sustained abstinence June 2017 3 Growth changes after inhalant abuse and Published by Human and toluene exposure: a systematic review and Experimental Toxicology on 31 meta-analysis of human and animal studies July 2018 4 Adolescent inhalant abuse results in adrenal Accepted for publication in dysfunction and hypermetabolic phenotype Neuroendocrinology on 13 with persistent growth impairments September 2018 5 General discussion Unpublished material not submitted for publication App. 1 Altered body weight associated with substance Published by Addiction abuse: a look beyond food intake Research and Theory on 20 March 2018 App. 2 A proposed clinical cohort exploring the growth Unpublished material not and energy balance effects of adolescent submitted for publication inhalant abuse

Contribution of persons in multi-authored publications

Chapter 2 Sheree Cairney: assistance with data analysis, interpretation, and manuscript preparation.

Andrew Lawrence: assistance with data interpretation and manuscript preparation.

Jhodie Duncan: assistance with data analysis, interpretation, and manuscript preparation.

Chapter 3 Andrew Lawrence: assistance with data interpretation and manuscript preparation.

Zane Andrews: assistance with data interpretation and manuscript preparation.

Leonid Churilov: assistance with meta-analytic techniques, data interpretation and manuscript preparation.

Jhodie Duncan: assistance with data interpretation and manuscript preparation.

4 Rose Crossin (737900)

Chapter 4 Zane Andrews: assistance with experimental design, data interpretation, and manuscript preparation.

Natalie Sims: assistance with skeletal micro-computed tomography and data interpretation.

Terence Pang: assistance with adrenal histology and data interpretation.

Michael Mathai: assistance with faecal bomb calorimetry.

Jonathan Gooi: assistance with bone mechanical testing.

Aneta Stefanidis: assistance with metabolic cage experiments.

Brian Oldfield: assistance with metabolic cage experiments.

Andrew Lawrence: assistance with experimental design, data interpretation, and manuscript preparation.

Jhodie Duncan: assistance with experimental design, data interpretation, and manuscript preparation.

Appendix 1 Andrew Lawrence: assistance with manuscript preparation.

Zane Andrews: assistance with manuscript preparation.

Jhodie Duncan: assistance with manuscript preparation.

Acknowledgement of funding sources The research was supported by the Australian National Health and Medical Research Council (NHMRC) (940835), the Australian Department of Education and Training, from whom Rose Crossin received a scholarship through the Research Training Program, and the Victorian Government’s Operational Infrastructure Support Scheme.

5 Rose Crossin (737900)

Acknowledgements I acknowledge with gratitude my supervisors; Jhodie Duncan, Andy Lawrence and Zane Andrews for their wisdom, support, tolerance and for giving me the freedom to let me develop this project myself. I especially need to acknowledge Jhodie, who has had a couple of tough years herself, but has been an amazing supervisor, mentor, and now I hope I can say, friend. I am also grateful to Julie Bernhardt and Song Yao from my advisory committee for their advice and encouragement.

Many thanks to the addiction lab group at the Florey, who have helped with experiments, provided feedback, and made being here fun. Especially I thank Robyn Brown, Christina Perry, Nikki Chen, and Luba Lee-Kardashyn for their practical advice and support during the tough times.

I acknowledge the contribution of the communities in Arnhem Land, who participated in the inhalant abuse epidemiology studies and who made such positive change to their communities. I also acknowledge Dr Chris Burns, who undertook the majority of the initial data collection, as part of this own thesis.

I worked with an amazing group of collaborators in the last three years and I thank them all for their help and the inspiration that it gave me to work with scientists of their calibre. I also thank the Howard Florey Laboratories animal house staff and Florey histology service for their assistance during my animal experiments.

I acknowledge with love and gratitude my family and friends, who have been there for me during the last three years. Thanks Mum and Dad for teaching me from a young age that things of value do not come easily. Thanks also to my grandparents (Fred and Ivy, Bob and Heather), who wanted to see all of their grandchildren succeed and take up opportunities that they did not have themselves. Especially, thank you with all my heart to my husband Enda, whose practical and emotional support made this endeavour possible. I could not have done this without you.

Lastly, and most importantly, I thank my daughters Zara and Isla, for being the inspiration that kept me here. When I started my PhD they were 1 and 3, and there were plenty of people who tried to discourage me from this path. But I wanted to prove to my girls that family life and careers can co-exist, and so, my darlings, I want you to remember this. When people tell you that you shouldn’t take on great challenges, don’t get angry, don’t get despondent, don’t get offended, get working.

“Energy rightly applied and directed will accomplish anything” – Nellie Bly 1889

(journalist, explorer, badass feminist)

6 Rose Crossin (737900)

This thesis is dedicated to my daughters

Zara and Isla

Thank you for inspiring me and I hope that I can inspire you to follow your dreams and pursue greatness, with hard work, passion, and the support of all those who love you

(Dad and I have got your backs, always and forever)

7 Rose Crossin (737900)

Contents Abstract ...... 1 Declaration ...... 3 Preface ...... 4 Publication status of chapters ...... 4 Contribution of persons in multi-authored publications ...... 4 Chapter 2 ...... 4 Chapter 3 ...... 4 Chapter 4 ...... 5 Appendix 1...... 5 Acknowledgement of funding sources ...... 5 Acknowledgements ...... 6 List of tables, figures, and illustrations ...... 10 Chapter 1 - General Introduction and Literature Review ...... 11 Inhalant abuse ...... 11 Adolescent inhalant abuse ...... 12 The effects of inhalant abuse on food intake, body weight, and growth ...... 13 Energy balance ...... 14 Central and peripheral regulation of energy balance ...... 15 Effects of inhalants on energy intake ...... 18 Effects of inhalants on energy expenditure ...... 18 Effects of inhalants on adipose tissue ...... 19 Effects of inhalants on bone...... 20 Effects of inhalants on the HPA axis and adrenal gland ...... 20 Effects in abstinence and the potential for adult-onset disorders ...... 22 Issues when studying inhalant abuse ...... 22 Summary ...... 23 Research aims and hypothesis ...... 24 Chapter 2 – Growth changes arising from adolescent inhalant abuse in a human cohort ...... 25 Chapter 3 – A systematic review and meta-analysis on the growth impacts of inhalant abuse 39 Chapter 4 – The energy balance consequences of adolescent inhalant abuse and potential causal mechanisms ...... 71 Supplementary Data for Chapter 4 ...... 120 Prelude ...... 120

8 Rose Crossin (737900)

Methods ...... 121 Results ...... 122 Discussion ...... 125 Chapter 5 - General Discussion ...... 127 Adolescent inhalant abuse results in growth suppression and energy balance alterations . 127 The growth impairments arising from adolescent inhalant abuse persist into sustained abstinence ...... 130 Growth changes arising from adolescent inhalant abuse are not attributable to under- nutrition but may be caused by adrenal insufficiency ...... 131 Translational applications of this research ...... 134 Future work ...... 134 Summary and conclusions ...... 136 Bibliography ...... 138 Appendix 1 – Review on the effects of substance abuse on body weight ...... 147 Appendix 2 – A human cohort of adolescent inhalant users ...... 157 Introduction ...... 157 Methods ...... 157 Current status ...... 160 Preliminary findings ...... 160 Appendix 3 – Research Protocol for the human cohort study ...... 162 Appendix 4 – Questionnaire for the human cohort study ...... 183

9 Rose Crossin (737900)

List of tables, figures, and illustrations Note: this list only includes the figures that are not already published in manuscripts, to avoid duplicated figure numbering. As such, this list includes figures contained within Chapter 1 (general introduction) and supplementary section of Chapter 4.

Figure / Title Section Table number Figure 1 Past year inhalant use in Australian secondary school Chapter 1 students, by gender, based on data sourced from the 2014 Australian Secondary School Drug Survey Figure 2 The energy balance equation Chapter 1 Table 1 Central and peripheral regulation of energy balance Chapter 1 Figure 3 Blood glucose homeostasis Chapter 1 Figure 4 Diagnosis of adrenal insufficiency Chapter 1 Figure 5 Whole body thermographic imaging Chapter 4 supplement Figure 6 IL-1β in the liver and hypothalamus Chapter 4 supplement Figure 7 Energy balance following corticosterone replacement Chapter 4 supplement

10 Rose Crossin (737900)

Chapter 1 - General Introduction and Literature Review

Inhalant abuse Substance abuse or substance use disorders are defined as “chronic, relapsing brain diseases that are characterised by compulsive drug-seeking and use, despite harmful consequences” [1]. Inhalant abuse is a form of substance abuse and involves the intentional inhalation of vapours from household or industrial products, in order to create a feeling of euphoria and an altered mental state [2]. Products that are commonly abused as inhalants include petrol, spray paint, thinners, aerosols, and glue. These products are cheap, legal to purchase, and readily accessible, which is thought to add to their attractiveness as substances of abuse [3].

While inhaled products are not technically considered ‘drugs’, they share the reinforcing nature of many drugs of abuse, due to the presence of toluene, which is common to all these products. Toluene is a volatile solvent; a colourless, water-insoluble liquid that is classed as an aromatic hydrocarbon. Toluene is considered to underlie the addictive drive to continue abusing inhalants, based on its ability to generate a conditioned place preference in rats [4, 5], and preferential inhalation by humans of pure toluene compared with mixed products containing toluene [6]. Therefore, toluene is commonly used as an experimental pre-clinical model for inhalant abuse, although it should be noted that inhaled products contain several harmful compounds [7-9]. Thus, for the purpose of this thesis, the terms “inhalants” and “toluene” are used interchangeably, depending on whether reference is being made to clinical or pre-clinical studies.

Inhalant abuse is commonly known as “chroming”, “huffing”, or “bagging” as common self- administration methods are to inhale concentrated vapours from spray cans, off solvent- soaked cloth, or directly from a bag, respectively [10]. When administered in these ways, the concentration of inhaled toluene is estimated to range from 3,000-15,000 parts per million (ppm) [11]. Once inhaled, the acute effects of toluene occur within minutes; characterised by locomotor stimulation and feelings of euphoria and exhilaration [12, 13]. At concentrations greater than 3,000 ppm, central nervous system (CNS) depression occurs, with acute symptoms including slurred speech, disorientation, weakness, and finally sedation [12, 14]. At its most severe, and usually observed when concentrations exceed 10,000 ppm, CNS depression may progress into cardiopulmonary failure, coma or death [15]. After inhalation, toluene is absorbed across the lungs into the bloodstream before being rapidly metabolised and excreted as urinary hippuric acid [16, 17]. The rapid metabolism of toluene out of the bloodstream makes it particularly difficult to detect, and there are no established blood, urinary or saliva tests for diagnosing or identifying either acute inhalant intoxication or chronic inhalant abuse. While the abuse of inhalants generally occurs for a short duration, ranging from minutes to one hour, for the majority of individuals this behaviour is repeated over time and extends for periods greater than one year, such that exposure becomes chronic but intermittent in nature [18].

The effects of inhalants on the CNS, including cognition, have been relatively well studied [19- 21] and do not form the basis of this thesis. However, inhalants can also have a wide range of peripheral effects throughout the body, including on the renal, cardiovascular, and pulmonary systems [12]. With repeated exposure, toluene accumulates in lipid-rich organs such as the

11 Rose Crossin (737900) brain, adipose tissue, kidneys and adrenal glands [22], and as such, is likely to affect the function of these organs. Additionally, one of the most profound, but under-explored, consequences of inhalant abuse is its effect on food intake, body weight, and growth; especially if exposure occurs during adolescence, suggesting inhalant abuse results in energy balance dysfunction (discussed further beginning page 13) [23-27]. However, understanding of the effects of inhalants on peripheral organs and processes, including energy balance, remains limited. Furthermore, most studies have investigated the acute effects of inhalants at, or shortly following, exposure but have neglected to investigate the long-term effects, even if abuse of inhaled products ceases. Thus, understanding the effects of inhalant abuse on energy balance and growth, including the persistence of these effects in abstinence, form the overarching aim of this thesis. The following review will highlight the current understanding of the effects of inhalants on growth and energy balance.

It is also important to note that, as well as intentional inhalation, exposure to inhaled toluene can also occur in industrial settings (e.g. in factories where toluene is used as a solvent). Both the patterns of use and concentrations of inhaled toluene differentiate inhalant abuse from occupational exposure, with occupational exposure characterised by sustained contact to concentrations of toluene less than 200 ppm [28, 29]. While this thesis focusses on inhalant abuse, it is acknowledged that some of the data pertaining to the toxicology of toluene has been determined using occupational exposure studies and this should be interpreted with caution relative to the effects in the abuse setting; these studies are highlighted when discussed.

Adolescent inhalant abuse Adolescence is defined as the period between puberty and adulthood, typically between the ages of 11 and 21 years [30]. Inhalant abuse is predominantly an issue in young adolescent populations (Figure 1), with prevalence rates approximately equal between males and females [27]. Inhalants are one of the first and most common substances abused by the 12-13 age group in Australia, with 14% of 12 year-olds having misused inhalants in the previous year, and of those, approximately 20% have misused inhalants 10 or more times in that year [27, 31]. Furthermore, this pattern of early adolescent use is consistent globally [32]. Of further concern, a recent survey by the Australian government found that, overall, inhalant abuse in Australia is increasing with rates of recent inhalant use more than doubling between 2007- 2016 compared to a 4% reduction in recent alcohol and cigarette use over the same period [31]. Statistics such as these highlight the growing popularity of inhalants as a form of substance abuse and are likely to reflect the increasing rates of inhalant abuse in young adolescents worldwide. For example the most recent US Monitoring the Future Survey also reported a significant increase in recent use in Grade 8 students (aged 13-14) along with a decline in the perceived risk of inhalant use [33].

12 Rose Crossin (737900)

Past year inhalant use in 2014

10 Male

5 Female % used in past year past in % used 0 12 13 14 15 16 17 Age

Figure 1 – Past year inhalant use in Australian secondary school students, by gender, based on data sourced from the 2014 Australian Secondary School Drug Survey. The data highlight that, amongst all adolescents, the predominant age group misusing inhalants is young adolescents, with use decreasing as age increases. Figure adapted from data provided in [27].

As well as an association with adolescence, inhalant abuse is further associated with vulnerable and disadvantaged groups [34], including indigenous populations. Inhalant abuse (particularly petrol sniffing) remains a significant public health issue for Indigenous Australians [35], with abuse rates of up to 60% reported in some remote communities [36]. Furthermore inhalant abusers are disproportionately represented in the mental health, juvenile justice, and protective services systems [3]. Therefore, given that much of the epidemiological data on inhalant abuse comes from secondary school surveys, it is likely that inhalant abuse rates are actually under-reported in Australia and worldwide. A history of adolescent inhalant abuse also has an association with a range of other behavioural and psychological harms. Adolescents who misuse inhalants are more likely to develop other substance dependence issues [37, 38], experience major depression [38], and to experience suicidal ideation or attempt suicide [38, 39]. It is important to note, however, that the directionality of these relationships is not known; these may be outcomes of inhalant abuse or arise from the common background of vulnerability and social disadvantage in general.

The effects of inhalant abuse on food intake, body weight, and growth In humans, adolescent inhalant abuse is associated with disordered eating [26], weight loss [24], and emaciation [40]. Inhalant abuse in both males and females has been associated with fasting, purging, and other behaviours associated with attempts to lose weight [26], with previous studies suggesting that inhalants may be used as a means of suppressing appetite and thus body weight [26, 41]. These findings are consistent with pre-clinical studies using rats, with adolescent and adult toluene exposure (repeated exposures, including intraperitoneal and inhalation administration, concentrations ranging from 3,000-16,000 ppm) associated with reduced food intake, decreased weight gain, and reduced fat deposition [25, 42, 43]. Indeed, the effect of inhalant abuse on body weight is so well known that it is listed as a warning sign of inhalant abuse [44].

13 Rose Crossin (737900)

In addition to effects on food intake and body weight, exposure to toluene is associated with impaired skeletal growth. In weanling rats, chronic toluene exposure at concentrations analogous to inhalant abuse (6,200 ppm, 15 mins/hour, 8 h/day, for 23 weeks) has been shown to impair skeletal growth, with a significant reduction in rump width and variable reduction in torso length [9]. Skeletal variations have also been shown in the offspring of rats where there was maternal exposure to toluene (up to 2,000 ppm, 6 h/day, 7 d/week for 80 days premating and 15 days of mating), as well as decreased body weight in the offspring [45, 46]. In humans, regular inhalant abuse (spray paint inhalation) during pregnancy is also associated with neonatal skeletal malformations and growth impairment [47].

As well as a reduction in overall skeletal growth, inhalant abuse is also associated with reduced bone density. A study of adolescent males with a history of chronic glue sniffing (mean duration of abuse 3.2 years) found that bone mineral density was significantly reduced in comparison to an age-matched non-sniffing group [48]. In this study, the mean height of the glue sniffing group was reduced in comparison to controls, though this result was not significant [48]. However, no other human studies on the effects of adolescent inhalant abuse on skeletal growth could be found. Collectively, these studies suggest that toluene exposure can lead to reductions in food intake, body weight suppression, and impaired growth. However, both the mechanisms underlying these consequences of adolescent inhalant abuse, and the persistence and long-term health effects of these changes in abstinence are currently unknown.

The impact of inhalants on food intake, body weight, and growth is of particular importance given the association between inhalant abuse and adolescence. During adolescence, critical maturational processes occur, including an intensive period of growth known as the adolescent growth spurt [49], where skeletal structures grow and mature [50]. Indeed, in humans, over half of the peak bone mass is attained during the adolescent growth spurt [50, 51]. However, the adolescent growth spurt can be disrupted by factors such as malnutrition [52] or exposure to estrogenic chemicals [53]. Growth impairments can potentially recover if the harmful factor is removed, but if exposure occurs during adolescence usually only partial recovery is possible [54]. Thus, any disruption to growth during adolescence has the potential to cause long-term effects on growth patterns [55]. Given that the highest rates of inhalant abuse are occurring during a critical maturational period, it highlights the importance of studying the impacts of inhalant abuse specifically in adolescents, including the long-term effects even if exposure ceases.

Energy balance Food intake, body weight, and growth are intrinsically linked and can be represented via the energy balance equation (Figure 2), with body weight a function of the energy stored in the body. Thus, an imbalance in this equation can lead to changes in body weight. Energy intake is primarily driven by food intake and moderated by absorption and excretion. Energy expenditure includes basal metabolic rate, physical activity and thermogenesis (subdivided into cold-induced thermogenesis and the thermic effect of food). Energy balance is controlled by the processes of metabolism, which encompasses a system that regulates eating behaviour and nutrient absorption, as well as the storage and release of energy – termed nutrient partitioning. Body weight will remain stable when there is a balance between energy intake,

14 Rose Crossin (737900) storage, and expenditure, known as energy homeostasis. Energy homeostasis is the balance between the pathways that stimulate food intake and promote weight gain, and the pathways that reduce food intake and promote weight loss [56]. The maintenance of body weight is buffered, i.e. a decrease in energy intake results in decreased expenditure and vice versa, therefore, any changes in body weight occur by overwhelming the buffering potential of homeostatic systems [57].

Figure 2 – The energy balance equation. Original figure by R Crossin.

A negative energy balance occurs when energy intake is less than energy expended, and this shortfall is made up by a release of energy from body stores (e.g. adipose tissue), leading to reduced body weight. If the difference between energy intake and energy expenditure is positive, then energy is stored and body weight increases (positive energy balance). Physiological processes strongly defend reductions in body weight arising from a negative energy balance, ultimately leading to weight regain [58]. In contrast, a positive energy balance that leads to weight gain does not initiate the same physiological defence, or resistance to that physiological defence develops, and weight gain can continue into obesity. This new increased body weight is subsequently defended [59]. Energy balance becomes particularly important in periods of rapid growth, such as the adolescent growth spurt, with energy intake one of the major determinants of growth [60, 61]. However, the effect of energy intake on growth is predominantly indirect, with growth impairments due to under-nutrition primarily attributable to endocrine changes such as increased circulating cortisol [62-64].

Central and peripheral regulation of energy balance Energy homeostasis, and thus body weight, is both centrally and peripherally regulated through the processes of metabolism [56], as described in Table 1. In order to understand the potential mechanisms behind inhalant-induced energy balance dysfunction, it is important to summarise the complex neuroendocrine system that regulates energy balance. The main short term fuel source of the body is glucose, which is maintained within highly-regulated levels [65]. Excess glucose in the blood is stored in the liver and muscles as glycogen, in response to insulin release, via the process of glycogenesis. Glycogen is then available for release in response to falling blood glucose levels, via the process of glycogenolysis, which is stimulated by the Hypothalamic-Pituitary-Adrenal (HPA) axis. Hypoglycaemia (low blood glucose) is a life- threatening condition, and thus the response is rapid and involves both central and peripheral regulation [65]. This counter-regulatory response activates both feeding behaviour and the release of glycogen stores, in order to rapidly normalise blood glucose levels [66], summarised in Figure 3.

15 Rose Crossin (737900)

Table 1 – central and peripheral regulation of energy balance

Key metabolic organ Compound Function Pancreas Insulin Decreases glucose levels in the blood and promotes lipogenesis (the storage of energy as adipose tissue) [67] Pancreas Amylin Slows gastric emptying and reduces appetite [68] Liver Insulin Responds to insulin secreted by the pancreas by taking up glucose from the blood [69] Kidney Insulin Clears and degrades circulating insulin [70] Gut Ghrelin Increases hunger and initiates feeding [71] Gut Mechanical Stretch receptors signal satiety to the brain via processes the vagus nerve [72] Gut Peptide YY (PYY) Slows gastric emptying and reduces appetite [73] Adipose tissue Leptin Opposes the actions of ghrelin, by sending an adiposity signal and inhibiting hunger [74] and also inhibits lipogenesis [75] Brain Receptors for gut Signals converge primarily on the brain stem hormones and hypothalamus, which are the primary brain regions involved in maintaining energy homeostasis [76] Brain Neuropeptide Y Increases food intake [77] (NPY) Adrenal gland Adrenalin and Responds to hypoglycaemia by inhibiting cortisol insulin release, and activates glycogenolysis and gluconeogenesis [65] Bone Osteocalcin Increases insulin production and sensitivity, and enhances glucose utilisation [78]

Insulin release also promotes the conversion of excess glucose to fatty acids, via the process of lipogenesis, which is then stored in adipose tissue [79]. Adipose tissue is thus the major site of long-term energy storage, noting however that some fatty acids can be stored ectopically in organs such as the liver. When energy homeostasis is achieved, the amount of energy stored as adipose tissue is stable [76], but during a negative energy balance, energy can be released by the breakdown of adipose stores via the process of lipolysis. The lipostatic model of energy balance states that food intake is not tied to the immediate energy needs, but rather it is regulated by signals that are proportional to the size of the body’s adipose stores [80]. Thus, weight loss caused by the depletion of energy stored as adipose tissue should increase food consumption. According to the lipostatic model, in response to weight loss that arises from insufficient energy intake, levels of the regulatory hormones insulin and leptin will decrease, which activates pathways that stimulate appetite and promote weight gain [56].

16 Rose Crossin (737900)

Figure 3 – Blood glucose homeostasis. Maintenance of blood glucose levels, whereby excess blood glucose is stored in tissue, and released via counter-regulatory responses to hypoglycaemia, image adapted from [81]

When metabolic processes are functioning correctly, feedback mechanisms correct disruptions to energy homeostasis, with both the hypothalamus and brain stem being key regions involved in sensing both short and long term changes in energy balance. The response to a negative energy balance is firstly the promotion of appetite / feeding behaviour, and the utilisation of glycogen stores (Figure 3), and when these are exhausted, the utilisation of adipose tissue stores, and finally the breakdown of lean muscle mass occurs. However, as described, adolescent inhalant abuse is associated with both decreased food intake, impaired weight gain, and fasting hypoglycaemia [23], which suggests that energy homeostasis is not being achieved and the feedback mechanisms, which should correct a negative energy balance, are dysfunctional.

Toluene has a complex pathway through the body. As previously discussed, once inhaled, toluene crosses rapidly from the lungs into the bloodstream, before being metabolised by the liver and excreted by the kidneys [16, 17]. With repeated exposure, toluene can accumulate in organs, primarily the brain, adipose tissue, and adrenal glands [22, 82]. As shown in Table 1, these are key regions for metabolic processes. Toluene can also cross [83] and disrupt [84] the blood-brain barrier and has a comparatively high accumulation in the brainstem and hypothalamus [85]. The brainstem and hypothalamus play a critical role in control of feeding and metabolism [76, 86], which raises the potential that inhalant-induced metabolic dysfunction may have both centrally and peripherally mediated components.

17 Rose Crossin (737900)

The symptoms of inhalant abuse (i.e. reduced food intake, body weight suppression, and impaired growth) are suggestive of a negative energy balance. However, it is not clear which components of the energy balance equation are altered, or how the central and peripheral regulation of energy balance may be being affected. Theoretically, inhalant abuse could reduce energy intake via a direct effect on appetite, or by increasing energy expenditure via changes to physical activity, basal metabolic rate, or thermogenesis. Inhalant abuse may also disrupt the metabolic signalling that would promote appetite and weight gain in response to a negative energy balance. Alternatively, exposure to inhalants could also be directly affecting body composition through effects on the absorption or deposition of fat, or skeletal composition, which may explain the observed growth impairments [9]. Each of these aspects of energy balance will be expanded upon below, as they are key components to explore in order to understand the energy balance and growth effects of inhalant abuse.

Effects of inhalants on energy intake In humans there is evidence that incidental inhalation of petrol fumes may have an acute anorectic effect, and the study’s authors hypothesised that in a food-constrained environment, a desire to suppress feelings of hunger could be a driver for intentional petrol sniffing behaviour [87]. These data confirmed previous anecdotal reports that petrol sniffing in Australian Indigenous communities is partly driven by a desire to suppress appetite [41], although the relationship between appetite suppression and inhalant abuse among street children in India is less consistent [88]. In adult rats, toluene-induced anorexia (intraperitoneal administration, 1.3 mL/kg/day, for 4 days) has been attributed in part to the observed decrease in the hypothalamic expression of Neuropeptide Y (NPY), a peptide which stimulates food intake [25]. Furthermore, energy intake is moderated by absorption of nutrients occurring in the gut, and disruptions to absorption are known to be associated with other types of substance abuse [89]; however, nutrient absorption across the gut following toluene exposure or inhalant abuse has never been assessed. This is a potentially significant knowledge gap, given that inhalant users frequently report gastrointestinal disturbances, including nausea, vomiting and diarrhoea [12, 90] and the potential effects on the gut microbiome have never been assessed in relation to inhalant abuse.

There is also evidence to suggest that inhalant abuse can impair metabolic signalling, because the normal responses to negative energy balance do not occur. Our lab has shown that food intake is reduced in rodents following exposure to toluene at abuse concentrations, with animals exposed to toluene consuming fewer kilo-calories than air-exposed animals at the same body weight [23]. This finding suggests that there is impairment to the metabolic feedback processes, which should act to restore homeostasis in response to insufficient energy intake. In rats, adolescent exposure to toluene (10,000 ppm, 1 h/day, 3 d/week, for 4 weeks) has been shown to alter levels of gut and pancreatic hormones, including insulin, amylin, and Peptide YY (PYY), as well as decrease hypothalamic leptin receptor mRNA expression [23]. Importantly, decreased amylin and PYY levels should act to increase appetite and food intake, but this response does not occur, supporting the idea of impaired metabolic signalling.

Effects of inhalants on energy expenditure Energy expenditure in an animal model of inhalant abuse has never been explicitly tested, but studies suggest that exposure to toluene affects components of energy expenditure (i.e.

18 Rose Crossin (737900) physical activity, basal metabolic rate, and thermogenesis). Toluene exposure in adolescents acutely increases locomotor activity (exposure by inhalation at 4,000 ppm) [7], although this effect is transient and the response to toluene is bi-phasic and locomotor activity is reduced at higher concentrations [91]. In relation to basal metabolic rate, a human study of occupational exposure to toluene found that chronic exposure (less than 100 ppm) was associated with increased food intake, though this was not reflected in an increased body mass index (BMI) [92]. The authors hypothesised that toluene exposure resulted in a higher basal metabolic rate, though this was not tested [92]. In contrast, animal studies (pregnant female rats, exposed to toluene by inhalation at 8,000, 12,000 and 16,000 ppm twice daily for 14 days) showed that pre-natal exposure to toluene resulted in reductions in metabolic rate and energy expenditure in the offspring, unfortunately this was not also assessed in the toluene-exposed individuals [42]. These studies also did not assess thermogenesis, in addition to physical activity or metabolic rate, but one study utilising mice (inhaled toluene at 2,000, 4,000 and 6,000 ppm once for 60 mins) found that toluene acutely reduced core temperature in a dose dependent manner [93]. Conversely, it is interesting to note that many inhalant users report a persistent low-grade fever [94] though a mechanism for this remains unknown.

Effects of inhalants on adipose tissue Adipose tissue is the major site of long-term energy storage in the body, and changes to adiposity will be reflected in changes to body weight. Therefore, adipose tissue is a key organ to consider in relation to adolescent inhalant abuse. In our laboratory’s rodent model of adolescent inhalant abuse (10,000 ppm, 1 h/day, 3 d/week, for 4 weeks), body composition was altered by toluene exposure. Male rats exposed to toluene had significantly reduced adiposity but no difference in lean mass or total water [23]. Interestingly, in offspring that had prenatal exposure to toluene (0-16,000 ppm, 15 mins twice/day, gestational days 8-20), body weight was also reduced, but toluene significantly increased adiposity in a dose-dependent way in male, but not female, offspring [42]. The increased adiposity was attributed to a compensatory developmental programming effect, whereby low birth weight alters hormonal and metabolic programming to maximise postnatal survival, and subsequently results in increased adiposity [95]. Furthermore, when there is a negative energy balance caused by insufficient food intake, adipose stores are utilised by the body to provide energy. Thus, the observed decrease in adiposity [23] associated with exposure to toluene may not be unexpected. However, alternative hypotheses as to why toluene reduces adipose levels are that it affects the absorption of ingested fats within the digestive tract, or the ability to depose lipids in adipose tissue, such as that observed in pancreatic disorders [96]. For example, both leptin and insulin play a crucial role in regulating lipogenesis [79] and therefore the observed disruptions to hormones involved in energy balance signalling, including insulin [23], may impair the deposition of lipids in adipose tissue. While the studies discussed above show that exposure to toluene has an effect on body composition, particularly adiposity [23], it has yet to be determined whether toluene has a direct effect on body composition, or if these changes are an indirect effect arising as a result of decreased food intake leading to a negative energy balance.

19 Rose Crossin (737900)

Effects of inhalants on bone The reduction in weight gain arising from toluene exposure may not only be due to reduced adiposity, but also due to skeletal growth impairments. Human studies investigating the effects of toluene on skeletal growth are extremely limited and therefore no firm conclusion can be currently reached on the acute and chronic impacts of adolescent inhalant abuse. However, alcoholism is also associated with impaired bone function and subsequent increased osteoporosis and fracture risk [97, 98]. Given that alcohol has a known metabolic effect [99] and similar modes of action to toluene [100, 101], this would suggest that the skeletal impacts of inhalant abuse warrant further study. Anorexia nervosa (AN) can provide another biologically relevant model for skeletal function, with symptoms of reduced food intake and fat deposition, occurring predominantly in adolescence [102], as has been observed in inhalant abusers. Impaired skeletal growth and function is a common symptom of AN [103], which has been associated with low leptin levels [104]; a common feature of AN [105]. Of major concern, the loss of bone density that is observed in adolescents with a history of AN does not resolve, even years after weight restoration [106, 107].

There are a number of potential mechanisms that may lead to impaired skeletal growth arising from adolescent inhalant abuse. Male adolescent mice exposed to toluene (300 ppm, 6 h/day, for 8 weeks) have been shown to have significantly reduced bone mineral density compared to controls [108]. As bone metabolism occurs as a continuous cycle, whereby osteoclasts resorb existing bone and osteoblasts then lay down replacement bone, the authors of this paper hypothesised that toluene may be disrupting both of these processes [108]. The metabolic perturbations that are observed in inhalant abuse could also be involved in the reported skeletal growth impairments. Leptin has a direct effect on bone cell function and acts to reduce bone fragility through anabolic effects on osteoblasts and the reduction of osteoclast formation [109]. Amylin increases bone mass by inhibiting bone resorption [110]. Insulin has an indirect effect on bone function, and increased circulating insulin is associated with increased bone density [111]. All of these hormones are affected following exposure to toluene [23, 112].

Bone itself can also exert reciprocal effects upon metabolism via osteocalcin, which increases insulin production and sensitivity, and enhances glucose utilisation [78, 113]. Osteocalcin can exert effects on both pancreatic β cells and adipocytes, making it a key component of metabolic regulation [114]. In combination, these developments led to the conclusion that there is reciprocal regulation between energy metabolism and bone metabolism [115-117]. Furthermore, this is mediated via the hypothalamus [118] which is known to be affected by exposure to toluene [85]. Given the metabolic functions of bone, any investigation into the metabolic disruptions caused by inhalant abuse should therefore also give consideration to the skeletal system. Furthermore, data from AN models suggest that if inhalant abuse does have a skeletal effect, it will be important to study the chronic impacts, as it is likely that the effects will persist in abstinence.

Effects of inhalants on the HPA axis and adrenal gland There is evidence to suggest that inhalant abuse may have an effect on the HPA axis, though the nature of this relationship is equivocal. Some studies have associated inhalant abuse with an up-regulation in HPA axis activity and associated behaviours [119-121], whereas others

20 Rose Crossin (737900) have found a blunted HPA response following inhalant abuse [122, 123]. It is also possible that there is a direct effect of toluene on the adrenal gland, which is where the hormones cortisol and adrenaline are produced as part of the stress response, in the zona fasiculata of the cortex and the medulla, respectively. The adrenal gland is vulnerable to toxicological injury due to high vascularity and fat content [124], particularly by aromatic hydrocarbons such as toluene [125]. Multiple pre-clinical studies have found that adrenal hypertrophy is evident following inhalant abuse [23, 126, 127], though only one clinical case study has associated inhalant abuse with an adrenal effect [128].

Adrenal hypertrophy can be an outcome of chronic stress or of adrenal insufficiency [129]. Adrenal insufficiency is a failure to produce the stress hormone cortisol by the cortex of the adrenal gland, either due to an issue in the adrenal gland (primary adrenal insufficiency) or impairment upstream in the HPA axis (secondary or tertiary adrenal insufficiency) [130]. Chronic stress is characterised by adrenal hypertrophy with elevated levels of cortisol, whereas adrenal insufficiency is characterised by adrenal hypertrophy with low or unchanged levels of cortisol, which is then further located within the HPA axis by the relative levels of the other HPA axis signalling hormones adrenocorticotropic hormone (ACTH) and corticotropic-releasing hormone (CRH) [129, 130], as shown in Figure 4.

Figure 4 – Diagnosis of adrenal insufficiency. The differentiation between chronic stress and adrenal insufficiency, and the localisation of adrenal insufficiency within the HPA axis by the relative quantities of cortisol (Cort), adrenocorticotropic hormone (ACTH) and corticotropic-releasing hormone (CRH), and responses to an insulin tolerance test (ITT). Original figured by R Crossin adapted from data in [129, 130].

21 Rose Crossin (737900)

When the adrenal gland has been directly investigated in relation to inhalant abuse, studies have found adrenal hypertrophy, particularly in the zona fasiculata of the adrenal cortex where cortisol (corticosterone in rats) is produced, and low/unchanged levels of basal corticosterone with high levels of basal ACTH [126, 127]. It is interesting to note that although these findings are diagnostically consistent with primary adrenal insufficiency [130], in both studies the authors attributed their findings to chronic stress [126, 127]. Adrenal insufficiency is a chronic disorder with symptoms of decreased food intake, impaired body weight and growth, non- diabetic fasting hypoglycaemia, persistent low-grade fever, fatigue, and gastrointestinal disturbances [130]. These symptoms are consistent with what has been observed following inhalant abuse [23, 24, 26, 90, 94, 131], though it is acknowledged that they also fit a range of other disorders. Nevertheless, there is potential that HPA axis dysfunction, particularly adrenal insufficiency, may underlie the growth and energy balance consequences of inhalant abuse.

Effects in abstinence and the potential for adult-onset disorders Pilot data from our laboratory suggest that some of the energy balance effects arising from adolescent inhalant abuse, including reduced body weight, persist into abstinence [23]. Severe and chronic weight suppression is associated with cognitive impairments [132] and health effects, including kidney and bone damage [133]. Additionally, growth impairments are associated with adverse psychological consequences, including social anxiety and poor self- image [134, 135]. Conversely, weight regain after a period of weight suppression is associated with increased central adiposity and insulin resistance, both of which are risk factors for the development of metabolic syndrome, obesity, and type-2 diabetes whereby tissues become resistant to insulin [58, 136]. It is important to note, however, that it is currently unknown whether weight suppression arising from adolescent inhalant abuse ultimately resolves. Nevertheless, adolescent inhalant abuse has the potential to initiate a sequence of long-term metabolic alterations and thus lead to a chronic disease burden for individuals, even if exposure ceases. However, the nature of that disease burden remains unknown and knowledge on the medical consequences of inhalant abuse in abstinence is scarce.

Current treatments for substance abuse focus primarily on cessation of use, and energy balance impacts are given limited treatment consideration beyond nutritional counselling in abstinence [137, 138]. Energy balance dysfunction is an under-explored aspect of chronic drug use, but inhalant abuse is not the only drug of abuse that has an energy balance impact. Nicotine, alcohol, methamphetamine, and opioids have all been linked to metabolic dysfunction [99, 139-141]. Furthermore, metabolic symptoms are relatively well reported in relation to substance abuse, including weight loss during active drug use, and weight gain in abstinence (see for review [142] and [143] provided as Appendix 1). Thus, understanding the link between adolescent inhalant abuse and energy balance dysfunction may provide important knowledge on the long-term consequences of both inhalants and other drug use and lead to treatment strategies that extend beyond drug use cessation.

Issues when studying inhalant abuse Understanding of the potential effects of inhalant abuse is constrained by limited data, and the diversity of exposure models and experimental subjects used. When interpreting human data, some caution must be used due to the confounds of other substance abuse [18]. This is particularly relevant as inhalant abuse is associated with an increased frequency of other drug

22 Rose Crossin (737900) use [144]. For example, the metabolism of toluene can be inhibited by the concurrent consumption of alcohol, which adds a confound to the understanding of whole-body effects [145]. Additionally, laboratory studies into the acute effects of inhalant abuse in humans are generally conducted in adults, despite adolescents being the predominant population abusing inhalants. This is important as the response of adolescents to drugs of abuse is not the same as that of adults [146, 147] and this also holds true for inhalant abuse. For example, adolescents have a decreased sensitivity to the locomotor effects of toluene in models of chronic toluene exposure between 2,000 and 16,000 ppm [7, 91] and interestingly, sensitivity continues to increase throughout adulthood, with oldest rats having the greatest sensitivity to the effects of toluene [148]. However, while adolescents are less sensitive to psychomotor effects of toluene, they are more vulnerable to its toxic properties [149]. Additionally, adolescents take longer than adults to recover from the acute intoxicating effects of toluene (single exposure at 5,000 ppm for 15 or 30 mins) [150]. Therefore it is difficult to directly extract data from studies using adult participants and relate the outcomes to adolescent inhalant users. Finally, the majority of human studies on the effects of toluene exposure have been predominantly conducted using low concentrations of toluene. In some cases this is because researchers were investigating occupational exposure to toluene and so the concentrations used were less than 200 ppm [28, 29], however, when deliberately inhaled, concentrations of toluene can be up to 15,000 ppm [11].

Therefore, it is important to specifically study adolescents in relation to inhalant abuse because:

- they are the predominant population abusing inhalants; - they need to inhale higher concentrations to achieve the same effect as adults, due to decreased sensitivity; - they are more vulnerable to the toxic properties of toluene; - they are abusing inhalants during a critical growth and maturational period; and - the long-term effects, even if exposure should cease, have not been explored.

Summary The cognitive and neurological consequences of inhalant abuse have dominated inhalant abuse research. Long-term cognitive impairment arising from intentional inhalation of inhalants containing toluene is well reported in both human [19-21] and animal [8] studies. In extended abstinence, these effects generally improve and in some cases normalise completely [43, 151, 152]. Although the primary research focus of inhalant abuse has historically been the brain, inhalant abuse is also associated with a range of symptoms that occur throughout the body. The inhalation of toluene is associated with gastrointestinal complaints, including nausea and vomiting [90], as well as more widespread issues such as fatigue [131] and persistent low-grade fever [94]. As discussed, inhalant abuse is associated with decreased food intake, reduced body weight, fasting hypoglycaemia, altered levels of gut hormones, as well as a preliminary link to impaired skeletal growth. Collectively, these symptoms suggest widespread energy balance dysfunction, which can impact growth, and may lead to an increased risk of chronic metabolic disorders e.g. obesity and type-2 diabetes in adulthood, even if abuse ceases. These potential energy balance and growth impacts are not well-

23 Rose Crossin (737900) researched, either acutely or in abstinence, and it is unknown what the long-term health consequences of inhalant-induced energy balance dysfunction may be.

Research aims and hypothesis Based on the information presented above, the aims of this research were to:

- characterise the energy balance and growth consequences of adolescent inhalant abuse; - understand how these effects may persist into abstinence; and - elucidate the mechanisms underlying these changes.

The hypothesis was that adolescent inhalant abuse will cause energy balance dysfunction and growth impairments that will persist into abstinence.

This research specifically addressed the following questions:

1. How does adolescent inhalant abuse affect the energy balance equation and metabolic signalling of energy balance? 2. Do the energy balance and growth consequences of adolescent inhalant abuse resolve or persist in abstinence? 3. Are the observed changes solely due to changes in food intake i.e. inhalant-induced under-nutrition, or are there additional mechanisms?

In order to fulfil these research aims, epidemiological data from an Australian Indigenous population who chronically abused inhalants during adolescence, followed by sustained abstinence of up to 15 years [144, 151, 152] was analysed. A systematic review and meta- analysis on the growth effects of inhalant abuse and toluene exposure was conducted, in order to determine the consistency of growth effects across studies and potential moderators of growth effects. The epidemiological data was analysed in conjunction with an established rodent model of adolescent inhalant abuse (10,000 ppm, 1 h/day, 3 d/week, for 4 weeks). This model mimics the chronic and intermittent nature of abuse seen in humans and has resulted in behavioural outcomes similar to those reported in human abusers [8]. The validity of using rats for inhalant studies is further supported by the fact that rats and humans have an equivalent sensitivity to toluene, and thus, the effects of toluene in humans may be estimated using rat models [153].

For the rodent studies, toluene was used in isolation as it is the common volatile solvent in abused products and is preferentially inhaled when available [6]. The combined power of epidemiological data with a well-established rodent model allowed characterisation of the energy balance and growth changes specifically associated with the inhalation of toluene and the mechanisms underlying these changes; providing a unique insight into the long-term consequences of adolescent inhalant abuse. Thus, this project addresses an important knowledge gap about the long-term energy balance and growth impacts of adolescent inhalant abuse, including whether these changes persist even if exposure ceases. Ultimately it is hoped that this information will further elucidate the long-term health risks associated with adolescent inhalant abuse and assist clinicians to treat individuals with a history of inhalant abuse, subsequently improving their quality of life.

24 Rose Crossin (737900)

Chapter 2 – Growth changes arising from adolescent inhalant abuse in a human cohort Chapter 2 of this thesis is based on a retrospective analysis of data collected from a cohort of Indigenous males in Australia, who sniffed petrol during adolescence. The aim of this study was to describe the growth changes arising from inhalant abuse, both during active inhalant abuse and after sustained abstinence. The rationale for this aim was two-fold. Firstly, weight impairments arising from adolescent inhalant abuse were relatively well-described in both clinical and pre-clinical studies, however given the adolescent growth spurt is characterised by dramatic increases to both weight and height, it was reasonable to hypothesise that inhalant abuse may also deleteriously affect height. Yet very few studies had investigated the effects of inhalant abuse on linear growth, with inconsistent findings between clinical and pre-clinical studies. This was described in the first paper: Adolescent inhalant abuse leads to other drug use and impaired growth; implications for diagnosis.

Additionally, data from pre-clinical studies had suggested that the weight impairments from inhalant abuse persisted into sustained abstinence; however, this had not been investigated in a human cohort of inhalant users. This was investigated in a sub-set of the same cohort of inhalant users, who had been abstinence from inhalant abuse for two years. This was described in the second paper: The persistence of growth impairments associated with adolescent inhalant abuse following sustained abstinence.

Paper citations:

Crossin, R., Cairney, S., Lawrence, A. J., & Duncan, J. R. (2017). Adolescent inhalant abuse leads to other drug use and impaired growth; implications for diagnosis. Australian and New Zealand Journal of Public Health, 41(1), 99-104.

Crossin, R., Cairney, S., Lawrence, A. J., & Duncan, J. R. (2018). The persistence of growth impairments associated with adolescent inhalant abuse following sustained abstinence. Addiction Research & Theory, 26(3), 183-186.

25 Adolescent inhalant abuse leads to other drug use and impaired growth; implications for diagnosis

Rose Crossin,1 Sheree Cairney,2,3 Andrew J. Lawrence,1,4 Jhodie R. Duncan1,4

nhalant abuse involves the intentional Abstract inhalation of vapours from products, such as petrol and glue, to create a feeling Objective: Abuse of inhalants containing the volatile solvent toluene is a significant public I 1 of euphoria and an altered mental state. health issue, especially for adolescent and Indigenous communities. Adolescent inhalant These products are cheap, legal and readily abuse can lead to chronic health issues and may initiate a trajectory towards further drug accessible, adding to their attractiveness use. Identification of at-risk individuals is difficult and diagnostic tools are limited primarily to as substances of abuse. A recent stratified measurement of serum toluene. Our objective was to identify the effects of adolescent inhalant and randomised selection survey of more abuse on subsequent drug use and growth parameters, and to test the predictive power of than 26,000 people aged 12 years and older growth parameters as a diagnostic measure for inhalant abuse. from across Australia, by the Australian Methods: We retrospectively analysed drug use and growth data from 118 Indigenous males; government, found a 23% increase in inhalant 86 chronically sniffed petrol as adolescents. abuse,2 highlighting its growing popularity. Results: Petrol sniffing was the earliest drug used (mean 13 years) and increased the likelihood In Australia inhalant abuse, especially petrol and earlier use of other drugs. Petrol sniffing significantly impaired height and weight and was sniffing, remains a significant issue for associated with meeting ‘failure to thrive’ criteria; growth diagnostically out-performed serum Indigenous communities with rates up to toluene. 60% in some remote communities.3 However Conclusions: Adolescent inhalant abuse increases the risk for subsequent and earlier drug in recent years, inhalant abuse has also use. It also impairs growth such that individuals meet ‘failure to thrive’ criteria, representing an become prevalent among young adolescents, improved diagnostic model for inhalant abuse. irrespective of ethnicity, as determined by both the household drug survey and Implications for Public Health: Improved diagnosis of adolescent inhalant abuse may lead to a specialised survey of 25,000 Australian earlier detection and enhanced health outcomes. secondary school students aged 12-17.2,4 This Key words: petrol sniffing, height, weight, Failure to Thrive is concerning as adolescence encompasses a critical period of maturation, which for a diagnosis to be made, an individual metabolised, which limits its diagnostic can be disrupted by exposure to harmful needs to be in the care of a psychological capabilities.8 As creatine kinase is elevated in substances.5 Inhalant abuse during this or health professional. Therefore, the first response to petrol sniffing it has also been period is associated with long-term adverse level of recognition of inhalant abuse often suggested as a potential diagnostic tool, health outcomes affecting both central and relies on a high index of suspicion from though its relevance to inhalants other than peripheral processes, including cognitive parents/guardians, teachers or health care petrol remains unknown.9 For most users, dysfunction, renal tubular acidosis, hepatic workers before referral to an appropriate the pattern of inhalant abuse is chronic but dysfunction and bone marrow suppression.6 health professional can occur, which intermittent,10 which highlights the need for Despite its prevalence and associated makes diagnosis problematic. Additionally, diagnostic markers that are more responsive harms, inhalant abuse remains a poorly inhalant abuse rates are typically highest in to sustained use. Chronic inhalant abuse recognised problem.6 There are reported populations of vulnerable youth, which have is associated with long-term health and warning signs to identify inhalant abuse, but reduced contact with the primary health psychological harm including increased these refer predominantly to acute effects, care system.7 Toluene is the primary volatile use of other drugs,11 cognitive decline and which are highly transient.6 The Diagnostic solvent in abused products, and therefore brain injury12 and adverse psychological and Statistical Manual of Mental Disorders serum toluene can be used as an acute consequences including suicide.13 Inhalant IV contains criteria for inhalant abuse, but diagnostic test. However, toluene is rapidly abuse, especially during adolescence,

1. Behavioural Neuroscience, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Victoria 2. Centre for Remote Health, Flinders University, Northern Territory 3. Ninti One Limited, Northern Territory 4. Department of Anatomy and Neuroscience, University of Melbourne, Victoria Correspondence to: Mrs Rose Crossin, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, Vic 3010; e-mail: [email protected] Submitted: March 2016; Revision requested: May 2016; Accepted: June 2016 The authors have stated they have no conflict of interest. Aust NZ J Public Health. 2016; Online; doi: 10.1111/1753-6405.12595

can also affect peripheral body systems, 2274). Data were collected from 1992 to 12 hours of their most recent petrol sniffing including those that regulate metabolism 1994, from 118 Indigenous males from two activity, which was verbally confirmed with and growth.14-16 Thus a more robust remote Indigenous communities in northern participants by the researcher during the diagnostic method, which can be applied Australia. Eighty-six participants had a history interview. Samples for blood hydrocarbons to chronic inhalant abuse, may increase of petrol sniffing as adolescents (both leaded (e.g. toluene) were collected into 7mL sterile, detection and provide an opportunity for and unleaded petrol) and 32 participants EDTA Vacutainers; all other haematological intervention strategies, to limit adverse from the same communities with no history measures collected into sterile (EDTA, heparin neuropsychological, metabolic and growth of petrol sniffing acted as age-matched clotted Vacutainers). All haematological outcomes. controls. Informed written consent was given variables and analysis methods are shown in Profiles of subsequent drug use and growth by all participants, who could withdraw from Supplementary Table 1. impairment arising from adolescent inhalant the study at any time. At the time of data abuse are compelling targets for further collection the average age of participants was Parameters of growth research, given that their incidence may 20.6±0.4 years. Height (cm) and weight (kg) were measured contribute to psychological harm and by the research team at the time of the chronic health effects. This study aimed to Drug use survey interview and body mass index (BMI) characterise patterns of other drug use and Classification into petrol sniffing or control calculated (weight (kg)/height squared (m2)). effects on growth in a cohort of adolescent groups was based upon a consensual Weight was measured on an electronic scale inhalant abusers from a population methodology, requiring corroboration from and height was measured using a tape. of Indigenous males. It also aimed to the individual, community health workers, These measures were undertaken by a single understand whether growth parameters research assistants and local health records. trained individual, to prevent inter-rater could represent a diagnostic tool to identify Each participant’s self-reported history was variability. Growth percentiles were calculated chronic inhalant abuse in adolescents, compared with local community clinic health using US Centre for Disease Control (CDC) including whether the effects were severe records, an assessment by a community 2000 growth charts (males aged 0-20 years19) enough to meet clinical ‘failure to thrive’ (FTT) health worker, and an assessment of petrol and the CDC National Health and Nutrition criteria. sniffing behaviour by research assistants Examination Survey (males > 20 years20). The recruited from the same language group. If criteria for meeting a FTT diagnosis were rd Method data was consistent across all these sources, weight below the 3 percentile or 20% below it was deemed to be an accurate reflection ideal weight for height, when compared to This study retrospectively analysed an of petrol sniffing history and the participant standard growth charts,21 only one criterion historical dataset, collected with the aim of was included in the study12 and furthermore, had to be met. investigating the social and neurological consistency between the consensual outcomes of adolescent inhalant abuse in assessment and self-assessment was high.18 Pooled groups Indigenous populations. Data collection All participants in the petrol sniffing group In initial studies using these participants; methods have been previously described met the criteria for inhalant abuse from the individuals with a history of petrol sniffing 9,12,17 in detail, but are summarised here and Diagnostic and Statistical Manual of Mental were separated based on the presence or discussed in the following paragraphs. Data Disorders IV. Drug use data were collected absence of lead encephalopathy. For this was collected from two remote Indigenous through an interview survey administered study, growth metrics were not different communities, where petrol sniffing had by the same researcher in conjunction with between these individuals, nor individuals been prevalent for many years. All subjects an Indigenous Research Assistant recruited between the two communities, and therefore were interviewed about petrol sniffing and from the same language group. This included individuals were not maintained in separate other drug use behaviour using a consensus age of onset/age petrol sniffing commenced groups for this study. To investigate the need methodology with survey instruments (years), duration (years), and severity (scale to control for the effects of other drug use developed specifically for this project, and of 1-5: 1/2=light/intermittent, 3=moderate, on growth, groups were stratified into the underwent a health check that included 4/5=heavy). The severity score was self-rated following; controls who use no other drugs neurological and cognitive examinations, based upon petrol sniffing behaviour in (n=15), controls who use one or two other collection of a blood sample, and height and the previous 12 months and was validated drug type use (n=13), controls who use three weight measurements. through a significant positive correlation or more other drugs (n=4) petrol sniffing with with the volume of petrol that participants no other drug use (n=9), petrol sniffing who 18 Participants reported having sniffed each week. use one or two other drugs (n=66), petrol Ethical approval for data collection was Participants were also surveyed regarding sniffing who use three or more other drugs obtained from the institutional ethics other drug use including alcohol, cigarette (n=11). These stratifications did not result committee, an independent Indigenous smoking, kava and marijuana. in any significant intra-group differences in ethics committee, and town councils from the height or weight (p>0.05). Thus individuals communities in the study. Ethical approval Haematological measures were pooled to form two groups; petrol for this analysis was granted by the Human Blood specimens were collected by sniffing and non-petrol sniffing control. Research Ethics Committee of Northern venepuncture. For those in the petrol sniffing Territory Department of Health and Menzies group, samples were collected within School of Health Research (HREC-2014-

Data analysis Table 1: Age of commencement and severity of any drug use. Statistical analysis was conducted using IBM Non-petrol sniffing controls Petrol sniffing group p value SPSS Statistics 22 and power calculations Petrol sniffing (n) N/A 86 - were conducted using G*Power 3.1.9.2. Data Age commenced (years) N/A 12.9 ± 0.4 - for each variable was assessed for normality, Severity score (0-5) N/A 4.1 ± 0.4 - and normalised as necessary using square Cigarette smoking (n) 9 63 root transformation. Group differences were Presence of smoking (%) 28.1 73.3 0.000** investigated using independent sample Age commenced (years) 18.3 ± 1.5 15.3 ± 0.4 0.033* t-tests. The linear relationship between Severity score (0-5) 1.9 ± 0.3 2.4 ± 0.1 0.106 continuous variables was investigated Alcohol drinking (n) 15 58 using linear regression and 2-tailed Pearson Presence of alcohol drinking (%) 46.9 67.4 0.053^ Correlation coefficients. Chi-square tests were Age commenced (years) 19.3 ± 1.3 16.8 ± 0.4 0.055^ used to investigate the relationship between Severity score (0-5) 2.8 ± 0.4 2.9 ± 0.2 0.788 categorical variables. Significance was set at Kava (n) 1 21 p≤0.05; trends reported at p>0.05 but ≤0.10. Presence of kava use (%) 3.1 24.4 0.007** Results are reported as mean±SEM. Age commenced (years) 28.0 17.0 ± 0.9 0.012* Severity score (0-5) 1.0 2.5 ± 0.3 0.315 Marijuana (n) 7 41 Results Presence of marijuana use (%) 21.9 47.7 0.011* Petrol sniffing Age commenced (years) 22.0 ± 3.0 17.3 ± 0.7 0.084^ Severity score (0-5) 1.7 ± 0.4 2.3 ± 0.2 0.283 The age of commencement for petrol sniffing ^p>0.5 but ≤0.1; *p≤0.05; **p≤0.01. was 12.9±0.4 years. The average duration Groups were non-petrol sniffing controls (n=32) and petrol sniffing individuals (n=86). Group differences between the presence/absence of drug use (%) were was 8.6±0.8 years and severity score 4.1±0.4. assessed using chi-square tests. Group differences between age of commencement and severity score for each drug use were assessed using independent There were significant linear correlations sample t-tests (mean±SEM). There is no SEM for kava use in the non-petrol sniffing control group, as n=1 respondent. Petrol sniffing commenced earlier than other drug use and was associated with increased and earlier use of other drugs, but not with increased severity of other drug use. between all three metrics of petrol sniffing behaviour (Supplementary Table 2), with the Haematological data percentile (p=0.001) between petrol sniffing severity of petrol sniffing being predicted by (19th±2nd percentile) and non-petrol sniffing both duration and age of commencement Results for haematological variables are groups (39th±5th percentile) and in weight-for- (p=0.012, linear regression). There was also a shown in Supplementary Table 1. There was a age percentile (p=0.009, Figure 2). Increased significant correlation between the severity significant difference in serum toluene levels duration of petrol sniffing was linearly score and blood lead levels (p=0.014), (p<0.000) between petrol sniffing (0.06±0.02 associated with reduced height percentile confirming the validity of the severity score. µg/mL) and non-petrol sniffing groups (0.00±0.00 µg/mL). However, 73.5% of petrol (p=0.024, linear regression, Figure 3), but not weight percentile. Age of commencement Other drug use sniffing individuals had 0.00 µg/mL serum toluene. and severity of petrol sniffing were not Petrol sniffing was predominantly the correlated with either height or weight There was a significant difference in creatine first drug used, with 48.1% of all petrol percentiles. When stratified into sub-groups kinase levels (p=0.002) between petrol sniffing individuals reporting that as their by other drug use, there were no significant sniffing (420.2±39.3 U/L) and non-petrol first substance; cigarette smoking was the intra-group differences in height or weight. sniffing groups (212.7±22.4 U/L). There was a next most common first substance (6.5%). Furthermore, there was no significant significant group difference in exceeding the In comparison, cigarette smoking was difference between groups for BMI (Figure creatine kinase reference range (p=0.035, chi the most common first substance in the 1C) and BMI percentile. control group (17.6%); alcohol consumption square). Neither serum toluene nor creatine kinase levels were correlated with any of the the next most common (11.8%). Petrol Failure to thrive sniffing was associated with an increased three metrics of petrol sniffing behaviour Of the petrol sniffing group, 49% were less proportion of all other drug use (trend for (p>0.05). than 3rd percentile for weight and thus met alcohol, p=0.053) and with earlier other drug FTT criteria compared to 25% in the non- use, which was significant for the average Growth data sniffing group (p=0.022, chi square). Of the 48 age of commencement of cigarette and There was a significant difference in height individuals who met FTT criteria, 83% were kava use, with trends for earlier alcohol (Figure 1A) and weight (Figure 1B) between petrol sniffing individuals. (p=0.055) and marijuana (p=0.084) use petrol sniffing and non-petrol sniffing groups, with petrol sniffing individuals on average (Table 1). Furthermore, the age that sniffing Diagnosing petrol sniffing commenced was positively linearly correlated 5 cm shorter (p<0.000, with power of 0.95) with age of other drug use commencement and 7 kg lighter (p=0.001, with power of The existing diagnostic model of serum for cigarette smoking (p=0.022), kava 0.86). The regression co-efficient for weight toluene had a sensitivity of 27% and a (p=0.048) and marijuana (p=0.040), but not as predicted by height was 0.70 (p<0.000) in specificity of 100% with an overall accuracy of alcohol. The severity of other drug use was the petrol sniffing group and 0.84 (p=0.003) 47%, reflecting the high rate of false negatives not significantly different between petrol in the non-petrol sniffing group. There was with no false positives. Creatine kinase had sniffing and non-sniffing groups (Table 1). also a significant difference in height-for-age a sensitivity of 35% and a specificity of 84%

Figure 1: Petrol sniffing is associated with decreased (A) height (cm) and (B) weight (kg), but there was no change in (C) body mass index (BMI) (kg/m2).

Data are presented as minimum/maximum boxplots. n=32 for non-petrol sniffing controls and n=82 for petrol sniffing individuals, as not all participants had measurements for these variables. **p≤0.01. with overall accuracy of 49%. Using FTT as toluene and creatine kinase as diagnostic is associated with neurocognitive pathology a diagnostic model resulted in a sensitivity models of chronic inhalant abuse, and and impairment5 and is a risk factor for of 49% and a specificity of 75% with overall FTT had the highest sensitivity and overall ongoing psychological harm.23 A history accuracy of 56%, reflecting a lower rate of accuracy for chronic petrol sniffing, which of drug use also increases the risk of false negatives. represents a novel and practical diagnostic psychotic outcomes including suicide.24 tool for this behaviour. Thus adolescent inhalant abuse may have a Discussion In our cohort, petrol sniffing commenced in direct contribution to lifelong psychological early adolescence and was predominantly impacts, and also contribute indirectly by We explored the effects of chronic petrol the first substance misused; a consistent increasing subsequent drug use. These sniffing during adolescence on subsequent finding with other demographic studies findings highlight the need for early diagnosis drug use and growth in a cohort of in Australia and overseas.2,22 Additionally, and intervention of inhalant abuse, ideally Indigenous males. We expanded initial commencement age was positively before individuals’ progress to other drugs 18 observations from this dataset to show that correlated with severity and duration of of abuse. Petrol sniffing during adolescence petrol sniffing was associated with increased petrol sniffing. This relationship is important significantly impaired weight. In humans, and earlier onset of subsequent drug use, and as duration of petrol sniffing significantly adolescent inhalant abuse is associated 15 the commencement age of petrol sniffing increases the magnitude of neurological and with weight loss and emaciation. These also correlated positively with that of other cognitive deficits.12 The findings of increased findings are recapitulated in rodent studies drug use. Petrol sniffing also impaired both and earlier drug use after petrol sniffing are and associated with reduced food intake in height and weight, and weight impairments consistent with other studies suggesting the absence of anhedonia, decreased weight 14 were so severe that half the petrol sniffing that inhalant use appears to establish a gain and reduced fat deposition. In humans individuals met FTT diagnosis. In this cohort, trajectory towards other drug use (see for there is evidence that incidental inhalation growth impairment out-performed serum review Dinwiddie 1994).11 This is of significant of petrol fumes may have an acute anorectic 25 concern as increased drug use in adolescence effect, suggesting that weight impairment Figure 2: Weight percentile difference between may be driven by decreased food intake, groups, presented as minimum/maximum boxplots while others suggest that toluene exposure Figure 3: Increased duration of petrol sniffing 26 with line on y-axis at 3rd percentile, denoting results in a higher basal metabolic rate or behaviour (years) is correlated with decreased 14 meeting failure to thrive criteria. alters appetite regulating hormones. Body height percentile (p=0.024), n=38 due to not all weight is driven by a complex relationship respondents providing data on duration of petrol between food intake, metabolic function, sniffing behaviour. body composition and energy expenditure;27 our dataset does not allow mechanistic interpretation. Nevertheless, the observed weight impairments were so severe that nearly 50% of petrol sniffing individuals met FTT criteria. FTT is not a disease, but rather a symptomatic description of poor growth attributable to many causes, including malnutrition or psychosocial factors including poverty.28 Whilst the baseline prevalence Petrol sniffing is associated with decreased weight percentile (p=0.009). A of FTT is higher in Indigenous populations break in the y axis below the 10th percentile is utilised to clearly display 29 left-skewed data. n=32 for non-petrol sniffing controls and n=82 for Data are presented as a scatter plot with a calculated regression line compared to non-Indigenous populations, petrol sniffing individuals. **p≤0.01. showing the relationship between the variables. we identified differences in the FTT rates

within these two communities, based upon facility, given their requirement for blood historic cohort of Indigenous males from a history of petrol sniffing in adolescence. collection. Since inhalant abuse is associated two remote communities. However, data Importantly, childhood FTT is associated with low socioeconomic status and inhalant from human inhalant abuse is difficult with an increased risk of subsequent abusers are disproportionately represented to obtain and, as drug use, growth and disease development, including obesity and in the mental health and protective services FTT are common measures of health and cardiovascular disease.30Adolescent petrol systems, these populations have less development, this study provides relevant sniffing also significantly impaired height, a regular contact with the medical system.7 In observations. One limitation of this study novel finding. Height was also a significant addition, petrol sniffing among Indigenous is that birth weight data was not available; predictor of weight in both petrol sniffing and people occurs predominantly in remote in order to assess growth trajectories prior non-petrol sniffing individuals, suggesting communities, where access to health services to petrol sniffing. However, the use of a that the weight impairments may be partially is limited. Therefore we investigated whether control group from the same communities attributable to impaired height. Adolescence the effects of inhalants on growth could be in this study provides some mitigation encompasses an intensive period of growth incorporated into a novel diagnostic tool for for this potential confound. Furthermore, (adolescent growth spurt) where over half of identifying chronic inhalant abuse. In our while low socioeconomic status and the peak bone mass is attained.31 For males, study, height percentile was significantly disadvantage are risk factors for FTT, and the growth spurt initiates around 12 years impaired by petrol sniffing and importantly, are associated with inhalant use, the control of age and ends at about 20 years of age,32 was also negatively correlated with increased group also provides mitigation for the risk which coincides with the primary time of duration of petrol sniffing behaviour, of these factors confounding our findings. petrol sniffing behaviour in this study. As this suggesting this metric may be sensitive to In addition, our ability to extrapolate these growth spurt can be disrupted by factors such identifying the nuances of chronic abuse. findings to females or to urban populations as malnutrition33 it is likely that inhalant abuse Furthermore, individuals with a history is limited. There is also a potential confound could disrupt this intensive growth period, of petrol sniffing met FTT criteria more of lead in this cohort, as individuals sniffed including skeletal growth. Indeed chronic frequently than non-sniffers. Thus growth a combination of leaded and unleaded toluene exposure impairs both skeletal metrics may be a clinically relevant tool for petrol, which has also been associated with growth and bone mineral density in rats16,34 identifying chronic adolescent inhalant abuse impaired weight and height.41 However, and bone mineral density in adolescent males having higher sensitivity and overall accuracy we suggest that these findings are relevant with a history of chronic glue sniffing.35 Any than any of the existing diagnostic tools. In to other inhalants, as toluene is a common disruption to growth during adolescence addition, growth metrics were almost twice solvent in all inhaled products and animal has the potential to cause long-term effects as sensitive as serum toluene in identifying studies using toluene have shown similar on stature.36 As both height and weight chronic inhalant abuse. For a screening test, growth impairments as in this study.14,16 were reduced, there was no change in BMI sensitivity is more valuable than specificity, as While further work will be needed to identify between petrol sniffing and non-petrol the focus for inhalant abuse needs to be on the mechanisms underlying these growth sniffing groups. This highlights the risk of improving the identification of a serious but impairments, particularly the effect of using BMI as a primary metric of growth. treatable condition. inhalant-induced appetite suppression, the Current diagnostic tools for inhalant abuse The use of growth metrics as a diagnostic tool persistence of these effects into abstinence, are limited, and restricted to presentation to is feasible, given that body growth is routinely and the health consequences of growth a medical facility under acute intoxication, measured by an individual or a parent/ impairment during adolescence, we suggest and the subsequent measurement of serum guardian, it does not change rapidly and that height impairment should be added to toluene or creatine kinase levels. In this measurement does not require contact with advisory notes as a warning sign of inhalant study, 73.5% of the petrol sniffing individuals the primary health care system or specialised abuse. had a zero reading of serum toluene within equipment. This may be especially valuable This is the first study that has established a 12 hours of petrol sniffing. Creatine kinase for disadvantaged and remote populations. It link between FTT and adolescent inhalant levels were elevated by petrol sniffing and is unknown how quickly growth is impaired abuse, potentially providing a novel may provide a more temporally stable by inhalant abuse; therefore, this approach diagnostic technique for chronic inhalant measure than serum toluene.9 However, the is more likely to be applicable to chronic abuse thus increasing the opportunity for relevance of creatine kinase to inhalants use. Growth impairments arising from intervention. Early intervention is effective other than petrol is unknown. Furthermore, adolescent inhalant abuse may also provide in relation to petrol sniffing as cognitive creatine kinase levels can be increased by the basis of a powerful health education impairments can resolve with sustained other environmental factors such as high message. Evidence suggests that some abstinence.17 Furthermore, early intervention ambient temperature37 or other drug use.38 inhalant users are motivated by a desire to can reduce subsequent drug use, adverse Neither serum toluene nor creatine kinase restrict their weight;39 however, it may be less psychological outcomes and other high-risk correlated with severity or duration of petrol desirable to also restrict height. This health behaviours.42 We suggest that inhalant abuse sniffing behaviour and were therefore not promotion message may also be relevant to should be added to the differential diagnosis helpful metrics for diagnosing chronic the study group we investigated, adolescent guidance for FTT in high risk populations and inhalant abuse. Indigenous males, where size and strength age groups. 40 Importantly, the measures indicated above are highly valued personal attributes. Our require acute presentation to a medical data were retrospectively analysed from a

6. Kurtzman TL, Otsuka KN, Wahl RA. Inhalant abuse by 27. Woods SC, Seeley RJ, Porte D, Schwartz MW. Signals that Conclusion adolescents. J Adolesc Health. 2001;28(3):170-80. regulate food intake and energy homeostasis. Science. 7. Kang M, Bernard D, Booth M, Quine S, Alperstein G, 1998;280(5368):1378-83. This study shows that petrol sniffing is Usherwood T, et al. Access to primary health care for 28. Shah MD. Failure to thrive in children. J Clin Gastroenterol. typically the first substance misused and is Australian young people: Service provider perspectives. 2002;35(5):371-4. associated with increased and earlier use of Br J Gen Pract. 2003;53(497):947-52. 29. Northern Territory Department of Health and Families. 8. Meulenbelt J, De Groot G, Savelkoul T. Two cases of acute Child Malnutrition / Failure to Thrive. Alice Springs other drugs. This misuse commences during toluene intoxication. Br J Ind Med. 1990;47(6):417-20. (AUST): Government of Northern Territory; 2008. a critical period for maturational processes. 9. Burns CB, Powers JR, Currie BJ. Elevated serum creatine 30. Stein AD, Thompson AM, Waters A. Childhood kinase (CK-MM) in petrol sniffers using leaded or growth and chronic disease: Evidence from countries Petrol sniffing significantly impaired weight unleaded fuel. Clin Toxicol. 1994;32(5):527-39. undergoing the nutrition transition. Matern Child Nutr. and height, independent of other drug use, 10. Lubman DI, Yücel M, Lawrence AJ. Inhalant abuse 2005;1(3):177-84. among adolescents: Neurobiological considerations. 31. Bonjour J-P, Theintz G, Buchs B, Slosman D, Rizzoli to the extent that many individuals met the Br J Pharmacol. 2008;154(2):316-26. R. Critical years and stages of puberty for spinal and criteria for a FTT diagnosis. In combination, 11. Dinwiddie SH. Abuse of inhalants: A review. Addiction. femoral bone mass accumulation during adolescence. these findings suggest that adolescent 1994;89(8):925-39. J Clin Endocrinol Metab. 1991;73(3):555-63. 12. MaruffP, Burns C, Tyler P, Currie B, Currie J. Neurological 32. Tanner JM. Growth and maturation during adolescence. inhalant abuse can initiate a trajectory of and cognitive abnormalities associated with chronic Nutr Rev. 1981;39(2):43-55. substance misuse and growth impairment, petrol sniffing. Brain. 1998;121(10):1903-17. 33. Dreizen S, Spirakis CN, Stone RE. A comparison of 13. Yip PS, Liu K-Y, Lam T, Stewart SM, Chen E, Fan S. skeletal growth and maturation in undernourished and thus highlights the need for early Suicidality among high school students in Hong Kong, and well-nourished girls before and after menarche. J diagnosis. Using growth parameters as SAR. Suicide Life Threat Behav. 2004;34(3):284-97. Pediatr. 1967;70(2):256-63. 14. Dick A, Simpson A, Qama A, Andrews Z, Lawrence 34. Atay AA, Kismet E, Turkbay T, Kocaoglu M, Demirkaya a diagnostic tool for inhalant abuse may A, Duncan J. Chronic intermittent toluene inhalation E, Sarici SU, et al. Bone mass toxicity associated with be clinically valuable as they are simple in adolescent rats results in metabolic dysfunction inhalation exposure to toluene. Biol Trace Elem Res. to implement and can be translated into with altered glucose homeostasis. Br J Pharmacol. 2005;105(1-3):197-203. 2015;172(21):5174-87. 35. Dündaröz MR, Sarici S, Turkbay T. Evaluation of bone populations where inhalant abuse is more 15. Glaser HH, Massengale ON. Glue-sniffing in children: mineral density in chronic glue sniffers. Turk J Pediatr. prevalent. Given the current difficulties Deliberate inhalation of vaporized plastic cements. 2002;44:326-9. JAMA. 1962;181(4):300-3. 36. Martorell R, Khan LK, Schroeder DG. Reversibility of in identifying inhalant misuse, our study 16. Pryor GT. A toluene-induced motor syndrome in rats stunting: Epidemiological findings in children from provides a novel opportunity for improving resembling that seen in some human solvent abusers. developing countries. Eur J Clin Nutr. 1994;48:S45-57. Neurotoxicol Teratol. 1991;13(4):387-400. 37. Knochel JP, Dotin LN, Hamburger RJ. Heat stress, diagnostic capabilities thus leading to 17. Cairney S, Maruff P, Burns CB, Currie J, exercise, and muscle injury: Effects on urate metabolism earlier intervention, in order to alter the Currie BJ. Neurological and cognitive recovery and renal function. Ann Intern Med. 1974;81(3):321-8. trajectory of ongoing drug use and prevent following abstinence from petrol sniffing. 38. Swartz CM, Breen KJ. Elevated serum CK in long Neuropsychopharmacology. 2005;30(5):1019-27. abstinent cocaine abusers. Am J Drug Alcohol Abuse. chronic adverse psychological and health 18. Burns C, d’Abbs P, Currie B. Patterns of petrol sniffing 1993;19(3):327-35. consequences. and other drug use in young men from an Australian 39. Pisetsky EM, May Chao Y, Dierker LC, May AM, Aboriginal community in Arnhem Land, Northern Striegel‐Moore RH. Disordered eating and substance Territory. Drug Alcohol Rev. 1995;14(2):159-69. use in high‐school students: Results from the Youth Acknowledgements and funding 19. Centre for Disease Control. Growth Charts, United States, Risk Behavior Surveillance System. Int J Eat Disord. 2000. Atlanta (GA): National Center for Health Statistics 2008;41(5):464-70. The authors thank Dr Chris Burns and National Center for Chronic Disease Prevention and 40. Mellor D, McCabe M, Ricciardelli L, Ball K. Body Professor Paul Maruff for data collection. Health Promotion; 2000. image importance and body dissatisfaction among 20. Fryar CD, Gu Q, Ogden CL. Anthropometric reference Indigenous Australian adolescents. Body Image. The research was supported by NHMRC data for children and adults: United States, 2007-2010. 2004;1(3):289-97. (940835), of which AJL is a Principal Research Vital Health Stat 11. 2012;(252):1-48. 41. Schwartz J, Angle C, Pitcher H. Relationship between 21. Perrin EC, Cole C, Frank D, Glicken SR, Guerina N, Petit childhood blood lead levels and stature. Pediatrics. Fellow (1020737), the Australian Research K, et al. Criteria for determining disability in infants and 1986;77(3):281-8. Council (DP 110100379) of which JRD was a children. Evid Rep Technol Assess (Summ). 2003;(72):1-5. 42. Toumbourou JW, Stockwell T, Neighbors C, Marlatt Future Fellow during the time of the study 22. Dell CA, Gust SW, MacLean S. Global issues in volatile G, Sturge J, Rehm J. Interventions to reduce harm substance misuse. Subst Use Misuse. 2011;46 Suppl associated with adolescent substance use. Lancet. (100100235) and the Victorian Government’s 1:1-7. 2007;369(9570):1391-401. Operational Infrastructure Support Scheme. 23. Armstrong TD, Costello EJ. Community studies on adolescent substance use, abuse, or dependence and psychiatric comorbidity. J Consult Clin Psychol. 2002;70(6):1224. Supporting Information References. 24. Borges G, Walters EE, Kessler RC. Associations Additional supporting information may be 1. Alcohol and other Drugs Council of Australia. Policy of substance use, abuse, and dependence with Position - Inhalants. Canberra (AUST): ADCA; 2010. subsequent suicidal behavior. Am J Epidemiol. found in the online version of this article: 2. Australian Institute of Health and Welfare. 2010 National 2000;151(8):781-9. Drug Strategy Household Survey Report. Canberra 25. Jackson CE, Currie BJ, Cairney S, Maruff PT, Snyder Supplementary Table 1: Methods and (AUST): AIHW; 2011. PJ. Hunger and the perception of the scent of petrol: results of haematological variables. 3. Cairney S, Maruff P, Burns C, Currie B. The A potential neurobiological basis for increased neurobehavioural consequences of petrol (gasoline) risk of petrol inhalation abuse. Addict Res Theory. Supplementary Table 2: Relationship sniffing.Neurosci Biobehav Rev. 2002;26(1):81-9. 2009;17(5):518-24. between metrics of petrol sniffing behaviour. 4. Cancer Council of Victoria. Australian Secondary School 26. Wang D-H, Ishii K, Seno E, Yane S, Horike T, Yamamoto Students’ Use of Tobacco, Alcohol, and Over-the-counter H, et al. Reduced serum levels of ALT and GGT and and Illicit Substances in 2011. Melbourne (AUST): high carbohydrate intake among workers exposed to Australian Department for Health and Ageing; 2012. toluene below the threshold limit values. Ind Health. 5. Lubman DI, Yücel M, Hall WD. Substance use 1998;36(1):14-9. and the adolescent brain: A toxic combination? J Psychopharmacol. 2007;21(8):792-4.

Supplementary Table 1 – Methods and results of haematological variables

Result (mean±SEM) Variable Unit A measure of Method of analysis Non-sniffing Petrol sniffing p value controls individuals Hemoglobin g/L Complete blood count Coulter S+4 counter 146.8±2.0 141.4±2.0 0.114 Red cell count 106/µL Complete blood count Coulter S+4 counter 5.2±0.1 5.1±0.1 0.299 Hematocrit / packed L/L Complete blood count Coulter S+4 counter 0.422±0.005 0.414±0.005 0.386 cell volume Mean cell volume fL Complete blood count Coulter S+4 counter 80.8±0.7 80.8±0.7 0.920 Red cell distribution Percentage Complete blood count Coulter S+4 counter 13.8±0.3 14.1±0.2 0.433 width Platelets 109/L Complete blood count Coulter S+4 counter 274.3±14.1 255.0±8.6 0.239 White cell count 109/L Complete blood count Coulter S+4 counter 9.3±0.4 9.2±0.3 0.778 Kodak “dry slide” Creatine kinase U/L Muscle inflammation 212.7±22.4 420.2±39.3 0.002 ** technique Smooth muscle Kodak “dry slide” Creatine kinase B U/L 1.8±0.4 3.0±0.3 0.075^ inflammation technique Iron µmol/L Iron panel Kodak E 700 14.5±0.8 16.2±0.8 0.202 Transferrin g/L Iron panel Beckman array 3.3±0.1 3.5±0.1 0.094^ Transferrin saturation % Calculation (% transferrin Iron panel 0.22±0.02 0.22±0.01 0.970 rate Saturation saturated with iron) Ferritin µg/L Iron panel Baxter Stratus 89.4±10.4 70.1±6.7 0.126 Blood lead µM/L Petrol exposure Standard additions method 0.27±0.03 1.52±0.10 0.000 ** Gas chromatograph mass Blood toluene µg/mL Petrol exposure 0.00±0.00 0.06±0.02 0.012 * spectrometry Gas chromatograph mass Blood benzene µg/mL Petrol exposure 0.000±0.000 0.011±0.003 0.057^ spectrometry

Consequences of adolescent inhalant abuse

The group differences between non-petrol sniffing controls and petrol sniffing individuals were measured using independent sample t-tests. For all haematological variables n=115 (32 controls and 83 petrol sniffing individuals). Results are presented as mean±SEM for each group. ^ p>0.5 but ≤0.1; * p≤0.05; ** p≤0.01.

Consequences of adolescent inhalant abuse

Supplementary Table 2 – relationship between metrics of petrol sniffing behaviour

Metric Age sniffing Duration of Severity of commenced sniffing sniffing Age sniffing commenced (years) 1 -0.426 -0.254 (0.008)** (0.040)* Duration of sniffing (years) -0.426 1 0.486 (0.008)** (0.003)** Severity of sniffing (scale 1-5) -0.254 0.486 1 (0.040)* (0.003)**

Only individuals in the petrol sniffing group (n=86) were asked this question.

Participation in all questions were optional and for this question n=72 for age sniffing commenced, n=66 for severity, and n=40 for duration. Results are reported as a two-tailed

Pearson’s correlation coefficient with associated p value in brackets. There is a significant correlation between all three metrics of petrol sniffing behaviour. *p≤0.05; **p≤0.01.

ADDICTION RESEARCH & THEORY, 2017 https://doi.org/10.1080/16066359.2017.1339229

RESEARCH ARTICLE The persistence of growth impairments associated with adolescent inhalant abuse following sustained abstinence

Rose Crossina , Sheree Cairneyb,c,d, Andrew John Lawrencea,e and Jhodie Rubina Duncana,f aAddiction Neuroscience, Florey Institute of Neuroscience and Mental Health, Parkville, VIC, Australia; bCentre for Remote Health, Flinders University, Alice Springs, NT, Australia; cCooperative Research Centre for Remote Economic Participation (CRC-REP), Ninti One Limited: Information-Innovation-Ideas for Remote Australia, Alice Springs, NT, Australia; dMenzies School of Health Research, Casuarina, NT, Australia; eFlorey Department of Neuroscience and Mental Health, University of Melbourne, Parkville, VIC, Australia; fSchool of Medicine, University of Adelaide, Adelaide, SA, Australia

ABSTRACT ARTICLE HISTORY Background: Abuse of inhalants containing the volatile solvent toluene is a significant public health Received 8 January 2017 issue, especially for adolescent and Indigenous communities. We previously demonstrated that inhalant Revised 1 June 2017 abuse (petrol sniffing) during adolescence results in impairments to height and weight. The aim of this Accepted 3 June 2017 study was to understand whether these impairments resolve or persist into early adulthood, following sustained abstinence. KEYWORDS Methods: Baseline data were collected from 118 Indigenous males; 86 chronically sniffed petrol during Petrol sniffing; height; adolescence. Following 2 years sustained abstinence, data were again collected from a subset (n ¼ 40) weight; Indigenous health; of this population; 30 sniffed petrol during adolescence. This study is a retrospective analysis of data volatile solvent abuse; body collected after 2 years sustained abstinence. mass index Results: After 2 years abstinence, inhalant-induced impairments to height persisted (p ¼ 0.023) whereas weight impairments resolved (p ¼ 0.796). Conclusions: Adolescent inhalant abuse alters growth trajectories, even after 2 years of sustained abstinence. Despite the fact that individuals continue to get taller, there is no catch-up growth in those who abused inhalants. The persistence of height impairments demonstrates that adolescent inhalant abuse can impact individuals into adulthood, despite sustained abstinence. In contrast, weight impairments associated with inhalant abuse resolved in abstinence, however, it is unknown if this represents a normalisation of weight or a rapid and unhealthy gain in weight. Further research is required to determine the health impacts of the observed weight changes.

Abbreviations: BMI: Body mass index

Introduction (Bonjour et al. 1991). Therefore, an impact on height has the potential to affect growth trajectories into adulthood (Reed Inhalant abuse is the deliberate inhalation of solvents (e.g. & Stuart 1959). Additionally, dramatic weight fluctuations petrol or glue), to achieve an altered mental state. can adversely impact health (Norgan 2000; Dulloo et al. Adolescents comprise the predominant population of inhal- 2006). Therefore, we investigated whether our observed ant abusers (AIHW 2011). In human and/or animal studies, weight and height changes arising from adolescent inhalant appetite suppression, reduced weight gain (Glaser & abuse (Crossin et al. 2016) persist following abstinence; cur- Massengale 1962; Dick et al. 2015) and depressed skeletal rently an open question. growth (Pryor 1991) are reported consequences of adolescent inhalant abuse. We previously showed that petrol sniffing during adolescence impairs height and weight, with petrol Methods sniffing individuals on average 5 cm shorter and 7 kg lighter than non-petrol sniffing controls (Crossin et al. 2016). This We undertook retrospective analysis of a historical dataset, study was the first to show height impairments arising from collected from a cohort of Indigenous males from a remote chronic inhalant abuse during adolescence (Crossin et al. northern Australian community. The current study utilises 2016). data collected as a follow-up after two years sustained abstin- Despite awareness of acute consequences, the long-term ence from chronic petrol sniffing (a combination of leaded effects of adolescent inhalant abuse on growth remain and unleaded petrol), achieved through a community inter- unknown but require consideration, as the peak time of vention programme. Data collection methods and validation inhalant abuse coincides with the adolescent growth spurt of the community-led intervention have been previously

CONTACT Rose Crossin Florey [email protected] Institute of Neuroscience and Mental Health University of Melbourne, Parkville, Victoria 3052, Australia ß 2017 Informa UK Limited, trading as Taylor & Francis Group. 2 R. CROSSIN ET AL. described (Burns et al. 1995a; Maruff et al. 1998), and are there were no volatile hydrocarbons detected in the blood of summarised below. any participants after two years abstinence and no differences were detected in creatine kinase levels between the Participants groups (Burns et al. 1995a). The use of cigarettes, alcohol, and cannabis remained consistent between the two time Ethical approval for this analysis was granted by the points, with a significant reduction in kava use after two Human Research Ethics Committee of Northern Territory years (Burns et al. 1995a). Department of Health and Menzies School of Health Research (HREC-2014-2274). Informed written consent was Growth measures given by all participants, who could withdraw from the study at any time. Height (cm) and weight (kg) were measured by the research We previously described the metrics of the historic petrol team during the interview and body mass index (BMI) cal- sniffing behaviour (Crossin et al. 2016). In summary, the age culated (weight (kg)/height squared (m2)). Weight was meas- of commencement for petrol sniffing was 12.9 ± 0.4 years. ured on an electronic scale and height was measured using a The average duration was 8.6 ± 0.8 (minimum 1, maximum tape. These measures were undertaken by a single trained 21) years and severity score 4.1 ± 0.4 (scale from 0-5, 5 indi- individual to prevent inter-rater variability, replicating previ- cating heaviest use) and the initial study was conducted at ous methodology (Crossin et al. 2016). an average age of 20.6 ± 0.4 years (Crossin et al. 2016). When data were originally collected, petrol sniffing was quantified as it related to a 375 mL soft drink can, as well as the sever- Data analysis ity score. The mean volume of petrol sniffed was 250 mL per Statistical analysis was conducted using IBM SPSS Statistics night and there was a significant correlation (p < 0.001) 22. Data for each variable was assessed for normality, and between the quantification estimate and the severity score normalised as necessary using square root transformation. (Burns et al. 1995b). Therefore the severity score was then Group differences were investigated using independent sam- used as the primary measure of consumption for the ple t-tests, which have high accuracy with small sample sizes. analysis. Data presented as mean ± SEM. Significance was set at After two years confirmed abstinence from petrol sniffing, p 0.05, trends reported at p > 0.05 0.1. 40 participants, a sub-set of the original study cohort, were re-interviewed by the original research team and an Indigenous Research Assistant recruited from the same lan- Results guage group. 30 participants had a history of petrol sniffing in adolescence, 10 non-petrol sniffing age-matched individu- Growth data is summarised in Table 1. After two years als acted as controls. Whilst the initial study was conducted abstinence, individuals with a history of petrol sniffing (age across two communities, only one community was revisited 23.0 ± 0.7) were on average 6 cm shorter than non-petrol for the follow-up study at two years post-abstinence. Within sniffing controls (age 22.3 ± 1.5). In contrast, there was no that community the retention rate in the study was 40 of the significant difference in body weight or BMI between the 58 initial participants, with the predominant reason for two groups, though there was a trend towards increased per- dropping out of the study being that individuals had left the centage weight gain in petrol sniffing individuals. community in the intervening two years (Burns et al. 1995a). Abstinence was confirmed by self-assessment combined Discussion with a consensual methodology, cross-checked against hos- pital admission records and confirmed by Aboriginal health We explored the persistence of growth impairments arising workers (Burns et al. 1995a; Cairney et al. 2005). Abstinence from inhalant abuse in a cohort of Indigenous males who was further biochemically confirmed by the finding that chronically sniffed petrol during adolescence, followed by

Table 1. Growth data after two years abstinence. After two years abstinence Non-petrol sniffing controls Petrol sniffing individuals p value Age (years) 22.3 ± 1.5 23.0 ± 0.7 0.693 Height (cm) 177.0 ± 3.1 171.3 ± 1.0 0.023 Change in height (cm) 3.1 ± 0.8 3.8 ± 0.7 0.590 Change in height (%) 1.8 ± 0.5 2.3 ± 0.5 0.514 Weight (kg) 66.0 ± 6.1 64.6 ± 2.4 0.796 Change in weight (kg) 4.8 ± 3.3 9.2 ± 1.6 0.191 Change in weight (%) 7.2 ± 4.6 18.0 ± 3.4 0.100# BMI (kg/m2) 20.9 ± 1.6 22.0 ± 0.7 0.502 Change in BMI (kg/m2) 0.7 ± 1.0 2.3 ± 0.6 0.172 Change in BMI (%) 3.4 ± 4.0 12.9 ± 3.4 0.139 Height was reduced by an average of 6 cm in individuals who sniffed petrol during adolescence, whereas weight was not significantly different. n ¼ 10 for non-petrol sniffing controls and n ¼ 30 for petrol sniffing individuals. Data presented as mean ± SEM. #p > 0.05 0.1. p < 0.05. ADDICTION RESEARCH & THEORY 3 two years sustained abstinence. In this cohort we previously et al. 2013), opioids (Thornhill et al. 1976), methampheta- showed that petrol sniffing during adolescence was associ- mine (Williams et al. 2004) and inhalants (Crossin et al. ated with height impairment (Crossin et al. 2016). While 2016). This weight loss is often attributed to poor nutrition acknowledging the low sample size of the participant subset and appetite suppression (Islam et al. 2002; Petersson et al. in our follow up study as a potential limitation, our data 2004). However, many studies, in both animals and humans, show that the effects of inhalant abuse on height persist into have observed that the degree of weight loss is not directly early adulthood, despite sustained abstinence. Despite indi- comparable to measured changes in food intake, which sug- viduals in both the control and petrol sniffing groups con- gests that nutrition is not the sole mediator of drug-induced tinuing to get taller, there is no catch-up growth in those weight loss (Perkins 1992; Cooper & Van der Hoek 1993; who abused inhalants, suggesting that adolescent petrol sniff- Ersche et al. 2013). ing may alter height trajectory. The mechanism for inhalant- In contrast to active substance abuse, recovery/abstinence induced growth impairment is unknown, though one have a strong association with weight gain (Jackson & Grilo possible explanation is that growth impairments are a result 2002; Hodgkins et al. 2004), and our results are consistent of malnutrition, as inhalant abuse is associated with altered with this. Various hypotheses exist to explain this abstinence metabolism (Dick et al. 2015) and reduced food intake effect include; a homeostatic rebound effect to normalise (Moron et al. 2004). Alternatively, inhalant abuse throughout body weight (Nolan 2013), addiction switching from a drug the adolescent growth spurt may directly affect bone growth to palatable food (Brunault et al. 2015), or common reward- (Martorell et al. 1994). Regardless of the mechanism, persist- seeking phenotype between drugs and palatable food (Nolan ent impairments to height may have detrimental consequen- & Stolze 2012). However, there are limits to all these hypothe- ces including adverse psychological outcomes e.g. social ses and all focus on food intake as the mediator of the growth anxiety and poor self-image (Molinari et al. 2002), as well as changes. Further research is needed into the cause of weight adverse developmental and educational outcomes (Siegel regain after substance abuse before effective interventions can et al. 1990; Norgan 2000). be developed, to assist in the prevention of health issues such In contrast to height, there was no long term difference as insulin resistance (Dulloo et al. 2006) in abstinence. in body weight; however the percent weight change in the In conclusion, we show that petrol sniffing associated petrol sniffing group was more than double that of the non- height impairments persist into adulthood despite sustained petrol sniffing controls. It is unlikely that the rapid weight abstinence. This highlights the need for prevention and early gain during abstinence is due to withdrawal symptoms from intervention for adolescent inhalant abusers, as in contrast petrol sniffing, as these are characterised by anhedonia to cognitive and neurological consequences (Cairney et al. rather than increased food intake, and are transient (Shah 2005), abstinence alone cannot ameliorate the adverse et al. 1999), though we acknowledge that no data on with- growth consequences. Indeed, this height restricting effect drawal symptoms was collected as part of this study. Weight may present the opportunity for a prevention message about regain after weight loss is a common physiological response inhalant abuse that will be relevant to adolescents, for whom to re-establish homeostasis (Sumithran & Proietto 2013). size and strength are highly valued personal attributes However, rapid weight regain is associated with adverse con- (Mellor et al. 2004). We also show that abstinence from pet- sequences including increased fat deposition and the devel- rol sniffing results in a rapid weight gain, the cause of which opment of insulin resistance (Dulloo et al. 2006). is currently unknown, but has the potential to lead to further Furthermore, whilst the average BMI of study participants adverse health outcomes even though inhalant abuse behav- ‘ ’ was in the normal range after two years abstinence; it can- iours have ceased. not be assumed that a normal BMI indicates metabolic health. Individuals with a history of petrol sniffing may present as metabolically-obese normal-weight, a harmful pheno- Acknowledgements type associated with increased cardiovascular and diabetes The authors thank Dr Chris Burns and Professor Paul Maruff for data disease mortality (Carnethon et al. 2012) though this collection. The research was supported by NHMRC (940835), of which requires confirmation. A further potential confound is that AJL is a Principal Research Fellow (1020737), the Australian Research whilst individuals were in abstinence from petrol sniffing, Council (DP 110100379) of which JRD was a Future Fellow during the their use of other drugs such as cigarettes, alcohol, and mari- time of the study (100100235) and the Victorian Government’s juana continued, though these drugs were not previously Operational Infrastructure Support Scheme. Funding bodies had no involvement in the design, analysis and decision to publish. There are found to be predictors of growth outcomes in this cohort no conflicts of interest or financial disclosures in this work. (Crossin et al. 2016). Our results highlight the need for further research into the effects of inhalant abuse on body weight, particularly in abstinence, as body weight can signifi- Disclosure statement cantly influence long-term health outcomes (Dulloo et al. No potential conflict of interest was reported by the authors. 2006). Additional human cohorts with larger sample sizes would be valuable in this regard. The effect of petrol sniffing, and indeed, substance abuse Funding in general on body weight is poorly understood. Active sub- The research was supported by NHMRC (940835), of which AJL is a stance abuse is associated with impaired body weight for Principal Research Fellow (1020737), the Australian Research Council many different drug types, including cigarettes (Grebenstein (DP 110100379) of which JRD was a Future Fellow during the time of 4 R. CROSSIN ET AL. the study (100100235) and the Victorian Government’s Operational Hodgkins CC, Cahill KS, Seraphine AE, Frostpineda K, Gold MS. 2004. Infrastructure Support Scheme. Adolescent drug addiction treatment and weight gain. J Addict Dis. 23:55–65. Islam SN, Hossain KJ, Ahmed A, Ahsan M. 2002. Nutritional status ORCID of drug addicts undergoing detoxification: prevalence of malnutrition and influence of illicit drugs and lifestyle. BJN Nutr. Rose Crossin http://orcid.org/0000-0003-1814-1330 88:507–513. Jackson T, Grilo C. 2002. Weight and eating concerns in outpatient men and women being treated for substance abuse. Eat Weight Disord Obesity. 7:276–283. References Martorell R, Khan LK, Schroeder DG. 1994. Reversibility of stunting: epidemiological findings in children from developing countries. Eur AIHW. 2011. 2010 National Drug Strategy Household Survey Report. J Clin Nutr. 48:S45–S57. Canberra. Maruff P, Burns C, Tyler P, Currie B, Currie J. 1998. Neurological and Bonjour J, Theintz G, Buchs B, Slosman D, Rizzoli R. 1991. Critical cognitive abnormalities associated with chronic petrol sniffing. Brain. years and stages of puberty for spinal and femoral bone mass accu- 121:1903–1917. mulation during adolescence. J Clin Endocrinol Metabol. Mellor D, McCabe M, Ricciardelli L, Ball K. 2004. Body image import- 73:555–563. ance and body dissatisfaction among Indigenous Australian adoles- Brunault P, Salame E, Jaafari N, Courtois R, Reveillere C, Silvain C, cents. Body Image. 1:289–297. Benyamina A, Blecha L, Belin D, Ballon N. 2015. Why do liver Molinari E, Sartorio A, Ceccarelli A, Marchi S. 2002. Psychological and transplant patients so often become obese? The addiction transfer emotional development, intellectual capabilities, and body image in hypothesis. Med Hypotheses. 85:68–75. short normal children. J Endocrinol Invest. 25:321–328. Burns C, d'Abbs P, Currie B. 1995a. Patterns of petrol sniffing and Moron L, Pascual J, Portillo P, Casis L, Macarulla MT, Abecia LC, other drug use in young men from an Australian Aboriginal com- Echevarrı E. 2004. Toluene alters appetite, NPY, and galanin immu- munity in Arnhem Land, Northern Territory. Drug Alcohol nostaining in the rat hypothalamus. Neurotoxicol Teratol. Rev.14:159–169. 26:195–200. Burns CB, Currie BJ, Clough AB, Wuridjal R. 1995b. Evaluation of Nolan LJ. 2013. Shared urges? The links between drugs of abuse, eating, strategies used by a remote aboriginal community to eliminate petrol – sniffing. Med J Australia. 163:82–86. and body weight. Curr Obes Rep. 2:150 156. Nolan LJ, Stolze MR. 2012. Drug use is associated with elevated food Cairney S, Maruff P, Burns CB, Currie J, Currie BJ. 2005. Neurological – and cognitive recovery following abstinence from petrol sniffing. consumption in college students. Appetite. 58:898 906. Neuropsychopharmacology. 30:1019–1027. Norgan N. 2000. Long-term physiological and economic consequences Carnethon MR, De Chavez PJD, Biggs ML, Lewis CE, Pankow JS, of growth retardation in children and adolescents. Proc Nutr Soc. – Bertoni AG, Golden SH, Liu K, Mukamal KJ, Campbell-Jenkins B. 59:245 256. 2012. Association of weight status with mortality in adults with inci- Perkins KA. 1992. Effects of tobacco smoking on caloric intake. Br J – dent diabetes. JAMA. 308:581–590. Addict. 87:193 205. Cooper SJ, Van der Hoek GA. 1993. Cocaine: a microstructural analysis Petersson A, Olsson-Mortlock C, Thiblin I, Rajs J, Eksborg S. 2004. of its effects on feeding and associated behaviour in the rat. Brain Nutritional status of deceased illicit drug addicts in Stockholm, – Res. 608:45–51. Sweden: a longitudinal medicolegal study. J Forensic Sci. 49:1 10. Crossin R, Cairney S, Lawrence A, Duncan J. 2016. Adolescent inhalant Pryor GT. 1991. A toluene-induced motor syndrome in rats resembling abuse leads to other drug use and impaired growth; implications for that seen in some human solvent abusers. Neurotoxicol Teratol. – diagnosis. Aust NZ J Public Health. 41:99–104. 13:387 400. Dick A, Simpson A, Qama A, Andrews Z, Lawrence A, Duncan J. 2015. Reed RB, Stuart HC. 1959. Patterns of growth in height and weight – Chronic intermittent toluene inhalation in adolescent rats results in from birth to eighteen years of age. Pediatrics. 24:904 921. metabolic dysfunction with altered glucose homeostasis. Br J Shah R, Vankar G, Upadhyaya HP. 1999. Phenomenology of gasoline Pharmacol. 172:5174–5187. intoxication and withdrawal symptoms among adolescent in India: a Dulloo AG, Jacquet J, Seydoux J, Montani J-P. 2006. The thrifty ‘catch- case series. Am J Addict. 8:254–257. up fat’phenotype: its impact on insulin sensitivity during growth tra- Siegel P, Clopper R, Stabler B. 1990. Psychological impact of signifi- jectories to obesity and metabolic syndrome. Int J Obes Relat Metab cantly short stature. Acta Paediatr Scand Suppl. 377:14–18. Disord. 30:S23–S35. Sumithran P, Proietto J. 2013. The defence of body weight: a physio- Ersche KD, Stochl J, Woodward JM, Fletcher PC. 2013. The skinny on logical basis for weight regain after weight loss. Clin Sci. cocaine: insights into eating behavior and body weight in cocaine- 124:231–241. dependent men. Appetite. 71:75–80. Thornhill J, Hirst M, Gowdey C. 1976. Disruption of diurnal Glaser HH, Massengale ON. 1962. Glue-sniffing in children. Deliberate feeding patterns of rats by heroin. Pharmacol Biochem Behav. inhalation of vaporized plastic cements. JAMA. 181:300–303. 4:129–135. Grebenstein PE, Thompson IE, Rowland NE. 2013. The effects of Williams MT, Moran MS, Vorhees CV. 2004. Behavioral and extended intravenous nicotine administration on body weight and growth effects induced by low dose methamphetamine administra- meal patterns in male Sprague-Dawley rats. Psychopharmacology. tion during the neonatal period in rats. Int J Develop Neurosci. 228:359–366. 22:273–283. Rose Crossin (737900)

Chapter 3 – A systematic review and meta-analysis on the growth impacts of inhalant abuse The aim of this study was quantify the magnitude of the effect on inhalant abuse (in clinical studies) and toluene exposure (in pre-clinical models) on body weight and height using previously published literature. This took into account study characteristics, to identify potential moderators to these effects on body weight and height. The rationale for this aim was firstly to build upon Chapter 2, and to determine whether the findings from the human cohort were consistent with the body of literature. Secondly, to determine the magnitude of these effects on body weight and height, including whether there was a dose-response relationship evident, and third, to determine if effects varied by study characteristics such as sex, species, and age at first exposure. This study would provide beneficial information for the inhalant abuse field more broadly, and also enable future experiments in an animal model to be guided through determination of the factors that would need to be controlled for. This was described in the paper: Growth changes after inhalant abuse and toluene exposure: a systematic review and meta-analysis of human and animal studies.

Paper citation:

Crossin, R., Lawrence, A. J., Andrews, Z. B., Churilov, L., & Duncan, J. R. (2018). Growth changes after inhalant abuse and toluene exposure: A systematic review and meta-analysis of human and animal studies. Human & Experimental Toxicology, 0960327118792064.

39 Article

Human and Experimental Toxicology 1–16 ª The Author(s) 2018 Growth changes after inhalant abuse Article reuse guidelines: sagepub.com/journals-permissions and toluene exposure: A systematic DOI: 10.1177/0960327118792064 review and meta-analysis of human journals.sagepub.com/home/het and animal studies

R Crossin1,2 , AJ Lawrence1,3, ZB Andrews4 , L Churilov3 and JR Duncan1,5

Abstract Inhalant abuse is a significant public health issue, particularly for adolescents, the predominant group of inhalant users. Adolescence is a critical growth period, and inhalant abuse has been associated with growth impairments, including reduced body weight and height. However, the extent to which inhalant abuse affects growth remains unquantified, and potential moderators remain unknown. To address this knowledge gap, a systematic review and meta-analysis of clinical human and preclinical animal studies utilizing toluene exposure (the primary solvent in abused products) was conducted. Five-hundred and sixty-nine studies were screened; 31 met inclusion criteria, yielding 64 toluene-control comparisons for body weight and 6 comparisons for height. Toluene exposure was negatively associated with body weight (d ¼0.73) and height (d ¼0.69). Concen- tration of inhaled toluene, but not duration, moderated the effect of toluene exposure on body weight, with more severe impairments at higher concentrations. Differences in effect size for body weight were observed for study characteristic subgroups including sex, age at first exposure, administration route and species. However, these findings should be interpreted cautiously due to low study numbers. Growth impairments, particularly during adolescence, can cause long-term health consequences. These effects on growth are therefore an important clinical outcome for individuals with a history of inhalant abuse.

Keywords Weight, height, volatile solvent abuse, adolescence

Introduction 1 Inhalant abuse is the deliberate inhalation of products Addiction Neuroscience, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, Melbourne, containing volatile solvents, in order to achieve an 1 Victoria, Australia altered mental state. Demographic studies into inha- 2Eastern Health Clinical School, Monash University, Clayton, Vic- lant abuse show that usage patterns are similar across toria, Australia sexes2 and overwhelmingly indicate that abuse is 3Florey Institute of Neuroscience and Mental Health, University most commonly associated with adolescence (partic- of Melbourne, Parkville, Melbourne, Victoria, Australia 4Department of Physiology, Monash University, Clayton, Victoria, ularly young adolescence) patterns that are consistent 3–5 Australia worldwide. For example, population-based surveys 5School of Medicine, University of Adelaide, Adelaide, South Aus- indicate that 19.1% of Australian 12- to 13-year olds tralia, Australia self-report experimenting with inhalants.3 While large population-based surveys do not collect data from Corresponding author: R Crossin, Addiction Neuroscience, Florey Institute of Neu- children under the age of 12, and smaller cohort roscience and Mental Health, University of Melbourne, Parkville, studies have identified that children under the age of Melbourne, Victoria 3010, Australia. 6 12 may also misuse inhalants. Email: [email protected] 2 Human and Experimental Toxicology XX(X)

Preadolescence and adolescence is a critical period characteristics. Firstly, the range of products inhaled of growth, with the peak growth spurt for females by inhalant users is varied and opportunistic and occurring between 10 years and 14 years of age, and includes common household items such as glue, thin- a slightly later spurt for males occurring at 12–15 ners and aerosols. However, a common solvent to all years of age.7 Studies indicate that for many inhalant these products is toluene, which is a colourless, water- users, the duration of abuse occurs over periods insoluble, aromatic hydrocarbon. Inhalant users will greater than 10 years, with users reporting abuse pat- preferably inhale pure toluene if it is available,20 and terns equivalent to intermittent high usage,8 thus over- thus toluene exposure is utilized as an exemplar for lapping with the period when rapid growth is inhalant abuse in animal models in the majority of occurring in many young inhalant users.3 It is, there- studies.21 Study design must consider toluene dosage fore, unsurprising that inhalant abuse has an effect on which when the administration route is inhalation will growth, which has been observed in both males6,9 and be reflected as a concentration in parts per million females10; however, this effect is poorly understood. (ppm). Human inhalant users will often inhale via a Case studies from the 1960s report that inhalant solvent-soaked cloth or directly from a bag, and the users present as emaciated,9 and a reduction in weight concentration of inhaled toluene from this method is is well-accepted as a consequence of inhalant abuse, estimated to range from 3000 ppm to 15,000 ppm.21 It to the extent that it is often listed as a warning sign of is important to note that exposure to toluene may also this type of drug abuse.1 Indeed weight impairments occur in the industrial setting; however, this occurs in can be so severe in humans that inhalant abuse is a different pattern to abuse with exposure at lower associated with failure to thrive (FTT).6 Weight concentrations for extended durations. Secondly, impairments (either a loss of weight or reduction in many humans abuse inhalants in a chronic yet inter- weight gain) is also a common outcome in animal mittent nature.8 Thus, a study design must include a experiments that utilize toluene exposure as a model pattern of exposure in minutes/day and days/week for of inhalant abuse, occurring immediately after the a specified period of time. Despite the diverse range exposure period,11–13 and body weight is often of exposure patterns used experimentally, this can be included as a descriptive statistic, even if it was not quantified as a variable called ‘exposure hours’, the key outcome under consideration. However, the which can then be multiplied by the dosage to give effects of toluene exposure on body weight are not ‘total toluene exposure’. The third issue is that of consistent in effect size, suggesting that unknown administration route. Inhalation is the most reflective moderators may play a role in the effect of toluene of human usage but requires specialized equipment exposure on weight. and thus some models utilize intraperitoneal (IP) Given that the adolescent growth spurt is reflected injections or gavage to administer toluene. How the in a change to individuals’ weight and height, it is amount of toluene administered is reported will surprising that the effects of inhalant abuse on height depend on the administration route, with inhalation have not been studied to the same degree as weight. reported as a concentration in ppm, IP injections as Height impairments arising from inhalant abuse have milligrams/kilogram (mg/kg) and gavage as millili- been reported in both humans6,14 and animal models tres/kilogram (mL/kg). The final issue is that of utilizing toluene exposure,15 and in humans this effect experimental animal. Rats and mice are the predomi- has been shown to persist into sustained abstinence.16 nant species used in inhalant abuse models, with rats Again the size of this effect has never been quantified, having equivalent sensitivity with humans to and the potential moderators remain unknown. toluene,22 but choices must still be made about sex Despite our lack of understanding of the factors mod- and the age at which exposure will begin, bearing in erating inhalant-induced changes to growth para- mind the association between inhalant abuse and meters, the effects of inhalant abuse on weight and young adolescence. height are important outcomes. This is especially due To date, the extent to which inhalant abuse affects to the potential for long-term health impacts arising growth patterns have not been quantified, and the from alterations to growth,17–19 even if an individual effects of potential moderators are not well- with a history of abusing inhalants remains abstinent. understood. To address this knowledge gap, we The majority of the research into the physiological conducted a systematic review of human and animal consequences of inhalant abuse is derived from ani- studies and applied meta-analytic techniques to cal- mal models, which incorporate five main study culate an effect size d. The primary aim of the study Crossin et al. 3

Table 1. Key words (in greyed boxes) and synonyms (below) used as alternative terms for the search strategy. Inhalant abuse Adolescence Weight Height Body mass index Volatile solvent abuse Teenage Body weight Length BMI VSA Youth Emaciation Spine length Adiposity Toluene Young adult Malnutrition Stunting Waist circumference Petrol sniffing Overweight Linear growth Z-score Glue sniffing Obesity Weight-to-height ratio Huffing Underweight Cachexia Wasting Failure to thrive was to examine the effects of inhalant abuse (in studies, only those that utilized toluene as the inha- humans) and toluene exposure (in animals) on weight lant were included in this meta-analysis (excluding and height, hypothesizing that ‘as amount of toluene animal models of inhaled benzene or xylene), as this administered (in ppm, mg/kg or mL/kg) increases, directly links the experimental model to the products height and body weight will decrease’. A secondary typically abused by humans (e.g. glue, paint and sol- aim was to explore the effect on height and body vents). Given that growth data are often reported as a weight of a range of study characteristics that are com- descriptive statistic in studies, without being a pri- monly encountered in inhalant abuse studies, pattern of mary outcome, it was not possible to limit the search exposure (including exposure duration and total to keywords such as inhalant þ weight. Therefore, a toluene exposure), administration route and model range of strategies were utilized including keyword (including species, sex and age at first exposure). searches, searches using synonyms for describing growth patterns (Table 1) and keyword searches of Methods physiological outcomes relative to toluene exposure where it might be plausible that body weight would This study was registered with the open access data- be reported (e.g. reproduction, motor syndrome and base PROSPERO (international prospective register food intake). of systematic reviews) at inception (registration num- Each combination of keyword and synonym was ber CRD42017069002; https://www.crd.york.ac.uk/ used as search strings; this strategy meant that 84 prospero/display_record.php?RecordID¼69002) and 23 unique searches (provided as search strings in Online conducted in accordance with PRISMA guidelines. Supplemental Table 1) were performed in PubMed, The completed PRISMA flow chart is provided as which yielded a total of 622 potential studies; 53 of Online Supplemental material. those were duplicates, leaving 569 studies that were checked to determine whether they included a Systematic literature review description of height or weight. Of the 569 studies, Peer-reviewed studies of inhalant abuse and toluene only 52 reported on at least one of these outcome exposure were collated via extensive literature measures. Studies were then excluded if they did not searches, searched for in PubMed. To be included, provide the mean of height or weight along with a studies must have provided a measure of height or measure of variance, for both the treatment and the weight that was taken immediately or shortly after control groups. This criterion reduced the number of exposure, sample sizes in each group along with a eligible studies from 52 to 31, with 21 studies measure of variance such as standard deviation (SD) excluded. It was deemed that this number of studies or standard error of the mean (SEM), in order for an was sufficient to proceed from systematic review to effect size to be calculated. Studies that focused on meta-analysis, and these 31 studies were then used for height or weight effects occurring in sustained absti- the meta-analysis. Figure 1 shows the PRISMA flow nence (i.e. after exposure to toluene had ceased) chart; the 21 excluded studies with reasons box refer were not the focus of this study and were not to the 21 studies discussed above. Risk of bias within included. Only published, peer-reviewed studies the included studies was assessed using the SYRCLE were considered for inclusion. For the animal risk-of-bias tool for animal studies,24 acknowledging 4 Human and Experimental Toxicology XX(X)

Figure 1. /prismastatement/flowdiagram (PRISMA) flow chart showing study numbers. Source: Moher et al.23 For more information, see www.prisma-statement.org. that four of the included studies were clinical obser- statistics computed was Cohen’s d, by comparing the vational studies, however, it was not considered mean body weight and/or height from the toluene group appropriate to utilize multiple risk of bias tools. to the mean body weight and/or height from the control Results of the assessment are provided as Online Sup- group. Then, 95% confidence intervals were computed plemental Table 2. for each d. Heterogeneity is reported using the I2 statistic, which describes the percentage of variation across studies that is due to heterogeneity rather than chance. Meta-analysis techniques Following Higgins et al.,50 we consider I2 values of The software program Review Manager (RevMan) ver- 25%,50% and 75% as low, moderate and high, respec- sion 5.3, Copenhagen: The Nordic Cochrane Centre, tively. Note that these categories do not refer to the The Cochrane Collaboration, 2014, was used to analyse absolute amount of observed heterogeneity, but rather the data and conduct the meta-analysis. An effect size to the proportion of the observed effect variance that statistic was computed for each toluene-to-control com- would remain if the sampling error were to be elimi- parison; the effect size was estimated by pooling indi- nated, that is, if we were to be able to observe the true vidual effects using a random-effects model. The effect size for all studies in the analysis. Crossin et al. 5

When entering the data into RevMan, the reportedpffiffiffiffi postnatal day and/or body weight, from which the SEM values were converted to SD (SD ¼ SEM N). category was calculated.53 Pearson’s correlation was As many of the studies utilized multiple treatment used to explore the relationship between concentra- groups, often due to testing more than one concentration and exposure for the identified experimental tion of toluene, a multiple-comparisons correction models. There are limited published data available for was required.51 The approach taken was to compute toluene exposure, and owing to incomplete data in all an ‘effective N’ for the control group that was being categories, a meta-regression could not be underta- utilized for more than one treatment comparison by ken, which would allow the relationship between dividing N by the number of total comparisons. Thus, study characteristics to be explored more fully. The each toluene-to-control comparison was entered indi- small number of studies also did not provide sufficient vidually into RevMan, with 64 unique comparisons power for the inclusion of funnel plots to assess pub- for body weight and 6 unique comparisons for height lication bias. The approach taken was therefore arising from the 31 studies. exploratory in nature, estimating the effects of the The studies identified in the systematic review uti- various study characteristics. All p values for individ- lized a wide range of concentrations of toluene (min- ual comparisons (seven in total) were therefore imum 80 ppm, maximum 10,000 ppm). To analyse reported, without multiplicity correction, and all concentration effectively, these were collapsed into Cohen’s d results are reported with the 95% confi- the following groups: 0–500 ppm, 501–2000 ppm, dence interval in brackets. 2001–5000 ppm and >5000 ppm. These groupings were chosen for consistency with previous meta- analysis on toluene exposure52 and provided an Results approximately even pool of comparisons across the Overall effect sizes and study characteristics subgroups. Additional analysis was performed post A summary of the study characteristics and effect hoc comparing occupational exposure concentrations sizes is provided in Table 2. The overall effect size ( 1000 ppm) with inhalant abuse concentrations on body weight was 0.73 [95% CI 0.99, 0.47], (>1000 ppm). Studies using a non-inhalation route with I2 of 76% categorized as high heterogeneity such as IP injection or gavage were not assigned a (Figure 2), and the overall effect size on height was concentration. Exposure hours reflected the total 0.69 [95% CI 1.16, 0.21], with I2 of 65% cate- duration of exposure to toluene and were calculated gorized as moderate–high heterogeneity (Figure 3), by multiplying the exposure duration in hours and the indicating that exposure to toluene resulted in signif- total number of exposures. For example, 1 h/day and 3 icantly decreased body weight and height, though the days/week for 4 weeks would be reported as 12 expo- relatively high heterogeneity suggests that subgroups sure hours. Total toluene exposure (in ppm h) was may exist that affect the outcome.50 calculated by multiplying exposure hours and the concentration in ppm, and then collapsed into subgroups of low (0–50,000 ppm h), medium (50,001–200,000 The effect of toluene concentration on body ppm h), high (>200,000 ppm h), which were chosen to weight and height provide approximately even distribution of studies As hypothesized, toluene concentration was a clini- across the subgroups. Fourteen comparisons could not cally meaningful moderator of body weight ( 2 ¼ be allocated a concentration (due to either utilizing a 25.50, df ¼ 3, p < 0.00001), as shown in Figure 4, non-inhalation route or because it was a human study with a full forest plot in Online Supplemental Figure of inhalation and concentration was not measured) 1. Clear separation between the 95% confidence inter- and did not have exposure hours and, therefore, could vals was observed between the upper and lower con- not have total toluene exposure computed. centration subgroups, suggestive of a concentration– Each study characteristic was assigned subgroups response relationship, though there was overlap (e.g. sex: male and female), and 2 tests were used to between the confidence intervals of the two middle quantify the differences between subgroups for each subgroups. The lowest concentration subgroup (0– study characteristic. Characterization of age at first 500 ppm) had no observable effect on body weight, exposure was into either adolescent or adult and was with decreases in body weight observed in the three based on data extracted from studies; some studies upper concentration subgroups. Furthermore, when directly stated adult/adolescent, whereas others stated concentration was differentiated post hoc by 6 Table 2. Summary of study characteristics and effect sizes used in the meta-analysis.a Effect size (d) for body Effect size Toluene amount Study ID weight with [95% CI] (d) for height (ppm unless stated) Exposure hours Total toluene exposure Admin. route Species Sex Age at first admin. Bushnell et al.25 1-1 0.22 [1.11, 1.55] 100 300 Low Inhalation Mouse M Adult 1-2 0.30 [1.04, 1.63] 1000 300 High Inhalation Mouse M Adult 1-3 0.27 [1.61, 1.06] 3000 300 High Inhalation Mouse M Adult Chan et al.26 2-1 0.06 [0.99, 1.10] 500 mg/kg IP injection Rat M Adolescent 2-2 0.07 [1.12, 0.98] 500 mg/kg IP injection Rat F Adolescent 2-3 0.51 [0.56, 1.59] 500 mg/kg IP injection Rat M Adolescent 2-4 0.31 [1.36, 0.75] 500 mg/kg IP injection Rat F Adolescent Chen et al.27 3-1 0.04 [0.73, 0.80] 500 mg/kg IP injection Rat M Adolescent 3-2 0.17 [0.59, 0.93] 500 mg/kg IP injection Rat M Adolescent Chien et al.28 4-1 0.03 [1.08, 1.02] 500 mg/kg IP injection Rat M Adolescent 4-2 0.47 [0.60, 1.54] 500 mg/kg IP injection Rat M Adolescent Crossin et al.6 5-1 0.68 [1.09, 0.26] 0.79 [1.21, 0.37] Not stated Inhalation Human M Adolescent Dick et al.29 6-1 1.34 [2.43, 0.24] 10,000 12 Medium Inhalation Rat M Adolescent 6-2 2.57 [3.38, 1.76] 10,000 12 Medium Inhalation Rat M Adolescent Dick et al.30 7-1 0.17 [0.97, 0.64] 3000 12 Low Inhalation Rat M Adolescent 7-2 0.20 [0.60, 1.00] 3000 24 Medium Inhalation Rat M Adolescent Dick et al.31 8-1 0.40 [1.35, 0.54] 3000 20 Medium Inhalation Rat M Adolescent 8-2 1.95 [2.86, 1.03] 10,000 12 Medium Inhalation Rat M Adolescent Dick et al.11 9-1 1.92 [2.68, 1.16] 10,000 12 Medium Inhalation Rat M Adolescent Duncan et al.12 10-1 0.18 [1.06, 0.70] 3000 12 Low Inhalation Rat M Adolescent 10-2 1.21 [2.18, 0.24] 3000 24 Medium Inhalation Rat M Adolescent Duncan et al.32 11-1 3.14 [4.80, 1.49] 10,000 9 Medium Inhalation Rat M Adolescent Dundaroz et al.33 12-1 0.15 [0.68, 0.38] 0.25 [0.78, 0.28] Not stated Inhalation Human M Adolescent Funada et al.34 13-1 1.33 [0.39, 3.05] 350 1.5 Low Inhalation Mouse M Adult 13-2 0.15 [1.70, 1.40] 700 1.5 Low Inhalation Mouse M Adult 13-3 0.25 [1.31, 1.80] 2500 1.5 Low Inhalation Mouse M Adult 13-4 0.27 [1.29, 1.82] 3200 1.5 Low Inhalation Mouse M Adult Furlong et al.35 14-1 1.84 [3.61, 0.08] 10,000 12 Medium Inhalation Rat M Adolescent Gospe et al.36 15-1 1.46 [2.42, 0.50] 1 mL/kg Gavage Rat F Adolescent Gotohda et al.37 16-1 1.11 [2.12, 0.10] 1500 28 Low Inhalation Rat M Adolescent Ishigami et al.38 17-1 4.58 [9.27, 0.12] 1500 80 Medium Inhalation Rat M Adolescent Jenkins et al.39 18-1 0.70 [0.23, 1.62] 1085 240 High Inhalation Rat M and F Adolescent 18-2 0.78 [0.18, 1.73] 107 2160 High Inhalation Rat M and F Adolescent 18-3 3.79 [5.27, 2.31] 1085 240 High Inhalation Guinea pig M and F Adolescent 18-4 0.60 [1.48, 0.28] 107 2160 High Inhalation Guinea pig M and F Adolescent 18-5 2.92 [5.76, 0.09] 1085 240 High Inhalation Dog M and F Adolescent 18-6 0.57 [2.27, 1.13] 107 2160 High Inhalation Dog M and F Adolescent 18-7 0.22 [1.61, 1.18] 1085 240 High Inhalation Monkey M and F Adolescent 18-8 1.19 [0.38, 2.76] 107 2160 High Inhalation Monkey M and F Adolescent Johnson40 19-1 0.76 [1.70, 0.18] 1000 160 Medium Inhalation Rat M Adolescent Martıńez-Alfaro et al.41 20-1 3.14 [5.29, 0.98] 3000 21 Medium Inhalation Rat M Adolescent Moron et al.13 21-1 1.59 [2.96, 0.22] 1.3 mL/kg IP injection Rat M Adult Olgar et al.14 22-1 1.70 [2.16, 1.23] 1.46 [1.91, 1.01] 7.63 + 5.69 tubes (50 mL/tube) Inhalation Human M Adolescent Ono et al.42 23-1 2.87 [5.31, 0.43] 4000 70 High Inhalation Rat M Adolescent 23-2 4.09 [7.25, 0.93] 6000 70 High Inhalation Rat M Adolescent Pryor15 24-1 2.70 [3.86, 1.55] 2000 616 High Inhalation Rat M Adolescent 24-2 4.91 [7.18, 2.65] 0.11 [1.24, 1.02] 2200 1288 High Inhalation Rat M Adolescent (continued) Table 2. (continued)

Effect size (d) for body Effect size Toluene amount Study ID weight with [95% CI] (d) for height (ppm unless stated) Exposure hours Total toluene exposure Admin. route Species Sex Age at first admin. 24-3 5.46 [7.91, 3.00] 0.31 [1.45, 0.83] 4400 644 High Inhalation Rat M Adolescent 24-4 6.00 [8.65, 3.35] 0.66 [1.82, 0.50] 6200 322 High Inhalation Rat M Adolescent Roberts et al.43 25-1 0.22 [1.13, 0.69] 100 570 Medium Inhalation Rat F Adolescent 25-2 0.43 [1.35, 0.50] 500 570 High Inhalation Rat F Adolescent 25-3 0.66 [1.63, 0.31] 2000 570 High Inhalation Rat F Adolescent 25-4 0.23 [1.14, 0.69] 2000 570 High Inhalation Rat F Adolescent 25-5 1.18 [2.19, 0.17] 2000 570 High Inhalation Rat F Adolescent Roberts et al.44 26-1 0.16 [0.75, 1.06] 250 54 Low Inhalation Rat F Adult 26-2 0.05 [0.87, 0.96] 750 54 Low Inhalation Rat F Adult 26-3 0.17 [0.75, 1.09] 1500 54 Medium Inhalation Rat F Adult 26-4 1.09 [2.05, 0.14] 3000 54 Medium Inhalation Rat F Adult Saillenfait et al.45 27-1 0.50 [0.37, 1.37] 500 84 Low Inhalation Rat F Adolescent 27-2 0.43 [1.29, 0.42] 1500 84 Medium Inhalation Rat F Adolescent Schiffer et al.46 28-1 1.88 [3.39, 0.37] 5000 3 Low Inhalation Rat M Adolescent Uchino et al.47 29-1 0.84 [1.32, 0.36] Not stated Inhalation Human F Adolescent von Euler et al.48 30-1 0.24 [0.75, 0.27] 80 120 Low Inhalation Rat M Adolescent Yoon et al.49 31-1 0.25 [0.38, 0.87] 7000 4 Low Inhalation Rat M Adolescent

IP: intraperitoneal. aThirty-one studies were included in the meta-analysis, yielding 64 unique toluene-to-control comparisons for body weight and 6 unique comparisons for height. ID is the study number followed by the comparison number. The characteristics of each toluene-to-control comparison are outlined in Table 2, along with the mean effect size and 95% confidence interval in brackets. Boxes are greyed out when there are no data available. Total toluene exposure was first calculated by multiplying concentration and exposure hours (ppm h). These were then collapsed into categories based on low (0–50,000 ppm h), medium (50,001–200,000 ppm h), high (>200,000 ppm h). 7 8 Human and Experimental Toxicology XX(X)

Figure 2. Exposure to toluene impaired body weight. Bars represent the effect size for each study with a 95% confidence interval. Multiple toluene-to-control comparisons (c) existed within a single study, and these are noted as c1, c2, and so on. occupational-level exposure (1000 ppm) or inhalant p < 0.00001). Only inhalant abuse exposure had an abuse exposure (>1000 ppm), this was also a signifi- effect on body weight (Online Supplemental Figure cant moderator of body weight ( 2 ¼ 25.81, df ¼ 1, 2), with no observable effect on body weight for Crossin et al. 9

Figure 3. Exposure to toluene impaired height. Bars represent the effect size for each study with a 95% confidence interval. Multiple toluene-to-control comparisons (c) existed within a single study, and these are noted as c1, c2, and so on.

intervals were observed. However, there was a significant negative relationship (Pearson’s correlation coefficient ¼0.366, p ¼ 0.009) between concentration and exposure hours in the experimental models within the studies, and the highest concentrations only occurred in the lowest exposure subgroup (Figure 6). Total toluene exposure was computed to account for this relationship, and increased total toluene exposure was associated with decreased body weight ( 2 ¼ 21.55, df ¼ 2, p < 0.00001, Table 3, Figure 5 and Online Supplemental Figure Figure 4. As toluene concentration increased, body 4), with the low exposure subgroup having no very weight was more impaired. The mean effect size for each small effect on body weight; however, there was subgroup within toluene concentration is shown with a 95% confidence interval. Only studies that cited a specific overlap between the confidence intervals of the concentration of toluene (in ppm) were included, which medium and high subgroups, both of which showed excluded 14 comparisons. Inhaled concentration was a large effect sizes. significant moderator of body weight, with the highest Administration route affected body weight ( 2 ¼ negative effect size observed at the highest concentration. 18.42, df ¼ 2, p ¼ 0.0001, Table 3, Figure 5 and Online Supplemental Figure 5), with toluene adminis- occupational-level exposure. The 14 comparisons that tered via IP injection having no effect on body weight, could not be ascribed a concentration in ppm were not but decreases being observed for both inhalation and able to be included in these subgroup analyses. In gavage, with large effect sizes. Differences in the addition, due to the unavailability of data in all cate- effect of toluene exposure on body weight were also gories for height, these analyses could not be com- observed for both species ( 2 ¼ 14.95, df ¼ 5, p ¼ pleted for height as an outcome. 0.01, Table 3, Figure 5 and Online Supplemental Fig- ure 6) and sex ( 2 ¼ 5.32, df ¼ 1, p ¼ 0.02, Table 3, Figure 5 and Online Supplemental Figure 7). Body Estimating the effect of study characteristics on weight decreases, with large effect sizes, were body weight observed for humans and rats following toluene expo- The effect of all study characteristics on body weight sure, but results were inconclusive for dogs and gui- was investigated, and the findings are summarized in nea pigs exacerbated by the low number of studies for Table 3 and Figure 5 (full forest plots in Online Sup- comparison, and results for monkeys and mice plemental Figures 3 to 8). trended towards an increased body weight following There was not a clear relationship between expo- toluene exposure, despite all these species undergoing sure hours and body weight ( 2 ¼ 0.65, df ¼ 3, p ¼ exposure via inhalation. While both males and 0.89, Table 3, Figure 5 and Online Supplemental females showed a decrease in body weight following Figure 3), and overlaps between the confidence toluene exposure, the effect size was greater for males 10 Human and Experimental Toxicology XX(X)

Table 3. The effects of toluene exposure on body weight, by subgroup. Moderator kI2 (%) Effect size (d) [95% confidence interval] p Value for subgroup differences Overall 64 76 0.73 [0.99, 0.47] N/A Inhaled concentration 0–500 ppm 11 21 0.05 [0.28, 0.38] <0.00001 501–2,000 ppm 16 72 0.78 [1.33, 0.23] 2001–5000 ppm 14 75 1.12 [1.79, 0.45] >5000 ppm 9 86 2.18 [3.18, 1.18] Exposure hours 0–20 15 80 0.91 [1.53, 0.29] 0.89 21–100 13 66 0.66 [1.21, 0.11] 101–500 10 82 0.94 [1.82, 0.06] >500 12 81 0.98 [1.74, 0.22] Total toluene exposure Low 14 19 0.06 [0.35, 0.23] <0.00001 Medium 16 74 1.22 [1.75, 0.70] High 20 83 1.36 [2.09, 0.64] Administration route Inhalation 54 78 0.85 [1.14, 0.55] 0.0001 IP injection 9 0 0.01 [0.32, 0.33] Gavage 1 N/A 1.46 [2.42, 0.50] Species Human 4 85 0.85 [1.45, 0.24] 0.01 Rat 47 78 0.83 [1.15, 0.51] Mouse 7 0 0.23 [0.33, 0.78] Guinea pig 2 92 2.14 [5.26, 0.99] Dog 2 49 1.46 [3.70, 0.78] Monkey 2 42 0.44 [0.94, 1.81] Sex Male 41 80 0.95 [1.30, 0.59] 0.02 Female 15 33 0.42 [0.70, 0.14] Age at first exposure Adolescent 52 79 0.88 [1.17, 0.59] 0.001 Adult 12 16 0.09 [0.48, 0.29]

IP: intraperitoneal; k ¼ the number of effect sizes (i.e. the number of toluene-to-control comparisons); I2 ¼ heterogeneity.

than females. Age at first exposure also had an effect (2001–5000 ppm and >5000 ppm) and exposure hours on body weight ( 2 ¼ 10.28, df ¼ 1, p ¼ 0.001, Table (101–500 h and >500 h). These results are not shown, 3, Figure 5 and Online Supplemental Figure 8), with due to the incompleteness of the data set. adolescents showing decreased body weight following toluene exposure, whereas there was no clear effect on body weight for adults, and the confidence Discussion interval crossed zero. We conducted a systematic review and meta- analysis of the inhalant abuse literature, incorporat- ing the use of toluene exposure as an experimental Estimating the effect of study characteristics on model of inhalant abuse and including both human height and animal studies. We demonstrate that inhalant Because of the small number of comparisons for abuse in humans and toluene exposure in animal height, full analysis of all study characteristics could models are associated with significant impairments not be undertaken. No characteristic had all subgroups to both body weight and height, with effect sizes that represented, and the only characteristics for which would be categorized as medium/large according to there was more than one subgroup represented for Cohen.54 Furthermore, the results indicate a concen- height were species (human and rat), concentration tration–response relationship between inhaled Crossin et al. 11

growth have the potential to affect an individual’s long-term health outcomes, through both physiological and psychological consequences.17–19 The finding that toluene concentration is a clinically meaningful moderator of the effect of toluene on body weight is juxtaposed by the non-significant effect of exposure hours, with similar medium–large effect sizes observed across the four subgroups. This may be due to the experimental design, where exposure to the highest concentrations have shorter exposure hours, thus this result may be confounded by the uneven distribution of concentrations within those experimental models. We attempted to over- come this issue by computing total toluene exposure, as a way of combining both toluene concentration and the duration of exposure. Total toluene exposure affected body weight more significantly than exposure hours (time) alone, a finding consistent with Bowen et al.55; however, there was no expected dis- tinction between the medium and the high exposure Figure 5. The effect size of toluene exposure on body weight differed by study characteristics, such as exposure subgroups. This suggests that while both toluene con- method, species and age at first exposure. The effect of centration and time in combination (i.e. total toluene each study characteristic on body weight was investigated exposure) affect body weight, the relative importance and the mean effect size for each subgroup within the study of these two factors requires further exploration. characteristic is shown with a 95% confidence interval. For The additional study characteristics of administra- exposure hours and total toluene exposure, only studies tion route, species, sex and age at first exposure meet that cited a specific concentration of toluene (in ppm) were the threshold for significance with p values less than included, which excluded 14 comparisons. 0.05; however, these results should be interpreted with caution. Some findings are consistent with the literature; for example, toluene exposure has been previously associated with weight gain in mice.52 It would also be expected that toluene would have a larger effect during adolescence, when growth is accelerating, as opposed to adulthood when growth (particularly height) is more stable. However, data are not consistently available across all subgroups. For example, in the systematic review, the studies with mice utilized only adult animals, whereas the guinea pig, dog and monkey data all came from adolescent animals. Therefore, we cannot identify whether it is Figure 6. There is an inverse relationship between species, age at first exposure or both that is affecting ¼ ¼ exposure hours and concentration (r 0.3657 p 0.009, the body weight. Furthermore, having identified a Pearson’s correlation test) within the experimental models concentration–response relationship, we found dif- used for the identified studies, demonstrating that the highest concentrations of toluene have the shortest ferences in the concentrations of toluene that differ- exposure hours, and vice versa. This analysis excluded the ent subgroups within each study characteristic were 14 comparisons that did not report a toluene concentra- exposed to. For example, in the highest inhaled con- tion (in ppm). centration subgroup of >5000 ppm, there are only males and no females, only adolescents and no toluene concentration and body weight impairment, adults and only rats and no other animal species. with low concentrations (<500 ppm) having no sig- Given that higher effect sizes were observed in nificant effect on body weight. These changes to males, adolescents and rats, it is possible that this 12 Human and Experimental Toxicology XX(X) is simply reflecting the response to the high concen- individual than to a developing foetus. In contrast to tration of toluene, rather than sex, age and species our findings, Callan et al. did not identify a linear moderating the relationship between toluene expo- concentration–response relationship and observed a sure and growth changes. reversal of the weight impairment trend at the highest It is important that valid animal models of toluene concentration of inhaled toluene but did find a linear exposure exist, which reflect human patterns of inha- relationship between total toluene exposure and lant abuse and which yield comparable outcomes. The weight impairments.52 This may reflect the fact that, differences between effect sizes for administration due to the distribution of the available data, the same route are of concern, particularly as toluene given via subgroups for total toluene exposure could not be IP injection resulted in no significant effect on body used between the two meta-analyses. It is also possi- weight. The data for gavage is difficult to interpret ble that for the exposed individual, toluene concen- due to the low number of studies, but exposure via tration is a more meaningful variable, whereas for a inhalation appears to yield results more consistent developing foetus, it is the total toluene exposure that with clinical outcomes. Rats are the most common causes the greatest effect on birth weight. Consistent animal model used for inhalant abuse, and it is posi- with our findings, Callan et al. found that toluene tive to note the large overlap in the confidence inter- exposure resulted in increased birth weight for mice, vals between rats and humans, suggesting that they compared to rabbits and rats, suggesting that mice are appear to be a valid experimental model. However, as not a good experimental model for exploring the noted, at the highest concentrations, only male rats are growth effects of inhalant abuse.52 represented. Given that inhalant abuse has similar To understand the relationship between inhalant prevalence across sexes,56 future animal experiments abuse and growth more fully, there are a range of need to consider whether the concentration–response experiments that would assist in filling the identified relationship between male and female rats is equiva- knowledge gaps, and while studies may measure lent. This is particularly relevant as only one human weight, weight data are not always included in the inhalant abuse study could be found that directly com- publication but is a valuable output to report on. pared the growth effects of inhalant abuse between Firstly, experiments should be conducted to under- males and females and showed that weight impair- stand the relative importance of concentration versus ments in females were more severe than in males.47 exposure time, within total toluene exposure. Sec- The lack of studies using females is particularly con- ondly, experiments should be done to understand cerning because a weight impairment and any associ- whether sex, age at first exposure and species do mod- ated loss of adiposity11 during adolescence can cause erate the effect of toluene exposure on body weight, disruptions to pubertal and metabolic processes due to by conducting experiments in which all subgroups the endocrinological activity of adipose tissue.57 have equivalent total toluene exposure. Thirdly, it Furthermore, as shown by Callan et al., exposure to would be beneficial for future experiments to also toluene during pregnancy can impair developmental consider height as a growth outcome, which can be body weight,52 and inhalant abuse is most prevalent in easily measured in experimental animals (crown-to- females during childbearing years.2,3 Thus, inhalant rump length) but is often missed as a descriptive out- abuse in females has the potential to affect not only come. This would allow better understanding of how the individual but also her offspring. toluene exposure affects height and what moderates The effects of inhalant abuse in females have been this effect. explored before in a meta-analysis that considered Weight impairment and emaciation can result in a whether toluene exposure during the prenatal period diverse range of harmful outcomes including affected birth weight of the offspring.52 This meta- impaired cognition, renal failure and osteoporosis.17 analysis only included animal studies, and the only Weight impairment may normalize to re-establish growth outcome was birth weight; however, many homeostasis if the factor suppressing weight is of the study characteristics explored were the same. removed.58 However, this process is associated with The average effect size for weight in the individual rapid weight regain (predominantly through increased from this meta-analysis was 0.73, compared to fat deposition), which then increases the risk of sub- 0.39 for the effect on birth weight following prena- sequent health issues such as visceral adiposity and tal exposure.52 This suggests that the growth impacts insulin resistance.18 Thus, weight impairments can of inhalant abuse may be more severe in the exposed cause long-term health impacts to individuals and are Crossin et al. 13 therefore an important clinical outcome for individu- resistance. This study does not provide mechanistic als with a history of inhalant abuse. Future studies explanation for the effect of toluene on body weight should focus on adolescence as the predominant time and height, which means that we are unable to com- for inhalant use,2,3 overlapping with a critical growth ment on whether the effect on height and weight are period, and seek to incorporate subsequent outcomes common correlated outcomes of the same mechanism such as adiposity and insulin resistance, as well as the or distinct outcomes arising from different mechan- persistence of weight and height effects in abstinence, isms. However, understanding the mechanisms to enable a better understanding of the long-term behind these effects on body weight and height, health consequences of inhalant abuse. including the impact of changes to food intake fol- It should be noted that one limitation of this lowing toluene exposure11,13 is necessary and will meta-analysis is the low number of comparisons, provide information that is relevant to the treatment particularly for the height outcome, and the uneven of those with a history of inhalant abuse. distribution of comparisons across the various sub- In conclusion, this systematic review and meta- groups, which makes interpreting the effects of study analysis forms a comprehensive summary of the characteristics difficult. As we were unaware at the human and animal research literature on the growth time of PROSPERO registration whether there would effects of inhalant abuse and toluene exposure (used be sufficient studies to conduct a meta-analysis and/or as a model of inhalant abuse). We have shown a meta-regression, the methodology was not prospec- medium–large negative effect of toluene exposure tively registered, which is a further limitation of this on body weight, which confirms weight impairment study. Overall, the risk of bias was difficult to assess, as a warning sign of inhalant abuse. Indeed the effect with unclear results for the majority of studies within on body weight is such that it has been previously the performance domain, predominantly attributable associated with FTT in humans, and furthermore, to lack of detailed methods on allocation procedures, identification of FTT can be a strong diagnostic mar- housing protocols and blinding. Due to lack of pre- ker for inhalant abuse in relevant clinical popula- registered protocols for any of the included studies, tions.6 We also show that toluene exposure has a there is a risk of bias due to selective outcome report- medium–large negative impact on height, and despite ing; however, for the majority of the studies, body limited studies, the effect size is similar to that of weight and height were not the primary outcomes of body weight. It is therefore recommended that height investigation; therefore, this may not be a major lim- impairment be added to the warning signs of inhalant itation for this study. We also acknowledge the diffi- abuse and as a potential deterrent to adolescents, for culties inherent in comparing animal studies to human whom height impairment compared to their peers may studies, particularly due to the range of substances be an undesirable outcome. We show that rats and (e.g. glue, paint, petrol, thinners, etc.) that humans exposure by inhalation show results consistent with may inhale. Although all these products contain human findings, but other species and administration toluene, there may be unknown substance-specific routes should be utilized with caution, until future effects. Defining the level of exposure is difficult, experimental research has tested whether these study given the multiple combinations between the amount characteristics do moderate the effect of toluene on of toluene exposure, the duration of that exposure and growth. The effect of sex is also under-studied, mak- how long it occurred for. Although we attempted to ing it difficult to extrapolate these findings to females address these issues through computation of exposure in the clinical setting. Additional outcomes that arise duration and total toluene exposure, this approach from growth changes, for example, insulin resistance does not take into account peak blood concentrations and adiposity should be included in future studies. of toluene (data that would be valuable but is not cited Thus, knowledge gaps need to be filled in order to in these studies), and thus a detailed analysis of the understand potential long-term health risks, stemming pharmacokinetics of toluene under the variety of from growth impairments, for individuals with a his- exposure models could not be undertaken. A further tory of inhalant abuse. limitation of this study is tlack of long-term health impact data arising from the changes to growth, which Author contributions means that while we can quantify the effect size on The study was conceived by RC, AJL, ZBA and JRD. The weight and height, we are unable to relate that to any systematic review was undertaken by RC. RC and LC long-term health outcome such as adiposity or insulin designed the meta-analysis, which was conducted by RC 14 Human and Experimental Toxicology XX(X) and checked by LC. All authors were involved in reviewing velocity, and stages of puberty. Arch Dis Child 1976; the manuscript drafts and approved the final version. 51: 170–179. 8. Lubman D, Yu¨cel M and Lawrence A. Inhalant abuse Declaration of Conflicting Interests among adolescents: neurobiological considerations. Br The authors declared no potential conflicts of interest with J Pharmacol 2008; 154: 316–26. respect to the research, authorship, and/or publication of 9. Glaser HH and Massengale ON. Glue-sniffing in chil- this article. dren: deliberate inhalation of vaporized plastic cements. JAMA 1962; 181: 300–303. Funding 10. Malm G and Lying-Tunell U. Cerebellar dysfunction The authors disclosed receipt of following financial support related to toluene sniffing. Acta Neurol Scand 1980; for the research, authorship and/or publication of this arti- 62: 188–190. cle: This research was supported by the Australian National 11. Dick A, Simpson A, Qama A, et al. Chronic intermit- Health and Medical Research Council (NHMRC) tent toluene inhalation in adolescent rats results in (940835), of which AJL is a Principal Research Fellow metabolic dysfunction with altered glucose homeosta- (1116930) and ZBA is a Career Development Fellow sis. Br J Pharmacol 2015; 172: 5174–5187. (1084344) and the Victorian Government’s Operational 12. Duncan JR, Dick ALW, Egan G, et al. Adolescent Infrastructure Support Scheme. RC is funded by a federal toluene inhalation in rats affects white matter matura- RTP scholarship. Funding bodies had no involvement in tion with the potential for recovery following absti- the design, analysis and decision to publish. nence. PloS One 2012; 7: e44790. 13. Moro´n L, Pascual J, Portillo P, et al. Toluene alters appe- ORCID iD tite, NPY, and galanin immunostaining in the rat R Crossin http://orcid.org/0000-0003-1814-1330 hypothalamus. Neurotoxicol Teratol 2004; 26: 195–200. ZB Andrews http://orcid.org/0000-0002-9097-7944 14. Olgar S, Oktem F, Dindar A, et al. Volatile solvent abuse caused glomerulopathy and tubulopathy in street Supplemental material children. Hum Exp Toxicol 2008; 27: 477–483. Supplemental material for this article is available online. 15. Pryor GT. A toluene-induced motor syndrome in rats resembling that seen in some human solvent abusers. References Neurotoxicol Teratol 1991; 13: 387–400. 1. Alcohol and Other Drugs Council of Australia. Policy 16. Crossin R, Cairney S, John Lawrence A, et al. The position – inhalants. Canberra: Alcohol and other persistence of growth impairments associated with Drugs Council of Australia, 2010. adolescent inhalant abuse following sustained absti- 2. Australian Institute of Health and Welfare. National nence. Addict Res Theory 2017; 26: 183–186. drug strategy household survey 2016. Canberra: 17. Mehler PS and Brown C. Anorexia nervosa–medical Australian Institute of Health and Welfare, 2016. complications. J Eat Disord 2015; 3: 11. 3. Cancer Council Victoria. Australian secondary school 18. Dulloo AG, Jacquet J, Seydoux J, et al. The thrifty students’ use of tobacco, alcohol, and over-the-counter ‘catch-up fat’ phenotype: its impact on insulin sensi- and illicit substances in 2011. Melbourne: Cancer tivity during growth trajectories to obesity and meta- Council of Victoria for Australian Government bolic syndrome. Int J Obes (Lond) 2006; 30: S23–S35. Department for Health and Ageing, 2012. 19. Molinari E, Sartorio A, Ceccarelli A, et al. Psycholo- 4. Nguyen J, O’Brien C and Schapp S. Adolescent inha- gical and emotional development, intellectual capabil- lant use prevention, assessment, and treatment: a liter- ities, and body image in short normal children. ature synthesis. Int J Drug Policy 2016; 31: 15–24. J Endocrinol Invest 2002; 25: 321–328. 5. Dell CA, Gust SW and MacLean S. Global issues in 20. Villatoro JA, Cruz SL, Ortiz A, et al. Volatile sub- volatile substance misuse. Subst Use Misuse 2011; 46: stance misuse in Mexico: correlates and trends. Subst 1–7. Use Misuse 2011; 46: 40–45. 6. Crossin R, Cairney S, Lawrence AJ, et al. Adolescent 21. Bowen SE, Batis JC, Paez-Martinez N, et al. The last inhalant abuse leads to other drug use and impaired decade of solvent research in animal models of abuse: growth; implications for diagnosis. Aust N Z J Public mechanistic and behavioral studies. Neurotoxicol Ter- Health. 2016; 41: 99–104. atol 2006; 28: 636–647. 7. Tanner J and Whitehouse R. Clinical longitudinal stan- 22. Benignus VA, Boyes WK, Kenyon EM, et al. Quanti- dards for height, weight, height velocity, weight tative comparisons of the acute neurotoxicity of Crossin et al. 15

toluene in rats and humans. Toxicol Sci 2007; 100: 35. Furlong TM, Duncan JR, Corbit LH, et al. Toluene 146–155. inhalation in adolescent rats reduces flexible behaviour 23. Moher D, Liberati A, Tetzlaff J, et al. Preferred report- in adulthood and alters glutamatergic and GABAergic ing items for systematic reviews and meta-analyses: signalling. J Neurochem 2016; 139: 806–822. the PRISMA statement. PLoS Med 2009; 6: e1000097. 36. Gospe SM, Saeed DB, Zhou SS, et al. The effects of 24. Hooijmans CR, Rovers MM, de Vries RB, et al. SYR- high-dose toluene on embryonic development in the CLE’s risk of bias tool for animal studies. BMC Med rat. Pediatr Res 1994; 36: 811–815. Res Methodol 2014; 14: 43. 37. Gotohda T, Tokunaga I and Kubo S. Toluene 25. Bushnell PJ, Evans HL and Palmes E. Effects of inhalation-induced adrenocortical hypertrophy and toluene inhalation on carbon dioxide production and endocrinological changes in rat. Life Sci 2005; 76: locomotor activity in mice. Fundam Appl Toxicol 1929–1937. 1985; 5: 971–977. 38. Ishigami A, Tokunaga I, Kubo S, et al. Immunohisto- 26. Chan MH, Tang YC, Chien TH, et al. Toluene expo- chemical study of rat spermatogenesis after toluene-in- sure during the brain growth spurt reduced behavioral halation. Leg Med 2005; 7: 42–46. responses to nicotine in young adult rats: a potential 39. Jenkins L, Jones R and Siegel J. Long-term inhalation role for nicotinic acetylcholine receptors in fetal sol- screening studies of benzene, toluene, o-xylene, and vent syndrome. Toxicol Sci 2008; 101: 286–293. cumene on experimental animals. Toxicol Appl Phar- 27. Chen HH, Lin YR and Chan MH. Toluene exposure macol 1970; 16: 818–823. during brain growth spurt and adolescence produces 40. Johnson AC. Auditory sensitivity in rats exposed to differential effects on N-methyl-D-aspartate receptor- toluene and/or acetyl salicylic acid. Neuroreport mediated currents in rat hippocampus. Toxicol Lett 1992; 3: 1141–1144. 2011; 205: 336–340. 41. Martı´nez-Alfaro M, Ca´rabez-Trejo A, Gallegos- 28. Chien TH, Chan MH, Tang YC, et al. Toluene expo- Corona MA, et al. Thinner inhalation effects on sure during the brain growth spurt reduces behavioral oxidative stress and DNA repair in a rat model of responses to noncompetitive N-methyl-D-aspartate abuse. J Appl Toxicol 2010; 30: 226–232. receptor antagonists in adult rats. Psychopharmacol- 42. Ono A, Kawashima K, Sekita K, et al. Toluene inhala- ogy (Berl) 2005; 182: 468–474. tion induced epididymal sperm dysfunction in rats. 29. Dick AL, Axelsson M, Lawrence AJ, et al. Specific Toxicology 1999; 139: 193–205. impairments in instrumental learning following 43. Roberts L, Bevans A and Schreiner C. Developmental chronic intermittent toluene inhalation in adolescent and reproductive toxicity evaluation of toluene vapor rats. Psychopharmacology (Berl) 2014; 231: in the rat: I. Reproductive toxicity. Reprod Toxicol 1531–1542. 2003; 17: 649–658. 30. Dick AL, Lawrence AJ and Duncan JR. Chronic inter- 44. Roberts L, Nicolich M and Schreiner C. Developmen- mittent toluene inhalation initiated during adolescence tal and reproductive toxicity evaluation of toluene in rats does not alter voluntary consumption of ethanol vapor in the rat: II. Developmental toxicity. Reprod in adulthood. Alcohol 2014; 48: 561–569. Toxicol 2007; 23: 521–531. 31. Dick ALW, Pooters T, Gibbs S, et al. NMDA receptor 45. Saillenfait AM, Gallissot F, Sabate´JP, et al. Develop- binding is reduced within mesocorticolimbic regions mental toxic effects of ethylbenzene or toluene alone following chronic inhalation of toluene in adolescent and in combination with butyl acetate in rats after rats. Brain Res 2015; 1624: 239–252. inhalation exposure. J Appl Toxicol 2007; 27: 32–42. 32. Duncan JR, Gibbs SJ and Lawrence AJ. Chronic inter- 46. Schiffer WK, Lee DE, Alexoff DL, et al. Metabolic mittent toluene inhalation in adolescent rats alters correlates of toluene abuse: decline and recovery of behavioural responses to amphetamine and MK801. function in adolescent animals. Psychopharmacology Eur Neuropsychopharmacol 2014; 24: 480–486. 2006; 186: 159–167. 33. Dundaroz M, Sarici S, Turkbay T, et al. Evaluation of 47. Uchino A, Kato A, Yuzuriha T, et al. Comparison bone mineral density in chronic glue sniffers. Turk J between patient characteristics and cranial MR find- Pediatr 2002; 44: 326–329. ings in chronic thinner intoxication. Euro Radiol 2002; 34. Funada M, Sato M, Makino Y, et al. Evaluation of 12: 1338–1341. rewarding effect of toluene by the conditioned place 48. von Euler M, Pham TM, Hillefors M, et al. Inhalation preference procedure in mice. Brain Res Protoc 2002; of low concentrations of toluene induces persistent 10: 47–54. effects on a learning retention task, beam-walk 16 Human and Experimental Toxicology XX(X)

performance, and cerebrocortical size in the rat. Exp 54. Cohen J. Statistical power analysis for the behavioral Neurol 2000; 163: 1–8. sciences. 2nd ed. Hilsdale, NJ: Lawrence Earlbaum 49. Yoon JH, Seo HS, Lee J, et al. Acute high-level toluene Associates, 1988. exposure decreases hippocampal neurogenesis in rats. 55. Bowen SE, Batis JC, Mohammadi MH, et al. Abuse Toxicol Ind Health 2016; 32: 1910–1920. pattern of gestational toluene exposure and early post- 50. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring natal development in rats. Neurotoxicol Teratol 2005; inconsistency in meta-analyses. BMJ 2003; 327: 27: 105–116. 557–560. 56. Australian Institute of Health and Welfare. 2010 51. Vesterinen H, Sena E, Egan K, et al. Meta-analysis of National drug strategy household survey report. Can- data from animal studies: a practical guide. J Neurosci berra: Australian Institute of Health and Welfare, Methods 2014; 221: 92–102. 2011. 52. Callan S, Kott J, Cleary J, et al. Changes in develop- 57. Siervogel RM, Demerath EW, Schubert C, et al. mental body weight as a function of toluene exposure: Puberty and body composition. Horm Res 2003; a meta-analysis of animal studies. Hum Exp Toxicol 60: 36–45. 2016; 35: 341–352. 58. Sumithran P and Proietto J. The defence of body 53. Sengupta P. The laboratory rat: relating its age with weight: a physiological basis for weight regain after human’s. Int J Prev Med 2013; 4: 624–630. weight loss. Clin Sci (Lond) 2013; 124: 231–241. The effects of inhalant abuse on growth: a meta-analysis

Supplementary Figure 1 As toluene concentration increases, weight decreases. Bars represent the effect size for each study with a 95% confidence interval. Where multiple toluene-to-control comparisons (c) existed within a single study, these are noted as c1, c2, etc. The effects of inhalant abuse on growth: a meta-analysis

Supplementary Figure 2 When toluene concentration is differentiated between occupational level exposure and inhalant abuse exposure, weight decreases only for inhalant abuse exposure. Bars represent the effect size for each study with a 95% confidence interval. Where multiple toluene-to-control comparisons (c) existed within a single study, these are noted as c1, c2, etc.