Probable neuropsychiatric toxicity of polyethylene glycol: roles of media, internet and the caregivers

To investigate the nature, origin and extent of reported neuropsychiatric side effects of polyethylene glycol 3350 (PEG, MiraLax©) leading to avoidance of this laxative by parents and caregivers.


| INTRODUC TI ON
A sustained negative perception about Polyethylene glycol 3350 (PEG; MiraLax©) among paediatric caregivers has long been known but never documented. PEG is a well-established treatment of constipation in children. 1,2 Several studies have confirmed its efficacy, safety and superiority over other commonly used laxatives in children. 1,3-6 However, due to a lack of long-term safety data in children and reports of minute amounts of toxic metabolites in the blood, the Federal Drug Administration (FDA) had issued a warning on the safety of long-term use of polyethylene glycol (PEG) in children. 7 The North American Society of Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) put out a statement in 2015 clarifying the FDA statement. 8 In February 2017, television segments and social media postings reported anecdotal neuropsychiatric side effects of PEG in children, such as change in behaviour, anger, etc. [9][10][11][12] Sporadic reports surfaced that many parents and caregivers were now reluctant to use PEG thereby limiting treatment options for constipation in children. This appeared to be a health care controversy with little information on the existence, extent and scope of the problem and scientific data to support either side. In this study, we tried to look objectively at the available information on the extent and genesis of this controversy. Our goal was to document if aversion to use of PEG was real and if so if it was driven by the internet or scientific data.
We analysed the FDA data bank on reported side-effects of PEG; trends, timing and geographical locations of internet activity, as well as the opinions of paediatric gastroenterologists via questionnaire.
To the best of our knowledge, no other report examined the existence and dynamics of PEG refusal as it exists in the public domain.
Studies sponsored by FDA or National Institute of Health (NIH) do not address this important clinical issue.

| MATERIAL S AND ME THODS
For this study we analysed three different sets of data: We obtained the FDA Adverse Effect Reporting System (AERS) data on PEG 3350 between 2007 and 2017 in subjects between 0 and 21 years of age. For analysis of the data, we divided the symptoms into neurological and non-neurological categories. The neurological side effects were subdivided into purely neurological where there were physical signs and symptoms indicating organic brain disease such as seizure, sensory disturbance, etc, and neuropsychiatric involving psychological or behaviorial components such as anxiety, anger, mood swings, etc.
We did a time line analysis of media reporting and internet activity using Google Search © and Google Trends © tools. with PEG 3350 administration. The second part requested data prospectively from a work week. In this section more details on the use of PEG and adverse effects were sought. A professional bulletin board, personal emails and phone calls were used to request participation.
Those who consented to participate and eligible were issued individualised invitations. The data was downloaded and statistically analysed.

| Statistical analysis
We used descriptive statistics to understand the basic features of the reported adverse effect symptoms by gender and age with Neuro and Non-neuro symptoms. For different Neuro and Non-Neuro symptoms a chi-squared test was done to determine if the variables were related to each other. P < 0.05 was considered as level of significance. Moreover, to control the false discovery rate we performed Benjamini-Hochberg procedure. Adjusting the false discovery rate controls the fact that sometimes small P-values (<5%) happens by chance, which can lead to incorrectly significant association between the variables. In other words, we have controlled the false discovery rate (Type I error). were neurological, and 770 of these were neuropsychiatric in nature (83.8%). A summary of neurological, non-neurological and neuropsychiatric symptoms is presented in Figure 1A-C respectively.

| FDA database analysis
Affective disorder was more common in females (P = 0.004) and in the 6 and 10 years' age group (P = 0.0003). Males were more likely to be autistic (P = 0.0062), especially in those 5 years of age and younger (P = 0.0015).
To explore if the timing of media reports and website searches correlated with an increase in the FDA's reported side effects, we plotted adverse effect report by month and year. The trend clearly identified a marked spike in reporting that coincided with the airing/ publishing of the press reports ( Figure 2).

| Timeline analysis
We found a clear spike in the number of visitors to such websites during February 19-25, 2017 ( Figure 3A). These dates  (Figures 2 and 3A).  were not. These PEG related chemicals were not associated with any significant side effects, especially neuropsychiatric ones. 24 In our study, we collected and analysed available objective data to understand the origin and possible scientific validity of information leading to this phenomenon. We documented, for the first time, that PEG aversion is real; about two third of paediatric gastroenterologists polled reported not using or discontinuing PEG at a caregiver's request for potential side effects. As to the cause of this phenomenon, our study provides an interesting overview. The internet data showed a significant spike of activity that coincided chronologically and geographically (within the US) with the anecdotal media reports. We also noted increased reporting of adverse reactions at We conclude that the recent PEG 3350 aversion was real. It was fuelled, at least in part, by media coverage of anecdotal experience and subsequent amplification by internet and social-media posts.