• No results found

Long-term effects of a synbiotic intervention in ADHD-patients: 18-month follow-up

N/A
N/A
Protected

Academic year: 2022

Share "Long-term effects of a synbiotic intervention in ADHD-patients: 18-month follow-up"

Copied!
45
0
0

Loading.... (view fulltext now)

Full text

(1)

Long-term effects of a synbiotic intervention in ADHD-patients

- 18-month follow-up

Author: Jaqueline Fricke Palmell

Supervisors: Thomas Nordström, Elin Skott &

Catharina Lavebratt

(2)

Abstract

A link between the gut and the brain has been proposed to influence psychiatric disorders. Probiotics have been suggested to modify the gut microbiota and thereby improve autism symptoms in children. Attention deficit hyperactivity disorder (ADHD) has high comorbidity with other neuropsychiatric diagnoses, including autism. This is a follow-up of the first study examining a synbiotic intervention in patients with ADHD (Skott et al., 2019). In the original study, 114 adults participated. In this study, 38 adults were evaluated. The aim was to examine if suggested improvements

remained 18 months post treatment. Specifically, if reductions were detected in comorbid autism symptoms, emotional dysregulation or functional

impairment. The endpoints were measured using questionnaires: Autism- Spectrum Quotient (AQ), Difficulties in Emotion Regulation Scale (DERS- 16) and Weiss Functional Impairment Rating Scale (WFIRS). No

Synbiotic2000-specific effect was detected. Synbiotic2000 and placebo improved emotion regulation and life skill-functioning equally well. More research is needed to draw reliable conclusions.

Keywords

Synbiotic intervention, probiotics, prebiotics, microbiota, ADHD, autism, emotion regulation, functional impairment, lactic acid bacteria,

inflammation, anti-inflammatory

Acknowledgements

I would like to thank Elin Skott and Miranda Stiernborg for the help with the statistical analyses. I would also like to thank Thomas Nordström and

Catharina Lavebratt for valuable guidance. My grateful thanks are also extended to Sven-Olof Söderlund and Jacob Söderlund for their help proofreading.

(3)

Sammanfattning

En koppling mellan magen och hjärnan har i studier antytts påverka psykiatriska tillstånd. Probiotika har föreslagits förändra mag- och

tarmkanalens bakterieflora och därigenom förbättra psykiatriska symtom hos barn med autism. ADHD har hög komorbiditet med andra neuropsykiatriska diagnoser, däribland autism. Detta är en långtidsuppföljning av RCT-studien som var först med att undersöka en synbiotika-intervention hos patienter med ADHD (Skott et al., 2019). I uppföljningen undersöktes 38 av de 114 vuxna som deltagit i originalstudien. Syftet var att undersöka om indikationerna till förbättring höll i sig 18 månader efter studieavslutet. Frågeställningen var om reduktion i komorbida autismsymtom, svårigheter med emotionsreglering eller funktionsnedsättning kunde identifieras. Detta undersöktes genom självskattningsskalor: Autism-Spectrum Quotient (AQ), Difficulties in Emotion Regulation Scale (DERS-16) och Weiss Functional Impairment Rating Scale (WFIRS). Ingen behandlingsspecifik effekt detekterades.

Förbättringar av samma grad identifierades av Synbiotic2000 och placebo, utifrån emotionsreglering samt delskalan färdigheter. Mer forskning på området krävs för att möjliggöra tillförlitliga slutsatser.

Nyckelord

Synbiotika, probiotika, prebiotika, mikrobiota, intervention, ADHD, autism, emotionsreglering, funktionsnedsättning, mjölksyrabakterier, inflammation, anti-inflammatorisk

Tack

Tack till Elin Skott och Miranda Stiernborg för hjälp med statistiska

analyser. Jag vill också tacka Thomas Nordström och Catharina Lavebratt för värdefull vägledning. Jag tackar också Sven-Olof Söderlund och Jacob Söderlund för hjälp med korrekturläsning.

(4)

Table of contents

1 Introduction 1

1.1 Theoretical background 1

1.1.1 ADHD and previous research 2

1.1.2 The original study 6

1.1.3 Aim and research question 8

2 Method 8

2.1 Participants 8

2.1.1 Participant recruitment 8

2.1.2 Participants characteristics 9

2.2 Instrument 10

2.2.1 Questionnaires used in the original study and this follow-up 10

2.2.2 Clinical endpoints 12

2.3 Procedure 12

2.3.1 Randomization and procedure 12

2.3.2 Intervention 13

2.3.3 Statistical analysis 14

2.4 Ethics 16

3 Results 17

3.1 Descriptive statistics 17

3.2 Autism symptoms (AQ) 19

3.3 Emotion reactivity (DERS-16) 20

3.3.1 Statistical analysis for the total scale 21 3.3.2 Statistical analysis of the subscale (goals) 22

3.4 Functional impairment (WFIRS) 23

3.4.1 Statistical analysis for the total scale 23 3.4.2 Statistical analysis for the subscale (life skill) 23

4 Discussion 25

4.1.1 Key findings and interpretation 25

4.1.2 Potential clinical implications and future research 31 4.1.3 Methodological strengths and limitations 31

4.1.4 Conclusions and future research 33

5 References 35

(5)

1 Long-term effects of a synbiotic intervention in ADHD-patients: 18-month follow-up

The aim was to examine if a time-limited intervention with

Synbiotic2000 resulted in long-term effects in adults with ADHD. Therefore, scales from the previous RCT called the BAMBA-study (subsequently referred to as the original study) (Skott et al., 2019) were assessed. In the follow-up, we evaluated scales where tendencies or significant results were indicated in the original study. The research question was if the suggested reductions in comorbid autism symptoms, emotional dysregulation or

functional impairment had persisted 18 months post treatment. In the original study, children (n=68) and adults (n=114) participated. The focus area was designed based on findings including children and adults in the original study, but this follow-up study exclusively evaluated adult participants (n=38).

1.1 Theoretical background

A common treatment for patients with ADHD is pharmacological agents taken on a long-term basis. These medications are reducing core symptoms of inattention and hyperactivity-impulsivity (Fredriksen, Halmøy, Faraone & Haavik, 2012; Fredriksen, Dahl, Martinsen, Klungsøyr, Haavik &

Peleikis, 2014). However, they do not seem to have major impacts on other difficulties related with ADHD, as comorbid psychopathology or emotional dysregulation (Fredriksen et al., 2012). There is documented short-term effects of pharmacological agents treating ADHD, but less is known about effects on long-term basis (Fredriksen et al., 2012).

The commonly used pharmacological agents are associated with negative side effects (Fredriksen et al., 2012; Fredriksen et al., 2014).

Examples of frequently reported side effects are decreased appetite,

headache, difficulties falling asleep, mood instability, agitation and anxiety (Bejerot, Rydén & Arlinde, 2010; Kooij et al., 2010; Adler et al., 2011;

(6)

Buitelaar et al., 2012; Fredriksen et al., 2014). Furthermore, blood pressure and heart rate may increase. Although, the side effects can be mild and transitory (Kooij et al., 2010). To keep in mind, the relevant literature is limited and primarily naturalistic (Fredriksen et al., 2012) and should therefore be interpreted with caution.

Since the pharmacological treatment is directed against core symptoms, and negative side effects are frequently reported as a reason to quit treatment (Skott et al., 2019; Fredriksen et al., 2014), dietary

supplements have been suggested to relieve some ADHD symptoms or related difficulties (Sinn & Bryan, 2007; Heilskov et al., 2015; Skott et al., 2019). Due to the associated side effects, it would potentially benefit many patients if synbiotics could relieve some ADHD symptoms or difficulties associated with ADHD.

A synbiotic is composed of prebiotics and probiotics (Bengmark, 2004). Prebiotics are fibers that provide nutrition for bacteria in the gut.

Probiotics are live microorganisms providing growth or activity for gastro- intestinal (GI) bacteria, that improve the gut flora and give beneficial effects (Bengmark, 2004). The beneficial effects include an increased efficiency of the immunological system and inhibition of pathogenic bacteria that provide a proper balance between pathogens and bacteria (Markowiak & Slizewska, 2017). The balance is essential for a normal function of the organism.

Furthermore, probiotics have known anti-inflammatory properties

(Bengmark, 2004; Olah, Belagyi, Poto, Romics & Bengmark, 2007; Rayes et al., 2007; Giamarellos-Bourboulis, Bengmark, Kanellakopoulou &

Kotzampassi, 2009, Bengmark, 2017) and enhance absorption of vitamins and minerals (Markowiak & Slizewska, 2017).

1.1.1 ADHD and previous research

Inattention and/or hyperactivity-impulsivity are characteristic

features of ADHD (American Psychiatric Association [APA], 2013). ADHD is a neurodevelopmental disorder that often persists during the lifespan

(7)

(Kooij et al., 2010). The estimated prevalence of adults with ADHD is in a range of 3-5 % (Fayyad et al., 2007; De Graaf et al., 2008). The symptoms entail functioning impairments in work or school, social relations, and family settings (APA, 2013; Austerman, 2015). ADHD is also associated with dysregulation of emotions (Barkley, 1997; Kooij et al, 2010; Lugo-Candelas, Flegenheimer, Harvey & McDermott, 2017; Skott et al., 2019). Accordingly, it is argued that emotional dysregulation should be considered as part of the core symptoms in ADHD (Kooij et al., 2010). The characteristic features of ADHD can have more subtle expressions in adults (Kooij et al., 2010).

The pathogenesis for ADHD is not fully understood, but heritability is considered to play a crucial role (Kooij et al., 2010). Although,

environmental factors are also proposed to influence the risk for developing ADHD. ADHD is a disorder with frequent comorbidity, for instance

regarding anxiety disorders, learning disorders and other neurodevelopmental diagnoses (or difficulties related with other neuropsychiatric conditions).

Thus, one common comorbid condition is autism (or autism symptoms) (Kooij et al., 2010; Skott et al., 2019).

As previously stated, some synbiotics have been suggested to modify the gut microbiota and thereby decrease autism symptoms in children (Xiang et al., 2019; Skott et al., 2019). Autism is defined as difficulties with social communication, social interaction and limited repetitive behaviours (APA, 2013). Since autism is a common comorbidity in patients with ADHD (Kooij et al., 2010; Skott et al., 2019) and earlier studies have proposed

improvements in autism symptoms using synbiotica, comorbid autism symptoms were measured in the original study (Skott et al., 2019).

Stimulant medication (methylphenidate and dexamphetamine) is the most effective pharmacological treatment for ADHD (Kooij et al., 2010).

However, an ideal treatment is multimodal. Medical treatment as sole treatment is usually not enough (Kooij et al., 2010). A multimodal treatment may include psychoeducation, cognitive behavioural therapy (CBT) or

(8)

supportive therapy, and pharmacological agents (Kooij et al., 2010).

Psychoeducation is the first step and involves information about ADHD symptoms and related difficulties. Most preferably, psychoeducation should also be suggested to family members (Kooij et al., 2010).

Psychotherapy is neither proposed as sole treatment for adults with ADHD. Although, the therapeutic interventions can be helpful working with acceptance of the disorder, strategies for organising, understanding of symptoms and emotional responses, relationship problems and low self- esteem (Kooij et al., 2010). To keep in mind, psychotherapy is not

considered to ameliorate the core symptoms of ADHD. Instead, the strive of psychological treatment is to develop strategies to improve daily functioning (Kooij et al., 2010). Moreover, CBT has strong evidence for treating some of the comorbidities related with ADHD, with moderate to large effect sizes.

Even so, limited controlled trials have been evaluating the effect of CBT in adults with ADHD (Kooij et al., 2010).

Dietary supplements have been suggested to improve

neuropsychiatric symptoms. For instance, omega-3 fatty acids have been suggested to improve some ADHD symptoms (Chang, Su, Mondelli &

Pariante, 2018). Although, with a small effect size. As earlier stated, autism is a common comorbidity for patients with ADHD (Kooij et al., 2010; Skott et al., 2019). A number of studies have indicated that interventions with synbiotics in children with autism improve GI and psychiatric symptoms (Parracho, Gibson, Knott, Bosscher, Kleerebezem & McCartney, 2010;

Grimaldi et al., 2018; Liu et al., 2019). However, few were blinded randomized trials (Xiang et al., 2019). Most of the previous studies were open-label trials, without a control for placebo effect. Thus, the design implies higher risk for confounding variables, compared to blinded randomized trials.

Even if synbiotic interventions in patients with autism have been studied, the original study was first to examine the effect of a synbiotic

(9)

intervention on comorbid autistic traits in patients with ADHD (Skott et al., 2019).

The gut-brain axis encompasses a bidirectional communication between the gut and the brain. This connection has been proposed to play a role for several psychiatric disorders (Aarts et al., 2017; Kang et al., 2019).

The gut bacterial flora, microbiota, has been identified as a potentially important contributor to the gut-brain axis (Aarts et al., 2017; Kang et al., 2019).

A recent trial indicates that fecal microbiota transplant to humans provide long-lasting reductions in GI and autism symptoms (Kang et al., 2017; Kang et al., 2019). Moreover, some studies on rodents have indicated that fecal transplant from patients with autism, depression or schizophrenia can create behaviors related with the disorders (Kelly et al., 2016; Sharon et al., 2019; Skott et al., 2019; Zheng et al., 2019).

Several studies indicate that there is a positive correlation between psychiatric symptoms and GI symptoms (McKeown et al., 2013; Kang, Wagner & Ming, 2014; Skott et al., 2019). GI symptoms are common in patients with ADHD and autism (McKeown et al., 2013; Kang, Wagner &

Ming, 2014). Some studies suggest a potential role of dysregulation in the immune system in persons with autism (Meltzer & Van de Water, 2017).

Proinflammatory markers have been proposed as increased in persons with ADHD (Mitchell & Goldstein, 2014). Although, the evidence basis regarding the immune system for patients with ADHD is limited (Skott et al., 2019).

Altogether, a proposed hypothesis is that the microbiota might be one mediator to the ADHD etiology, via the gut-brain axis (Aarts et al., 2017;

Skott et al., 2019). Therefore, a recent RCT (Skott et al., 2019) explored if a daily intake of prebiotics and probiotics (Synbiotic 2000) affected the microbiota and improved psychiatric symptoms in children and adults with ADHD. The study was the first to evaluate a synbiotic intervention in

(10)

persons with ADHD. This study is an 18-month follow-up of the aforementioned RCT (Skott et al., 2019).

1.1.2 The original study

In the original study, the active treatment was taken once daily, during a period of nine weeks. The active treatment was Synbiotic2000, which is a lyophilized mixture in powder form that contains four anti- inflammatory fibers and three anti-inflammatory lactic acid bacteria (Kotzampassi, Giamarellos-Bourboulis, Voudouris, Kazamias &

Eleftheriadis, 2006; Skott et al., 2019). The placebo group was taking maltodextrin once daily, during the same period (Skott et al., 2019).

Maltodextrin is commonly used as placebo (Kolida, Meyer & Gibson, 2007;

Waitzberg et al., 2012).

In the original study, questionnaires were evaluated assessing ADHD symptoms, autism symptoms, emotional dysregulation, functional

impairment, and dietary habits. The statistical analysis was made using two- way ANOVA. In the original study, suggestive statistical significance was defined as α=0.10 and statistical significance as α=0.050 for the ANOVA models (Skott et al., 2019).

In summary, there was no treatment specific effect reducing ADHD symptoms, comorbid autism symptoms, emotional dysregulation or

functional impairment in the total sample (Skott et al., 2019). As the researchers expected, autistic traits were common even if the sample consisted of patients with ADHD without an autism diagnosis (Skott et al., 2019). Even if no Synbiotic2000 specific effect was detected in the total sample, there was a significant treatment effect in repetitive and stereotyped behaviors among children (p = .050, ηp2 = 0.060). A tendency for a treatment specific effect was also found in the total SCQ scale (p = .098, ηp2 = 0.043).

Repetitive and stereotyped behaviors are a subset of symptoms of autism (APA, 2013) and constitute a subscale in SCQ. SCQ is a commonly used

(11)

self-reported psychiatric scale that measures autism symptoms in children (Rutter, Bailey & Lord, 2003).

Moreover, there was a tendency for an improvement in life skill- related functioning in children (p = .0.092, ηp2 = 0.045), a subscale in WFIRS. WFIRS is a validated questionnaire measuring functional difficulties in patients with ADHD (Thompson, Lloyd, Joseph & Weiss, 2017).

Sensitivity analyses were made, stratified for low-grade inflammation and ADHD medication. Tendencies of improvement in total autism score and life skill-related functioning were driven by children without ADHD

medication (Skott et al., 2019). However, there was a tendency of

improvement in emotion regulation for adults with low-grade inflammation (p = .0.054, ηp2 = 0.115). A tendency for improvement in goal-directed behaviour was also proposed (p = .0.070, ηp2 = 0.102), for adults with low- grade inflammation. Goal-directed behaviour is a subscale in DERS-16 that measures difficulties in emotion regulation (Bjuereberg et al., 2016). To conclude, all findings from the original sample needs replication in larger samples.

In the original study, a food-frequency questionnaire was used to allow adjustment for confounding by diet, at baseline and post treatment. The food questionnaire was based on the ETICS diet study questionnaire. The questionnaire consisted of items representing common food units or common food groups (Kautto, Ivarsson, Norstrom, Hogberg, Carlsson & Hornell, 2014; Skott et al., 2019). At the 18-month follow-up, a brief version was administered. Furthermore, ADHD symptoms were assessed in the original study, via Adult ADHD Self-Report Scale (ASRS) (Skott et al., 2019). The focus of this study was as previously stated to follow up results where tendencies or significant results were identified in the original study.

Therefore, ADHD symptoms were not assessed in the 18-month follow-up.

(12)

1.1.3 Aim and research question

Whether or not the proposed effects in the original study (Skott et al., 2019) are long-lasting needs to be investigated. As earlier mentioned, the aim of this study was to examine if the time-limited synbiotic intervention

resulted in long-term effects in adults with ADHD. Accordingly, the research question was: Are there any reductions in comorbid autism symptoms and/or emotional dysregulation and/or functional impairments 18 months after the synbiotic intervention?

Due to the results presented from the original study, we hypothesized that the nine-week intervention with Synbiotic2000 had long-term effects on adults with ADHD, regarding emotion regulation.

2 Method

For clarification purposes: when referring to the original study (Skott et al., 2019), the manuscript version that was submitted to Brain, Behavior and Immunity in 2020 was used. Due to the review process, the published version might differ from the submitted manuscript.

2.1 Participants

2.1.1 Participant recruitment

In the original study, participants were recruited via advertisement in a local newspaper and through predefined psychiatric out-patient clinics in Stockholm (Skott et al., 2019). The inclusion criteria for participation were:

an established ADHD-diagnosis based on International Statistical Classification of Diseases and Related Health Problems – tenth revision (ICD-10) or Diagnostic and Statistical Manual of mental disorders – fifth revision (DSM-V), age 5-55 years, unchanged pharmacological treatment during the last four weeks and the ability to read Swedish (Skott et al., 2019).

The exclusion criteria were a confirmed autism diagnosis, a GI-disorder

(13)

diagnosis (IBS excluded), antibiotic treatment during the last six weeks, diabetes and celiac disease (Skott et al., 2019).

In the 18-month follow-up, all adults that participated in the original study were invited to participate. Participants were contacted via text

message. The maximum number of reminders that we had permission for were used.

2.1.2 Participants characteristics

Participants characteristics at 18-month follow-up are illustrated in Table 1.1. In the original study, the adult participants consisted of n=114 (Skott et al., 2019). In the follow-up, which included adults with an ADHD- diagnosis, n=38 participated. Thus, 32 % of the adults from the original study participated.

In the follow-up 76 % were female, compared to 71 % adult females in the original study (Skott et al., 2019). The median age was 37 years in the follow-up, compared to 36 years in the original study. In the original study, the patients were allocated to active treatment (50 %) or placebo (50 %) via computerized randomization (Skott et al., 2019). In the follow-up, 41 % had been part of the active treatment group in the original study, while 59 % had been part of the placebo group.

Table 1.1.

Note. IQR = Interquartile range

In order to compare the demographic data of the follow-up with the data from the original study, a drop-out analysis was performed based on gender, age, and treatment group. The analysis consisted of two Chi-square

Clinical characteristics

Adults (n=37) Median (IQR)/N (%)

Age [years] All 37 (35–40)

Sex Female 28 (76%)

Male 9 (24%)

Treatment Active treatment 15 (41%)

Placebo 22 (59%)

(14)

tests, evaluating age and treatment group, and one t-test evaluating age. The analysis was particularly important due to the small size of the sample.

A Chi-square test of independence was conducted to compare the distribution of the treatment groups in the 18-month follow-up compared to the original study. The distributions of the treatment groups did not differ significantly. X2 (2, N = 114) = 2.339, p = .310.

Another Chi-square test of independence was performed to examine if there were any gender differences between the sample in the 18-month follow-up in comparison with the sample in the original study. The

proportion of gender did not differ significantly between the samples. X2 (2, N = 114) = 2.952, p = .229.

At last, an independent sample t-test was performed to examine the relation between age distribution in the follow-up study compared with the original study. There was no significant difference in the distribution of age.

t(149) = -1.192, p = .235.

In summary, participants did not differ significantly regarding the assessed demographic factors. However, the drop-out analysis was limited to assess only demographic factors. Due to that, no direct comparisons of the studies are possible, and the conclusions should therefore be interpreted with caution.

2.2 Instrument

2.2.1 Questionnaires used in the original study and this follow-up

Commonly used and validated self-reported psychiatric scales were distributed to all participants at baseline (the day before start of treatment), at the end of the nine-week intervention (within two weeks after end of

treatment) and in this 18-month follow-up study (Swanson et al., 2001;

Baron-Cohen, Wheelwright, Skinner, Martin & Clubley, 2001; Rutter et al, 2003; Kessler et al., 2005; Woodbury-Smith, Robinson, Wheelwright &

Baron-Cohen, 2005; Bjureberg et al., 2016).

(15)

As previously stated, a brief diet questionnaire was also administered based on the ETICS diet study questionnaire. The questionnaire consisted of 57 items representing common food units or common food groups (Kautto et al., 2014).

Comorbid autism symptoms were measured using Autism-Spectrum Quotient (AQ) (Baron-Cohen et al., 2001). AQ consists of 50 items with four optional answers: “definitely agree”, “slightly agree”, “slightly disagree” and

“definitely disagree” (Baron-Cohen et al., 2001). The likert scale includes five subscales: social skills, attention switching, attention to detail,

communication, and imagination. Higher scores are assumed to indicate more autism symptoms. To keep in mind, AQ is not diagnostic but can be helpful in identifying autistic traits (Baron-Cohen et al., 2001).

Difficulties with emotion regulation were measured using DERS-16 (Bjureberg et al., 2016). DERS-16 includes 16 items on a five-degree likert scale from 1 = “almost never” to 5 = “almost always”. The functionality of emotions and strategies for responding to emotional distress defined emotion regulation (Bjureberg et al., 2016). DERS-16 is a validated and reliable brief version of the original DERS. DERS-16 consists of five subscales: clarity, goals, impulse, strategies, and non-acceptance. Higher sum scores indicate more emotional dysregulation(Bjuereberg et al., 2016).

WFIRS was used to measure functional impairment. WFIRS contains 50 items and is based on a likert scale with five optional answers, from 0 =

“never or not at all” to 3 = “very often or very much”. “Not applicable” is also a possible answer (Thompson et al., 2017). WFIRS is the only instrument measuring functional impairment that has been validated in patients with ADHD, and it is based on domains usually influenced in ADHD (Thompson et al., 2017). WFIRS consists of the following subscales:

family, school and learning/work, life skills, self-concept, social activities, and risky activities. Higher scores correspond to higher degree of functional

(16)

impairment. A mean sum score of >1.5 was interpreted as clinically impaired (Thompson et al., 2017).

Furthermore, blood samples were collected from the participants at baseline in the original study. Proinflammatory markers were analysed in the blood samples. At the 18-month follow-up, we used the data of the

proinflammatory markers from the original study to adjust for confounding by inflammation at baseline.

2.2.2 Clinical endpoints

The endpoints were changes in autism symptoms (measured by AQ), emotion regulation (assessed using DERS-16) and functional impairment (assessed with WFIRS).

2.3 Procedure

2.3.1 Randomization and procedure

Questionnaires were administered at baseline (the day before start of treatment), post treatment (within two weeks after end of treatment) and in this 18-month follow-up. Active treatment and placebo lasted for nine weeks (Skott et al., 2019). Initially, patients got information about the study from a care provider. Afterward, three psychiatric research nurses examined if inclusion and exclusion criteria were met. The nurses performed interviews to examine if inclusion and exclusion criteria were met, recruitments,

measurements at baseline and post treatment, and treatment allocation (Skott et al., 2019). The randomization was performed by the Karolinska Trial Alliance, an independent unit. If initial inclusion criteria were met and if informed consent was provided, participants were randomized to either the active treatment group or to the placebo group in an allocation ratio 1:1, using block randomization. The treatments were parallel. Allocation structure, block size and individual distributions were concealed for all people involved in the study (Skott et al., 2019). At baseline, the participants answered questionnaires (ASRS, AQ, DERS-16, WFIRS and the diet

(17)

questionnaire) and provided a blood sample. Furthermore, information from participants was obtained about childbirth delivery, breastfeeding, body mass index (BMI) and recently used or current pharmacological treatment (Skott et al., 2019). Post treatment, the same questionnaires were administered.

18 months post treatment, patients were invited via text message to participate in the follow-up study. The messages included information about participation, General Data Protection Regulation (GDPR) and an inserted link where patients could decline participation. The text messages were sent based on identification numbers from the original study. The participants in the 18-month follow-up answered the questionnaires digitally using a web- based survey site. As previously noted, all questionnaires corresponded to those administrated at baseline and post treatment. The focus of this study was to follow up results where tendencies or significant results were identified in the original study. Therefore, ADHD symptoms were not assessed at the 18-month follow-up.

We sent not more than three reminders about participation to each person, which was the maximum number of reminders that we had

permission to send. Invited participants were given the optional response “do not want to participate”. Those who responded by that option, where not reminded any more. In total, 38 persons participated in the 18-month follow- up. Thus, the participation rate was 32 % based on the number of adults who participated in the original study.

2.3.2 Intervention

The active treatment (Synbiotic2000) and placebo (Maltodextrin) had the form of a powder. The treatments were taken orally for nine weeks, once a day. The same instructions were given to all participants (Skott et al., 2019). The participants were instructed not to change their diet. Furthermore, they were asked to fill out a questionnaire to note missed treatment days.

Post treatment, the questionnaires were submitted to the nurses. The

(18)

participants were administered sachets with powder of Syntibiotic2000 or maltodextrin for two to three weeks at a time.

Synbiotic2000 and placebo were matched based on packaging, volume, weight, colour of content, texture, flavour and had neutral fragrance.

Synbiotic2000 is a mixture of three lactic acid bacteria and four different fibers (Skott et al., 2019). Maltodextrin was used as placebo. Maltodextrin has no effect and is frequently used in controlled trials (Kolida, Meyer &

Gibson, 2007; Waitzberg et al., 2012).

2.3.3 Statistical analysis

At first, all the data were extracted from the web-based survey site to Excel to be prepared for the statistical analysis. The sum and the mean for each scale and subscale were calculated in Excel. Thereafter, all data from the 18-month follow-up were imported to SPSS where each scale was merged with data of the same scale from the original study. Boxplots were conducted based on data at baseline in comparison with data at 18-month follow up, and for all three timepoints (baseline, post treatment, and 18 months post treatment). Means for the demographic data were calculated in Excel. Before conducting statistical analyses in SPSS, we examined if the data set for each scale and subscale met the assumptions for ANOVA.

All analyses were conducted in IBM’s SPSS statistic software (version 26). The data at baseline were compared with the assessments 18 months post treatment through a two-way repeated measures ANOVA with a between-group analysis. Accordingly, the changes of symptoms between the treatment groups were evaluated based on within-subject effects

(time*treatment).

In addition, we controlled for the following covariates: age-group, gender and a pro-inflammatory marker called vascular cell adhesion molecule (VCAM). The covariates were also estimated based on within- subject effects (time*age-group, time*sex, time*VCAM). If the covariates

(19)

were assed not to impact the scale (AQ, DERS-16 or WFIRS), they were excluded. In that case, a new ANOVA was performed with remaining covariates. The cut-off for excluding covariates was set to p >0.40.

In order to evaluate the effect on symptoms of Synbiotic2000, comparisons were made between the active treatment group and the placebo group. Statistical analysis was performed for three different scales: AQ, DERS-16, and WFIRS. Additionally, two subscales were analysed: life skill- functioning (WFIRS) and goal related emotion regulation (DERS-16). When statistical assumptions for ANOVA were not met, supplementary tests were conducted with the statistical test Mann-Whitney U.

The effect sizes that are reported from the ANOVA models are based on partial eta squared (ηp2). However, the effect sizes reported from Mann- Whitney U-tests are based on eta squared (η2) and was calculated by hand with the following equation: η2 = Z2/N-1. Partial eta squared was evaluated based on Cohens (1992) proposed rule of thumb which assumes: small effect size (ES) = 0.01, medium ES = 0.06 and large ES = 0.14 (Cohen, 1992;

Levine & Hullett, 2002). Eta squared was assessed as: small η2 = 0.10, medium η2 = 0.30 and large η2 = 0.50 (Cohen, 1988).

In the original study, suggestive statistical significance was defined as α=0.10 and statistical significance as α=0.050 for the ANOVA models (Skott et al., 2019). Since this is a follow-up, the cut-off values remain unchanged to avoid analytical flexibility.

The covariate age-group were pre-defined in the original study to 19- 25 and 26-55 years due to knowledge of possible change in ADHD

symptoms during lifespan (Philipp-Wiegmann, Retz-Junginger, Retz &

Rösler, 2016).

In the original study, sensitivity analyses were performed based on the pro-inflammatory marker (VCAM) and ADHD-medication. At first, we planned to perform sensitivity analysis based on the same pro-inflammatory marker in the follow-up. However, the sample was too small to perform valid

(20)

sensitivity analysis. Therefore, we decided to control for VCAM as a

covariate in the ANOVA models. Information about ADHD-medication was not collected in the 18-month follow-up, which was the reason not to control for ADHD-medication in the ANOVA models.

2.4 Ethics

Approval for all procedures was obtained from the Stockholm Ethics Review Board, including the procedures for the 18-month follow-up

(2015/884-31, 2017/771-32) (Skott et al., 2019). The procedures were in line with ethical standards of committees on human experimentation and with the Helsinki declaration (World medical association, 1975). Information was provided to patients about voluntary participation, and the possibility to at any time stop initiated participation without further explanation. Oral and written informed consent was obtained. Moreover, all participants were informed about the purpose of the study.

A serial number was assigned to each participant, with an associated identification number that was inaccessible and locked in. The same numbers were used in the 18-month follow-up. The data analyses were performed at the centre of molecular medicine (CMM) at Karolinska Institute (KI), to protect personal data associated with identification numbers. In other words, the data material was safely stored at KI.

One guideline of The Swedish Research Council (2017) is striving to conduct research without doing harm. Thus, one scientific hesitation in this study was the collected blood samples. However, the type of blood sample that was used is associated with low risk. Furthermore, local pain relief was offered and participation without giving blood sample was offered. The advantages of the study were also considered, such as possible benefits regarding ADHD symptoms, emotion regulation and daily functioning.

The voluntary participation (The Swedish Research Council, 2017) can be questioned as psychiatric patients is a vulnerable group. This risk was

(21)

minimized through providing information and obtaining informed consent (orally and written). Another way to minimize the risk was to clarify that declined participation would not affect patient’s ordinary health care in any way. Furthermore, samples were not administered by patient’s ordinary care provider.

To enable reporting transparent procedures and results, this paper describes the procedures and results from both the original study and the 18- month follow-up. The purpose of summarising the original study is to facilitate understanding of results and procedures in the follow-up, to encourage openly accounting (The Swedish Research Council, 2017). To ensure that the research was organized (The Swedish Research Council, 2017) procedures were documented through the process. When the analyses were finished, documents were submitted to the research group at KI. Also, GDPR has been complied with (Skott et al., 2019).

3 Results

The 18-month follow-up strives to answer the following research question: Are there any reductions in comorbid autism symptoms and/or emotional dysregulation and/or functional impairment, 18 months after the synbiotic intervention in adults with ADHD? In order to answer the research question, two-way repeated measures ANOVA are reported. Results for the ANOVA models are reported for one scale at the time (AQ, DERS-16, WFIRS).

3.1 Descriptive statistics

Descriptive statistics for each scale (AQ, DERS-16, WFIRS) and subscale (goal-directed behaviour and life skill-related functioning) are illustrated in Table 2.1

(22)

Table 2.1.

Descriptive statistics

Treatment M SD N

AQ

Baseline Placebo 0.42 0.13 18

Active treatment 0.33 0.11 14

Total 0.38 0.13 32

Follow-up Placebo 0.39 0.13 18

Active treatment 0.34 0.13 14

Total 0.37 0.13 32

DERS-16

Baseline Placebo 49.59 14.04 17

Active treatment 40.54 14.25 13

Total 45.67 14.61 30

Follow-up Placebo 41.35 18.69 17

Active treatment 33.08 16.46 13

Total 37.77 17.95 30

Goals

Baseline Placebo 12.76 2.63 17

Active treatment 9.69 3.47 13

Total 11.43 3.35 30

Follow-up Placebo 10.12 4.52 17

Active treatment 7.92 4.44 13

Total 9.17 4.54 30

WFIRS

Baseline Placebo 0.98 0.33 18

Active treatment 0.84 0.19 14

Total 0.92 0.28 32

Follow-up Placebo 1.02 0.37 18

Active treatment 0.71 0.23 14

Total 0.88 0.35 32

Life skill

Baseline Placebo 1.32 0.43 18

Active treatment 1.14 0.38 14

Total 1.24 0.41 32

Follow-up Placebo 1.14 0.56 18

Active treatment 1.11 0.56 14

Total 1.13 0.55 32

(23)

3.2 Autism symptoms (AQ)

In the 18-month follow-up, changes of autism symptoms (assessed with AQ) were evaluated in adults with an established ADHD-diagnosis. The AQ scores were compared between the treatment and placebo group, based on assessments at baseline and at 18-month follow-up. The statistical analysis was performed for the total AQ scale through a two-way repeated measures ANOVA.

The assumptions for ANOVA were met for the data set that constituted the total scale of AQ. Accordingly, a two-way ANOVA was conducted to examine if there were any differences between the treatment groups regarding change in autism symptoms over time. There was no significant interaction between time and treatment regarding autism symptoms, F (1, 27) = 0.781, p = .385, ηp2 = 0.028. Neither was there any main effect for time (p = .949, ηp2 < 0.001) or for group (p = .078, ηp2 = 0.111).

The interaction between time*treatment, regarding comorbid autism symptoms, is illustrated in Figure 2.1.

(24)

Figure 2.1. Assessments of AQ scale at (1) baseline and (2) 18 months post treatment, for (A) placebo and (B) active treatment

To conclude, no Synbiotic2000-specific effect was identified for comorbid autism symptoms in adults with ADHD, 18 months post treatment.

In other words, there was no significant difference in effect between the active treatment group and the placebo group, in total AQ scale. It should be noted that the treatment groups differ at baseline, where placebo has higher values per scale question, in comparison with the active treatment group.

3.3 Emotion reactivity (DERS-16)

Changes in emotion regulation (assessed with DERS-16) were measured to evaluate if there were any Synbiotic2000 specific reduction in emotion reactivity among adults with ADHD, 18 months post treatment.

Statistical analysis was performed for the total scale of DERS-16, and for a subscale (goals) assessing the ability to execute goal-directed behaviour under emotional distress. This subscale was evaluated since a tendency of improvement was proposed for adults in the original study, in goal-directed behaviour. As previously stated, the focus area in this study was following up

(25)

results where tendencies or significant results were identified in the original study.

The assumptions for ANOVA were met for the data set that constituted the total scale of DERS-16. However, the data for the subscale did not meet the assumptions for ANOVA. Accordingly, a two-way repeated measures ANOVA was conducted for total DERS-16, whereas the

supplementary test of Mann-Whitney U was performed for goal-directed behaviour. The two analyses are reported separately below.

3.3.1 Statistical analysis for the total scale

A two-way ANOVA was conducted to examine if there were any differences between the treatment groups regarding changes in emotion regulation over time. There was no significant interaction between time and treatment based on changes in emotion regulation, F (1, 27) = 0.069, p = .795 ηp2 = 0.003. However, there was a statistic significant main effect for time (p

= .029, ηp2 = 0.165). The effect size (ηp2 = 0.165) is interpreted as large, based on Cohens (1992) rule of thumb (which assumes small ES = 0.01, medium ES = 0.06 and large ES = 0.14). There was no significant main effect for group (p = .056, ηp2 = 0.128).

The interaction between time*treatment considering emotion regulation, is illustrated in Figure 2.2.

(26)

Figure 2.2. Measurements of the total DERS-16 scale at (1) baseline and (2) 18 months post treatment, for (A) placebo and (B) active treatment

Based on the visual statistics, the significant main effect for time (p = .029, ηp2 = 0.165) seems to indicate an equal improvement in emotion

regulation for the active treatment group and the placebo group, 18 months post-treatment.

The sum scores at baseline differ between the placebo group and the active treatment group, where participants in the placebo group score higher sum scores at baseline.

3.3.2 Statistical analysis of the subscale (goals)

A Mann-Whitney U test was performed to investigate if there were any differences between active treatment and placebo regarding goal-related behaviour under emotional distress.

The Mann-Whitney U test indicated that there was no significant difference between the active treatment group and the placebo group

(27)

regarding sum scores in goal-related behaviour in DERS-16 (U = 86.550, z = -0.444, p = .657, η2 = 0.007).

To sum up, no Synbiotic2000 specific reduction was found in emotion reactivity 18 months post treatment, neither in the total scale (DERS-16) or in the subscale (goals).

3.4 Functional impairment (WFIRS)

The final analyses conducted in the 18-month follow-up were evaluating the questionnaire WFIRS in order to assess if there were any changes regarding functional impairment in the treatment groups. Two statistical analyses were performed, one for the total scale and one for the subscale (life skill). The subscale examines life skill-related functioning in patients with ADHD. Due to assessments of the statistical assumptions for ANOVA, one Mann-Whitney U test and one two-times repeated measures ANOVA were performed.

3.4.1 Statistical analysis for the total scale

A Mann-Whitney U test was performed to examine if there were any differences between the active treatment group and the placebo group based on total scores of WFIRS (assessing functional impairment in domains usually affected in ADHD) 18 months post treatment.

The Mann-Whitney U test indicated that there was no significant difference between the active treatment group and the placebo group based on the total WFIRS-scale (U = 114.000, z = -1.200, p = .230, η2 = 0.042).

3.4.2 Statistical analysis for the subscale (life skill)

A two-way repeated measures ANOVA was conducted to assess if there were any difference between the treatment groups in life skill-related functioning, 18 months post treatment. Life skill-related functioning includes abilities such as keeping clean, taking care of hygiene, eating, and sleeping without difficulties.

(28)

Based on the ANOVA model, there was no significant interaction between time and treatment based on changes in life skill-related functioning in adults with ADHD, F (1, 29) = 0.690, p = .413, ηp2 = 0.023. However, there was a significant main effect for time (p = .0050, ηp2 = 0.127). The effect size (ηp2 = 0.127) is interpreted as medium-sized, based on Cohens (1992) rule of thumb (which assumes small ES = 0.01, medium ES = 0.06 and large ES = 0.14). There was no significant main effect for group (p = .396, ηp2 = 0.025).

The interaction between time*treatment in the subscale of life skill- related daily functioning is illustrated in Figure 2.3.

Figure 2.3. Assessments of the subscale WFIRS life skill scale at (1) baseline and (2) 18 months post treatment, for (A) placebo and (B) active treatment

Accordingly, no Synbiotic2000 specific reduction was indicated for functional impairment 18 months post treatment. This was the case for both the total scale and the subscale measuring life skill-related functioning.

However, a medium-sized main effect was found for time which indicates

(29)

reduction in functional impairment for both treatment groups. As previously, the means at baseline were higher in the placebo group than the active treatment group.

To conclude, no Synbiotic2000 specific effect was found for reduction in autism symptoms, emotion reactivity or functional impairment (AQ, DERS-16, WFIRS) at the assessments 18 months post treatment. Thus, no support was found for long-term effects in reduction of emotion reactivity in favour of the active treatment group, which was hypothesized. However, improvements in emotion regulation (ptime=.029, ηp2 = 0.165) and life skill- related functioning (ptime = .0050, ηp2 = 0.127) were indicated independently of treatment group.

4 Discussion

4.1.1 Key findings and interpretation

This is the first long-term follow-up of a randomized placebo-

controlled trial exploring the effects of a synbiotic intervention on symptoms and functional impairments in patients with ADHD. The aim of the study was to examine if there are long-term effects of a time-limited intervention with Synbiotic2000 in adults with ADHD. The results of the 18-month follow-up did not provide support for Synbiotic2000 specific reductions at group level, neither in comorbid autism symptoms, emotion reactivity or functional impairments. Thus, no interaction for time*treatment was detected, which means that no significant difference in effect was indicated for the treatment groups. However, there was a main effect for time for the total scale of DERS-16 (p = .029, ηp2 = 0.165) and the subscale of WFIRS (p

= .005, ηp2 = 0.127). In other words, reduction in symptoms was seen both in the active treatment group and the placebo group, regarding emotion

reactivity and life skill-related functional impairment. Accordingly, those reductions could not be interpreted as a specific treatment effect of

(30)

Synbiotic2000, since reduction occurred in both treatment groups. Therefore, the Synbiotic2000 specific effect reported in the original study (Skott et al., 2019) regarding emotion regulation in adults (including goal-directed behaviour) did not seem to persist in the 18-month follow-up.

It should be noted that the scores at baseline differed for the active treatment group and the placebo group in the total scale of DERS-16, where placebo had higher values at baseline. As previously reported, maximum sum score for DERS-16 is 80, where higher points indicate more difficulties with emotion regulation (Bjuereberg et al., 2016). Based on the assessments performed at baseline, the estimated mean of the placebo group was 49.59 points, whereas the mean of the active treatment group was a sum score of 40.54. At the assessment 18 months post treatment, the mean of the placebo group was 40.28 points, whereas the mean for active treatment was 31.55 points. One possibility is that it could have been more difficult to detect reduction in the active treatment group due to the tendency of higher values at baseline for the placebo group. Although, the confidence interval for the estimated means did overlap.

An independent sample t-test was performed to examine if there were any statistical difference between the values at baseline. Based on the

statistical analysis, there was no significant difference between the active treatment group and placebo in the total scale of DERS-16 at baseline t(28)=1.738, p = .093. Thus, the difference between the groups at baseline was not significant.

It should be noted that no interaction between time*treatment was found (p = 0.795, ηp2 = 0.003). Therefore, the difference could be explained by the drop-out, due to the lack of performed intention to treat (ITT). Also, the small size of the sample can cause a more uncertain randomization. Due to controlling for demographic variables in the drop-out analysis, it is not likely that the tendency of difference could be explained by difference in demographic variables.

(31)

In the original study, a tendency for Synbiotic2000 specific effect in emotion regulation was found among adults with low-grade inflammation (p

= .054, ηp2 = 0.115). The effect size was interpreted as medium large based on Cohens (1992) rule of thumb. Thus, low-grade inflammation was stratified for in the sensitivity analysis in the original study. As previously stated, we could not perform corresponding sensitivity analysis in the long- term follow-up, due to the small size of the sample. As an alternative, we controlled for low-grade inflammation as a covariate.

Due to the absence of sensitivity analysis for low-grade inflammation in the 18-month follow-up, it cannot be ruled out that patients with low-grade inflammation at baseline had a persistent effect of Synbiotic2000, even if the results of the 18-month follow-up did not detect any significant difference in the general sample. Since ADHD and autism are suggested to be closely related neuropsychiatric disorders (Sokolova et al., 2017), this possibility is enhanced by the fact that long-term effects have been suggested in a recent long-term follow-up examining effects of a fecal microbiota transplant in autism patients with gastrointestinal problems (Kang et al., 2019). Even though the intervention for microbiota was differently designed than the intervention in the original study (Skott et al., 2019), the 24-month follow-up in the Kang et al. (2019) study suggests that improvements in GI symptoms were maintained whereas autism-related symptoms improved even more.

The suggested improvements involved autism related symptoms and the severity of autism, both dimensions assessed by the Childhood Autism Rating Scale (CARS). The microbial diversity significantly increased (including beneficial microbes) which may reflect that the intervention successfully altered the gut microbiota healthier and, in that way, generated long-term improvements in GI and psychiatric symptoms (Kang et al., 2019).

As in the original study (Skott et al., 2019) the researchers (Kang et al., 2019) suggested a link between the gut and brain via the gut-brain axis modulating the gut microbiota by probiotics and prebiotics and/or microbiota

(32)

transplant. As previously stated, the intervention used in the original study was an orally administered mixture of prebiotics and probiotics

(Synbiotic2000). Although, the aim of the fecal transplant intervention is the same as the oral synbiotic intervention: to improve microbiota and in that way affect behaviour via the proposed gut brain axis (Kang et al., 2019).

Furthermore, it should be noted that Kang's study (2019) is an open-label trial with 18 participants. Thus, the results should be interpreted with caution.

As far as we know, the follow-up by Kang et al. (2019) is the first published long-term result of a microbial intervention assessing psychiatric symptoms.

Even if long-term effects have not been evaluated, there are several trials suggesting effects of dietary supplementations on ADHD symptoms (Sinn & Bryan, 2007; Heilskov et al., 2015). For example, one trial examined effects of a 15-week polyunsaturated fatty acids- and micronutrients-

supplementation on children with ADHD. This study suggested an effect on core ADHD symptoms based on Conners parent rating scales (Sinn & Bryan, 2007). Furthermore, omega-3 had shown stable, but small, positive effects on ADHD symptoms (Helskov et al., 2015).

To conclude, there was no Synbiotic2000-specific reduction in emotion reactivity, but it is possible that patients with low-grade inflammation had persistent effects since stratified analysis was not conducted. If so, this could be related to Synbiotic2000 known anti-

inflammatory properties (Bengmark, 2004; Olah, Belagyi, Poto, Romics &

Bengmark, 2007; Rayes et al., 2007; Giamarellos-Bourboulis, Bengmark, Kanellakopoulou & Kotzampassi, 2009). Accordingly, the tendency of Synbiotic2000 specific effects (p = .054, ηp2 = 0.115) that was suggested for emotion regulation in adults in the original study, were primarily driven by patients with low-grade inflammation at baseline (Skott et al., 2019).

Another possible reason for not detecting significant effects of the orally administered Synbiotic2000 intervention could be that this kind of time-limited synbiotic intervention does not provide long-lasting effects in

(33)

patients with ADHD. As the research field of synbiotic interventions in the psychiatric field is quite new, there are yet limited published long-term results. Since Kang et al. (2019) suggested results in GI- and psychiatric symptoms of a fecal microbiota transplant, it is also possible that such microbiotic interventions are more effective than the orally administered synbiotics. Thus, the lack of detected long-term effects as opposed to Kang et al. (2019) findings, could be explained by differences in effect of the diverse forms of microbiotic interventions. For example, a recent review has stated that the composition of prebiotics and probiotics is crucial (Markowiak &

Slizewska, 2017). Also, the effectiveness of one specific composition (of prebiotics, probiotics or synbiotic) may vary due to which condition is treated (Markowiak & Slizewska, 2017). However, several studies have shown promising results for orally administered synbiotic, even if long-term assessments have not been performed (Bengmark, 2004; Kotzampassi, 2006;

Parracho et, 2010; Plaudis, Pupelis & Zeiza, 2012; Markowiak & Slizewska, 2017; Grimaldi et al., 2018). Even though it is possible that fecal

interventions have more persisting effects than the orally administered synbiotic intervention used in the original study. To conclude, it is uncertain to draw reliable conclusions based on the minor sample size and the limited evidence base (Kang et al., 2019).

Several studies have shown promising effects for probiotic interventions in children with autism (Xiang et al., 2019) whereas the

original study is the first study evaluating effects for children and adults with ADHD (Skott et al., 2019). Thus, another possibility is that effects of

synbiotic interventions differ in patients with ADHD and autism (Kang et al., 2019), even if the diagnoses are suggested being closely related (Markowiak

& Slizewska, 2017; Sokolova et al., 2017).

Moreover, results from a systematic review evaluating the role of prebiotics and probiotics in autism spectrum disorders indicate that synbiotic interventions are effective on children (Xiang et al., 2019). The long-term

(34)

follow-up by Kang et al. (2019) also evaluated effects in children with autism. Even in the original study (Skott et al., 2019) more effects were indicated for children than for adults. Altogether, this could reflect a limited window of treatment to young age for these interventions, as proposed in rodent models for treatment of autistic traits (Hsiao, 2013). Although, the findings need replication in human trials.

As previously stated, no Synbiotic2000 specific reduction was found in adults regarding functional impairment (p = .230) or autism symptoms (p

= .385) at the 18-month follow-up. The results are in accordance with those in the original study, where no significant differences in functional

impairment (p = .905) or autism symptoms (p = .754) were found (Skott et al., 2019).

In the original study, a tendency for Synbiotic2000 specific effect in total AQ scale was identified in children with low-grade inflammation (p = .073, ηp2 = 0.091). Moreover, a statistically significant reduction was found for repetitive and stereotyped behaviour (p = .050, ηp2 = 0.060). The fact that no specific effect was detected for autism symptoms in adults, neither in the original study nor in the follow-up, could also be due to the possibility for a window of treatment (Hsiao, 2013). Although, the interesting findings in autism symptoms in children (Skott et al., 2019) were not investigated. As previously stated, this study exclusively centred on adult participants.

Accordingly, it is not yet possible to draw conclusions about long-term effects in children.

To conclude, the absence of detected effect for adults in the 18-month follow-up goes in line with previous studies whereas effects of dietary

supplementation primarily have been shown on children (Sinn & Bryan, 2007; Parracho et al., 2010; Heilskov et al., 2015; Chang, 2018; Liu et al., 2019; Grimaldi et al., 2018).

(35)

4.1.2 Potential clinical implications and future research

Based on the results of this 18-month follow-up, it is too early to recommend Synbiotic2000 as a treatment option for adults with ADHD.

More research is needed to evaluate long-term effects of Synbiotic2000, and other analogous interventions for microbiota, particularly in adults with ADHD. To enable Synbiotic2000 as an option for treatment, more results showing specific effects for ADHD symptoms, comorbid symptoms, or related difficulties are needed. Due to promising results of microbiotic interventions in children (Sinn & Bryan, 2007; Parracho et al., 2010;

Heilskov et al., 2015; Chang, 2018; Grimaldi et al., 2018; Liu et al., 2019) it would be interesting to examine long-term effects on children with ADHD participating in the original study. If future results would confirm Synbioitc- 2000 specific effects, it could be a valid treatment option or addition for patients having negative side effects of pharmacological agents or having GI problems. Especially due to results indicating better treatment effects in patients without medication and with GI problems (Kang et al., 2019, Skott et al., 2019).

4.1.3 Methodological strengths and limitations

The main strength of the original study is the placebo-controlled design. Furthermore, the treatment time of nine weeks was long for being a probiotic intervention trial in neuropsychiatry (Skott et al., 2019). One strength in the follow-up is the performed drop-out analysis, providing information about demographic factors showing similar baseline

characteristics in the original study and this study. Another strength is the strive to use the same statistical analyses and procedures as in the original study, to the extent possible. Nevertheless, the study has limitations. First, the reason for the adjusted alpha-levels is ambiguous. As previously stated, suggestive statistical significance was defined as α=0.10 for the ANOVA models (Skott et al., 2019). Commonly, statistical significance (α=0.05) is

(36)

reported without suggestive statistical significance. Reporting tendencies of improvements may increase the risk of type 1 errors which implies the rejection of a true null hypothesis (Wilson & MacLean, 2011). Thus, it increases the risk of reporting false positive results. Even so, the alpha-levels were remained in the follow-up to avoid more analytical flexibility. It was considered difficult following up results with other alpha-levels than in the original study.

Furthermore, nutrient intake and medication were controlled for in the original study, whereas they were not analysed in the follow-up. This is a limitation due to the risk of changes in diet (Heilskov et al., 2015) or changes in pharmacological treatment as confounding variables. Moreover, the sample was limited which complicates the possibility of conclusive evidence due to the risk of uncertainty in the analysis. Another limitation was the restriction of only following up adult participants. The cause was limited time for statistical analyses. For the same reason, the focus was to follow up results where tendencies or significant results were identified in the original study, to examine if the suggested effects had persisted. This is a limitation because of the possibility that effects can emerge some time after the end of treatment. Therefore, we cannot draw any conclusions about long-term effects of core ADHD symptoms, since results of Adult ADHD Self-Report scale (ASRS) were not followed up. The same fact applies to subscales other than WFIRS life skill and DERS-16 goals. Furthermore, the absence of intention to treat analysis (ITT) makes it difficult to perform direct

comparisons of the results in the original study and the results in the follow- up study.

Even if the psychiatric scales of use are considered as validated and reliable (Baron-Cohen et al., 2001; Swanson et al., 2001; Rutter et al, 2003;

Kessler et al., 2005; Woodbury-Smith et al., 2005; Bjureberg et al., 2016) it is still risks for errors of measurement in self-reported psychiatric scales.

(37)

Furthermore, it is also possible that the scales could not provide enough sensitivity to detect treatment effects.

As previously reported, the link for participation was sent via text message. That means that we could not control the environment for participation. One of the core symptoms of ADHD is inattention (APA, 2013) which also could have implied the risk for concentration difficulties while responding the questionnaires in the long-term follow-up. Furthermore, reading difficulties in general and dyslexia in particular are common

comorbidities in patients with ADHD (Germanò, Gagliano & Curatolo, 2010). Thus, the self-reported psychiatric scales could have been an

aggravating factor. We cannot be sure that the participants fully understood the questions. Thereby, the validity may have been adversely affected. A solution could have been to interview the participants, which could have provided the possibility to control the study environment and answer possible questions about the content in the psychiatric scales.

As reported in the descriptive statistics, another limitation is that all participants did not complete every scale.

Altogether, the scientific evidence of this study is not conclusive which clarifies the need for replication in larger samples to explore effects of Synbiotic2000 in patients with ADHD.

4.1.4 Conclusions and future research

The 18-month follow-up did not detect Synbiotic2000-specific reductions in comorbid autism symptoms, functionality or emotion

regulation in adults with ADHD. Therefore, the results of this study do not provide support for specific effects of the active treatment (Synbiotic2000).

However, both treatment groups improved in emotion regulation and life skill-related functioning. More research is needed in larger samples to examine effects of orally administered synbiotic interventions and evaluate the possibility for Synbiotic2000 (and other synbiotic interventions) as a

(38)

treatment option in patients with ADHD. Because of the absence of sensitivity analysis regarding low-grade inflammation in the 18-month follow-up, it would be interesting to explore if adults with ADHD and low- grade inflammation got persisting effects of the treatment with

Synbiotic2000. Especially because of recently suggested long-term effects in autism patients with GI problems, after a microbiotic intervention (Kang et al., 2019). Due to previous research suggesting synbiotic effects for children, and more specific effects for children than for adults in the original study, and the absence of long-term follow-up, it would be particularly interesting to explore long-term effects in children with ADHD.

If orally administered synbiotics could ameliorate some ADHD- symptoms or reduce related difficulties that would potentially benefit many patients. Especially patients who experience negative side effects of the medical treatment or do not take ADHD-medication.

References

Related documents

Stöden omfattar statliga lån och kreditgarantier; anstånd med skatter och avgifter; tillfälligt sänkta arbetsgivaravgifter under pandemins första fas; ökat statligt ansvar

46 Konkreta exempel skulle kunna vara främjandeinsatser för affärsänglar/affärsängelnätverk, skapa arenor där aktörer från utbuds- och efterfrågesidan kan mötas eller

This result becomes even clearer in the post-treatment period, where we observe that the presence of both universities and research institutes was associated with sales growth

För att uppskatta den totala effekten av reformerna måste dock hänsyn tas till såväl samt- liga priseffekter som sammansättningseffekter, till följd av ökad försäljningsandel

To gain knowledge about how the students perceive the study environments that were refurbished in 2015- 2016, a survey with focus on a typology of learning space preference

The 1995 study used the Conflict Tactics Scale to measure physical and emotional violence by a current or former intimate partner, with an additional question on lifetime

Then, we related fear memory savings (e.g. [19]) to initial amygdala activity to deter- mine if the long-term effects of undisrupted reconsolidation can be predicted from activity

IS B N 9 78 -9 1- 62 9 -0 35 4 -1 (PR IN T )ISBN 978-91-629-0355-8 (PDF) http://hdl.handle.net/2077/54527Printed by BrandFactory, Gothenburg SAHLGRENSKA AC ADEMYINSTITUTE OF