Journal ofhttp://jad.sagepub.com/ Attention Disorders Vitamin−Mineral Treatment of ADHD in Adults: A 1-Year Naturalistic Follow-Up of a Randomized Controlled Trial Julia J. Rucklidge, Chris M. Frampton, Brigette Gorman and Anna Boggis Journal of Attention Disorders published online 7 May 2014 DOI: 10.1177/1087054714530557 The online version of this article can be found at: http://jad.sagepub.com/content/early/2014/05/07/1087054714530557
Published by: http://www.sagepublications.com
Additional services and information for Journal of Attention Disorders can be found at: Email Alerts: http://jad.sagepub.com/cgi/alerts Subscriptions: http://jad.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jad.sagepub.com/content/early/2014/05/07/1087054714530557.refs.html
>> OnlineFirst Version of Record - May 7, 2014 What is This?
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
530557
research-article2014
JADXXX10.1177/1087054714530557Journal of Attention DisordersRucklidge et al.
Article
Vitamin–Mineral Treatment of ADHD in Adults: A 1-Year Naturalistic Follow-Up of a Randomized Controlled Trial
Journal of Attention Disorders 1–11 © 2014 SAGE Publications Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1087054714530557 jad.sagepub.com
Julia J. Rucklidge1, Chris M. Frampton2, Brigette Gorman1, and Anna Boggis3
Abstract Objective: Despite widespread use, there is little data investigating the long-term impact of micronutrients on psychiatric disorders. This study investigated the naturalistic outcome 1-year post-baseline of a randomized controlled trials (RCT) that compared micronutrients with placebo in 80 adults with ADHD. Method: All participants were contacted and clinician-rated questionnaires completed. Results: A total of 72 (90%) of the sample participated; although there was significant regression in psychiatric functioning from the end-of-trial on all measures, outcomes remained significantly improved from baseline. Dominant treatment from the end-of-treatment to follow-up was investigated as a mediator of outcome; those staying on the micronutrients performed better than those who switched to medications or discontinued micronutrients. Cost was the most substantial reason why people stopped micronutrient treatment. Conclusion: For the small number of participants who stayed on micronutrients, the benefits conferred through the controlled trial were maintained. The results are limited by small sample, lack of blinding, expectation, and reliance on self-report of symptoms. (J. of Att. Dis. XXXX; XX(X) XX-XX) Keywords ADHD, micronutrients, minerals, vitamins Trials examining efficacy of treatments are typically shortterm, providing us with good information about whether a treatment works, but very little information on whether treatment effects are sustained over time. The short-term efficacy of micronutrients (vitamins and minerals) has now been demonstrated in a number of randomized controlled trials (RCTs) across a variety of disorders and psychological states (e.g., ADHD, autism and stress; Adams et al., 2011; Kennedy et al., 2008; Rucklidge, Frampton, Gorman, & Boggis, 2014); however, no study has followed participants naturalistically who have been in a controlled experiment testing the efficacy of micronutrients for the treatment of mental illness to determine what happens to them after the controlled trial is completed. ADHD, a chronic condition consisting of difficulties with attention, hyperactivity, and/or impulsivity, has been well documented to persist into adulthood with continued impairment across a range of areas, including psychiatric, educational, financial, and risk-taking behaviors (Biederman et al., 2006; Murphy & Barkley, 1996). Short-term treatments for ADHD tend to center on medications; however, the long-term effectiveness of these treatments has been challenged (Advokat, 2010; Molina et al., 2009; Parker, Wales, Chalhoub, & Harpin, 2013), with concerns expressed about the adverse effects (Thomas, Mitchell, & Batstra,
2013) and reduced responses due to comorbidities (Torgersen, Gjervan, & Rasmussen, 2008). As such, many individuals seek out alternative treatments for their symptoms (Baumgaertel, 1999; Nigg, 2011). In our double-blind RCT (Rucklidge et al., 2014), we demonstrated that over 8 weeks, micronutrient treatment induced statistically robust improvements in several indices, from ADHD symptoms to Global Assessment of Functioning (GAF), compared with placebo. This naturalistic follow-up aimed to (a) provide a naturalistic long-term follow-up of adults with ADHD treated with micronutrients to document how they were functioning both as compared with baseline and with the end of the trial and (b) determine whether dominant treatment status from the end of openlabel (OL) to follow-up influenced psychiatric functioning. Measures included standardized clinician-rated scales
1
University of Canterbury, Christchurch, New Zealand University of Otago, Christchurch, New Zealand 3 Canterbury District Health Board, Christchurch, New Zealand 2
Corresponding Author: Julia J. Rucklidge, Department of Psychology, University of Canterbury, Private Bag 4800, Christchurch, 8140, New Zealand. Email: [email protected]
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
2
Journal of Attention Disorders
capturing attention, hyperactivity/impulsivity, mood, quality of life, and overall psychiatric functioning.
Method The study, including the follow-up, received ethical approval from the Human Ethics Committee, University of Canterbury and the Upper South A Regional Ethics Committee. Written informed consent was obtained from all participants. The trial was prospectively registered: www.anzctr.org.au:ACTRN12609000308291
Participants and Entry Criteria Participants were recruited in Canterbury, New Zealand, from May 2009 to May 2012 via referrals from public services, private clinicians, and self-referrals based on advertisements. From 136 referrals, 80 adults (age ≥ 16 years) with ADHD were assigned in a 1:1 ratio to 8 weeks of treatment with either micronutrients or placebo (Figure 1). After 8 weeks, all participants were offered the micronutrients for a further 8 weeks as OL. Participants had to meet criteria for ADHD based on the Conners’ Adult ADHD Diagnostic Interview for Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994; CAADID; Epstein, Johnson, & Conners, 2002) or, for those under 18 years (n = 7), the Kiddie-Schedule for Affective Disorders and Schizophrenia (K-SADS-PL; Kaufman et al., 1997). Only individuals not taking psychiatric medications (medication-free for at least 4 weeks) were considered. See Rucklidge et al. (2014) for other exclusions and study details. Comorbidity was assessed using the Structured Clinical Interview for DSM (4th ed., text rev.; DSM-IV-TR; APA, 2000) Axis I Disorders, Research Version (SCID-I; First, Spitzer, Gibbon, & Williams, 2002). Clinical interviews were conducted by clinical psychologists or senior clinical psychology graduate students. Once eligibility was confirmed and baseline assessment completed, participants were allocated to either the micronutrients (EMPowerplus [EMP]—a formula consisting of 14 vitamins, 16 minerals, 3 antioxidants, and 3 amino acids) or placebo. The ingredients of EMP are available in Rucklidge et al. (2014). The placebo consisted of Fiber Acacia Gum, Maltodextrin, Cocoa Powder, and Riboflavin Powder. Participants titrated their dose over a week from 5 capsules per day to 15 capsules per day, in 3 doses of 5 capsules, taken with food and water. Safety of EMP is now well established and reported elsewhere (Rucklidge et al., 2014; Simpson et al., 2011).
Psychiatric Assessment All participants were monitored with face-to-face or telephone meetings at screening, baseline, and Weeks 1, 2, 4, 6,
and 8 (end of RCT), 10, 12, 14, 16 (end of OL), and 52 weeks (follow-up). At baseline, 8, 16, and 52 weeks, the following measures were completed: (a) the three DSM-IV subscales (inattentive, hyperactivity/impulsivity, and combined) of the Conners’ Adult ADHD Rating Scale–Clinician form (CAARS-O:SV; Conners, Erhardt, & Sparrow, 1999); (b) The Clinical Global Impressions-Improvement (CGI-I) Scales (Guy, 1976) producing two scores: CGI-I-Global (capturing global change in functioning from baseline, acknowledging that individuals with ADHD typically present with many psychiatric symptoms beyond ADHD), and CGI-I-ADHD (reflecting change in ADHD symptoms); (c) The Montgomery–Åsberg Depression Rating Scale (MADRS; Montgomery & Åsberg, 1979); (d) The GAF (APA, 2000); and (e) The Range of Impaired Functioning Tool (LIFE-RIFT; Leon et al., 2000). Treatment response at 52 weeks was defined in two ways: (a) a final CGI-I-Global Impression of either “much” or “very much” improved as compared with baseline and (b) ≥30% decrease as compared with baseline, which is a standard percentage change in ADHD ratings used in the ADHD literature (Wilens et al., 2005), on the CAARS clinician-rated DSM-IV combined subscale. We also identified “remitters” as those who obtained a T-score of 60 or below at 52 weeks on the clinician CAARS DSM-IV combined subscale.
Other Areas Assessed A brief diet intake questionnaire (modified from Baker et al., 2003) was used to assess dietary patterns. At follow-up, participants were asked whether they were still taking micronutrients and at what dose, taking psychotropic medications, or receiving other forms of treatment such as counseling, omega-3s, or other dietary supplements. This information was used to determine the dominant form of therapy over the period from the end of OL to follow-up. Also, if they had discontinued the treatment, they were presented with a list of items and were asked to rate them from 1 (not at all) to 7 (very much) how much that specific item affected their decision to stop taking the micronutrients post-trial. Areas covered included: experienced no symptom benefit, side effects, could not remember to take them, number of pills to take, the cost, and found another treatment.
Procedures Approximately 1-year post-baseline, all participants were contacted by phone and/or email, irrespective of whether they completed the trial or the OL phase, and the clinicianrated measures were completed either over the phone (n = 55) or in person (n = 17). Seven people could not be contacted and one declined to do the follow-up interview.
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
3
Rucklidge et al.
Assessed for eligibility (n = 136)
Enrollment
Not enrolled (n = 56) ♦ Declined to participate (n = 15) ♦ Did not meet inclusion criteria [e.g. not ADHD] (n= 12) ♦ On medications (n = 18) ♦ Abnormal blood work (e.g. sickle cell anemia) (n = 3) ♦ Moved away (n = 8)
1 Year Follow-up
Open Label
Allocation - RCT
Randomized (n = 80)
Allocated to EMP+ (n = 42) ♦ Completed full treatment (n = 38) ♦ Discontinued intervention (n = 4) Participant choice (n = 1, week 2) Perceived lack of efficacy (n = 2, weeks 4,6) Adverse event (n = 1, week 4)
♦ ♦ ♦
Allocated to placebo (n = 38) ♦ Completed full treatment (n = 36) ♦ Discontinued intervention (n = 2) Participant choice (n = 1, week 2) Perceived lack of efficacy (n=1, week 2)
♦ ♦ ♦
Enrolled in open label phase (n = 33) Completed open-label phase (n = 31) Discontinued intervention (n = 2) Adverse event (n = 2)
Enrolled in open label phase (n = 36) Completed open-label phase (n = 34) Discontinued treatment (n = 2) Adverse event (n = 1) Lost to follow-up (n = 1)
Contacted (n = 36) ♦ Lost to follow-up (n = 1) ♦ Refused (n = 1)
Contacted (n = 36) ♦ lost to follow up (n = 6) Dominant therapy: ♦ Micronutrients (n = 6) ♦ Medications (n = 8) ♦ No treatment (n = 22)
Dominant therapy: ♦ Micronutrients (n = 8) ♦ Medications (n = 9) ♦ No treatment (n = 19)
Figure 1. CONSORT flow diagram for participants through the trial. Note. EMP = EMPowerplus.
Statistical Analyses The changes from baseline to follow-up and the end of OL treatment to follow-up were independently compared using paired-sampled t tests (two-tailed). The sample was divided into three groups based on the dominant treatment through to follow-up and the changes were compared between these
groups using ANCOVA with the pre-change level as a covariate and summarizing the treatment effects as mean differences and 95% confidence intervals. Where the ANCOVA model indicated significant differences among groups these were further explored with pairwise comparisons using Fisher’s least significance difference (LSD) tests. Categorical outcomes were compared between groups
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
4
Journal of Attention Disorders
using chi-square tests. The p values less than .05 were considered statistically significant.
Table 1. Demographic and Clinical Features of Those Participants From the RCT Included and Not Included in the12Month Assessment.
Results
Followed up (72)
Not followed up (8)
M ± SD or n (%)
M ± SD or n (%)
Those Followed Versus Those Lost to Follow-Up Of the 80 participants who entered the trial, 72 (90%) were successfully followed up approximately 1 year from baseline (M = 11.0 months; range = 7-19 months). Figure 1 illustrates the retention of participants at each phase. Table 1 shows the baseline demographics and clinical features of those followed versus not followed up. Those not followed up were more likely to be from a lower socio-economic background (28.3 ± 3.0) than those followed up (45.4 ± 18.0), F(1, 79) = 7.2, p = .009, and tended to be older although this was not statistically significant.
Psychiatric Status at 52 Weeks Across most clinician-rated variables (DSM inattention, DSM hyperactivity/impulsivity, DSM combined, MADRS, GAF, CGI-I-ADHD, and CGI-I-Global), there was a significant regression from the end of OL to 52 weeks, although the participants were still significantly improved on all measures compared with baseline (Table 2). Compared with baseline, 45 (62.5%) were identified as “much” to “very much” improved and 23 (31.9%) showed a decrease of at least 30% based on the clinician-rated ADHD scales. Dietary patterns did not change significantly from 16 to 52 weeks as assessed in a subsample, t(32) = 0.70, ns.
Dominant Therapy From 16 to 52 Weeks All participants were asked whether they had stayed on the micronutrients, switched to medications, discontinued the micronutrients, and had not pursued other treatment options, seen a therapist or switched to other nutrients. Based on responses, participants were classified according to the dominant therapy over the time period from 16 to 52 weeks—three groups were identified as follows: (a) those who stayed on micronutrients, (b) those who switched to medications, and (c) those who stopped all forms of therapy. Only 14 (19%) participants had consistently stayed on the micronutrients, with a daily dose ranging from 4 to 15 pills. Seventeen (24%) had switched to medications; 12 were taking ADHD medications (e.g., atomoxetine, methylphenidate, dexamphetamine), 4 were taking medications for mood (e.g., lamotrigine, fluoxetine), and 1 was taking a medication for anxiety. All of those on medications had been taking them for a minimum of 2 months (16 had switched within 1 month of the OL phase). The remainder (n = 41; 57%) had stopped taking the micronutrients and had not switched to medications and the majority (37/41)
Characteristic
Demographics Age 34.2 ± 12.7 41.2 ± 19.0 Male, n (%) 48 (66.7) 5 (62.5) Estimated IQa 113.0 ± 13.9 105.1 ± 13.1 Incomeb $70,000 24 (33.3) 2 (25) Socio-economic statusc 45.4 ± 18.0 28.3 ± 3.0** Ethnic origin, n (%) New Zealanders of 57 (79.2) 7 (87.5) European descent Mäori 3 (4.2) 0 Other 12 (16.7) 1 (12.5) Married/common-law 22 (30.6) 1 (12.5) Clinical, n (%) ADHD type Inattentive 26 (36.1) 3 (37.5) Hyperactive/impulsive 5 (6.9) 1 (12.5) Combined 41 (56.9) 4 (50) Mood disorder (Major Depressive Disorder, dysthymia, or Bipolar Disorder) Current 17 (23.6) 1 (12.5) Past 43 (59.7) 4 (50) At least one anxiety disorder Current 30 (41.7) 3 (37.5) Past 54 (75) 5 (62.5) Alcohol/substance abuse or dependence Current 11 (15.3) 0 Past 23 (31.9) 2 (25) Learning disabilityd 12 (16.7) 3 (37.5) Any co-occurring disorder Current 41 (56.9) 4 (50) Past 54 (75) 5 (62.5) Clinician CAARS (T-scores) DSM inattention 72.2 ± 8.3 71.9 ± 7.5 DSM hyperactivity/ 65.0 ± 12.2 71.3 ± 9.0 impulsivity DSM combined 71.3 ± 9.0 71.5 ± 8.1 MADRS 15.9 ± 6.8 14.2 ± 9.3 Note. RCT = randomized controlled trial; CAARS = Conners’ Adult ADHD Rating Scale; DSM = Diagnostic and Statistical Manual; MADRS = Montgomery-Åsberg Depression Rating Scale. a Assessed using Block Design and Vocabulary subtests of the WAIS-III (Wechsler, 1997). b In New Zealand dollars. c Based on the NZSEI-96 which ranks occupations from 10 to 90 (Davis, McLeod, Ransom, & Ongley, 1997). d Defined as having at least one standard score below 85 on either reading or spelling of the WRAT3 (Wilkinson, 1993). **p < .01, significantly different between those included and not included.
were not taking other non-pharmacological treatments. Four of those who had stopped EMP had switched to other
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
5
Rucklidge et al. Table 2. Baseline, 16-, and 52-Week Data on Outcome Measures for Those Who Completed 52-Week Follow-Up. Total sample (n = 72)
Variable Clinician CAARS (T-scores) DSM inattention DSM hyperactivity/impulsivity DSM combined CGI-I-ADHDb CGI-I-Globalb MADRS total GAF Life-rift total
Baseline
16 weeks
52 weeks
M
SEM
M
SEM
M
SEM
72.2 65.0 71.3
1.0 1.4 1.1
15.9 60.1 11.5
0.8 0.9 0.3
60.9 54.1 58.9 2.5 2.5 9.4 68.4 9.8
1.3 1.3 1.2 0.2 0.2 0.9 1.2 0.3
64.1 60.1 64.1 3.0 3.1 12.6 64.1 10.4
1.3 1.4 1.4 0.1 0.2 1.0 1.2 0.4
Comparing baseline with 52 weeks Change from baseline
Confidence interval
Effect sizea
−8.3*** −4.7*** −7.2***
[−5.6, −10.9] [−2.2, −7.2] [−4.6, −9.7]
0.74 0.44 0.67
−3.4*** 4.0** −1.1**
[−1.5, −5.3] [1.5, 6.5] [−0.3, −1.8]
0.42 0.39 0.35
Comparing 16 weeks with 52 weeks Change from 16 weeks
Confidence interval
Effect sizea
3.2* 6.1** 5.2** 0.5** 0.6*** 3.2** −4.3** 0.6
[0.5, 5.9] [3.8, 8.5] [2.7, 7.6] [0.1, 0.9] [0.2, 1.0] [1.1, 5.3] [−7.0, −1.7] [0.2, 1.3]
0.28 0.62 0.50 0.32 0.41 0.36 0.39 0.19
Note. SEM = Standard Error of the Mean; CAARS = Conners’ Adult ADHD Rating Scale; DSM = Diagnostic and Statistical Manual; CGI-I = Clinical Global Impression Improvement; MADRS = Montgomery-Åsberg Depression Rating Scale; GAF = Global Assessment of Functioning. a Cohen’s d (effect size) measured as the mean difference in the change divided by the within-group SD of the change. b Assesses change so not measured at baseline. *p < .05. **p < .01. ***p < .001.
supplements such as fish oils, spirulina, evening primrose, and/or probiotics but were not taking them consistently such that these treatments were not considered as a dominant therapy. Six of the 72 (8%) had had a few sessions of psychotherapy intervention since stopping the trial (under 6 sessions); 3 (18%) from the medication group, and 3 (7%) in the group who had stopped taking micronutrients. There were no between-group differences in the demographic variables, age: F(2, 69) = 1.2, ns; IQ: F(2, 69) = 0.7, ns; socio-economic status (SES): F(2, 69) = 0.7, ns; gender: χ2(72) = 1.5, ns; or ethnicity: χ2(72) = 0.4, ns, indicating that it was not those who were more educated or had higher incomes who chose to stay on the micronutrients. Also, there was no group difference in time to follow-up, F(2, 69) = 0.8, ns. Table 3 shows the comparisons across these three groups for change in the clinician-rated measures from 16 to 52 weeks, illustrating that with the exception of two comparisons, there were significant group differences in changes in psychiatric functioning from the end of OL to 52 weeks. The pairwise analyses showed that for the CGI-I, MADRS, and GAF, those remaining on micronutrients did not show the decline in functioning seen in those who had discontinued or had switched to medications. It appears that those who switched to medications ended the OL phase more depressed and with more ADHD symptoms than those who stayed on EMP; however, those on medication still showed a worsening in MADRS scores at 52 weeks and showed no improvement in ADHD symptoms even after switching to medications. Figures 2 to 6 illustrate the pattern of change over time for these three groups.
Treatment Response Based on Treatment From the End of OL Therapy to Follow-Up Using the CGI-I-Global as an indicator of a responder (“much” to “very much” improved) compared with baseline functioning, a chi-square revealed a significant group difference, χ2(72) = 18.150, p < .001, with 12 (86%) of those staying on the micronutrients being identified as a responder, 6 (35%) of those on medications identified as a responder, and 27 (38%) of those who stopped the micronutrients being identified as a responder. Using a change of ≥30% on the clinician CAARS as an indicator of clinical improvement from baseline, there was a significant group difference, χ2(72) = 9.4, p = .009. Nine (64.3%) of those on EMP, 6 (35.3%) of those on medications, and 8 (20%) of those who stopped were identified as responders. We also determined how many participants now had a score within the normal nonclinical range (remitters) on the CAARS: 9 (64.3%) of those taking EMP, 5 (29.4%) of those taking medications, and 11 (27.5%) of those who stopped, χ2(72) = 6.5, p = .039.
Reasons for Discontinuing the Micronutrients Those who had stopped the micronutrients after the trial ended (n = 58) were asked to rate the extent to which different reasons affected their decision to stop. The highest mean was associated with the cost of taking the pills (4.3, SD = 2.3), followed by the number of pills (3.7, SD = 2.4) and then not experiencing symptoms benefit (3.6, SD = 2.5).
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
6
Journal of Attention Disorders
Table 3. Comparison of Change From 16 to 52 Weeks on Outcome Measures Based on Dominant Treatment at Follow-Up: (a) Stayed on EMP, (b) Switched to Medications, or (c) Stopped All Treatment. End of open-label: 16 weeks
Variable
Follow-up: 52 weeks
Stayed on Switched to Stopped all Stayed on Switched to EMP medications treatments EMP medications (n = 14); A (n = 17); B (n = 41); C (n = 14); A (n = 17); B
M (SEM)
Clinician CAARS (T-scores) DSM 55.9 (2.8) inattention 52.5 (2.9) DSM hyperactivity/ impulsivity DSM combined 55.1 (3.1) CGI-I-ADHD 2.1 (0.4) CGI-I-global 2.1 (0.4) MADRS total 8.9 (2.6) GAF 71.4 (3.2) Life-rift total 9.1 (0.9)
Change from 16 to 52 weeks
Stopped all treatments (n = 41); C
Stayed on EMP (n = 14); A
Switched to medications (n = 17); B
Stopped all treatments (n = 41); C
Difference (confidence interval)a
Difference (confidence interval)a
Difference (confidence interval)a
Fc
Post hoc analyses A, B < C
M (SEM)
M (SEM)
M (SEM)
M (SEM)
M (SEM)
65.5 (2.3)
60.7 (1.7)
55.6 (2.9)
63.8 (1.9)
67.2 (1.8)
−3.0 [−8.1, 2.1]
0.7 [−3.9, 5.3]
6.4 [3.5, 9.4]
6.09**
58.4 (2.8)
52.9 (1.6)
54.2 (3.1)
63.1 (2.7)
61.0 (1.8)
1.2 [−3.7, 6.0]
6.3 [1.8, 10.7]
7.8 [5.0, 10.7]
2.79
63.9 (1.8) 2.9 (0.2) 2.8 (0.2) 11.7 (1.6) 65.4 (1.9) 11.2 (0.6)
58.2 (1.6) 2.4 (0.2) 2.5 (0.2) 8.6 (1.2) 68.7 (1.5) 9.4 (0.4)
56.0 (3.2) 1.8 (0.2) 1.7 (0.2) 6.6 (1.1) 71.0 (2.5) 8.2 (0.8)
65.5 (2.3) 2.8 (0.3) 3.1 (0.3) 14.2 (2.2) 61.4 (2.2) 11.7 (0.8)
66.3 (1.7) 3.5 (0.2) 3.6 (0.2) 13.9 (1.2) 62.8 (1.5) 10.6 (0.6)
−0.6 [−5.6, 4.3] −0.6 [−1.1, −0.1] −0.7 [−1.3, −0.1] −2.6 [−6.3, 1.1] 1.7 [−3.1, 6.4] −1.1 [−2.7, 0.4]
3.5 [−1.0, 8.1] 7.9 [5.0, 10.8] 0.3 [−0.2, 0.7] 1.0 [0.7, 1.3] 0.5 [−0.0, 1.0] 1.1 [0.8, 1.4] 4.0 [0.6, 7.4] 4.9 [2.7, 7.0] −6.1 [−10.4, −1.8] −5.7 [−8.5, −2.9] 1.2 [−0.2, 1.7] 0.9 [0.0, 1.7]
4.92** 14.74*** 14.69*** 6.23** 3.95* 2.90
A
Published by: http://www.sagepublications.com
Additional services and information for Journal of Attention Disorders can be found at: Email Alerts: http://jad.sagepub.com/cgi/alerts Subscriptions: http://jad.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jad.sagepub.com/content/early/2014/05/07/1087054714530557.refs.html
>> OnlineFirst Version of Record - May 7, 2014 What is This?
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
530557
research-article2014
JADXXX10.1177/1087054714530557Journal of Attention DisordersRucklidge et al.
Article
Vitamin–Mineral Treatment of ADHD in Adults: A 1-Year Naturalistic Follow-Up of a Randomized Controlled Trial
Journal of Attention Disorders 1–11 © 2014 SAGE Publications Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1087054714530557 jad.sagepub.com
Julia J. Rucklidge1, Chris M. Frampton2, Brigette Gorman1, and Anna Boggis3
Abstract Objective: Despite widespread use, there is little data investigating the long-term impact of micronutrients on psychiatric disorders. This study investigated the naturalistic outcome 1-year post-baseline of a randomized controlled trials (RCT) that compared micronutrients with placebo in 80 adults with ADHD. Method: All participants were contacted and clinician-rated questionnaires completed. Results: A total of 72 (90%) of the sample participated; although there was significant regression in psychiatric functioning from the end-of-trial on all measures, outcomes remained significantly improved from baseline. Dominant treatment from the end-of-treatment to follow-up was investigated as a mediator of outcome; those staying on the micronutrients performed better than those who switched to medications or discontinued micronutrients. Cost was the most substantial reason why people stopped micronutrient treatment. Conclusion: For the small number of participants who stayed on micronutrients, the benefits conferred through the controlled trial were maintained. The results are limited by small sample, lack of blinding, expectation, and reliance on self-report of symptoms. (J. of Att. Dis. XXXX; XX(X) XX-XX) Keywords ADHD, micronutrients, minerals, vitamins Trials examining efficacy of treatments are typically shortterm, providing us with good information about whether a treatment works, but very little information on whether treatment effects are sustained over time. The short-term efficacy of micronutrients (vitamins and minerals) has now been demonstrated in a number of randomized controlled trials (RCTs) across a variety of disorders and psychological states (e.g., ADHD, autism and stress; Adams et al., 2011; Kennedy et al., 2008; Rucklidge, Frampton, Gorman, & Boggis, 2014); however, no study has followed participants naturalistically who have been in a controlled experiment testing the efficacy of micronutrients for the treatment of mental illness to determine what happens to them after the controlled trial is completed. ADHD, a chronic condition consisting of difficulties with attention, hyperactivity, and/or impulsivity, has been well documented to persist into adulthood with continued impairment across a range of areas, including psychiatric, educational, financial, and risk-taking behaviors (Biederman et al., 2006; Murphy & Barkley, 1996). Short-term treatments for ADHD tend to center on medications; however, the long-term effectiveness of these treatments has been challenged (Advokat, 2010; Molina et al., 2009; Parker, Wales, Chalhoub, & Harpin, 2013), with concerns expressed about the adverse effects (Thomas, Mitchell, & Batstra,
2013) and reduced responses due to comorbidities (Torgersen, Gjervan, & Rasmussen, 2008). As such, many individuals seek out alternative treatments for their symptoms (Baumgaertel, 1999; Nigg, 2011). In our double-blind RCT (Rucklidge et al., 2014), we demonstrated that over 8 weeks, micronutrient treatment induced statistically robust improvements in several indices, from ADHD symptoms to Global Assessment of Functioning (GAF), compared with placebo. This naturalistic follow-up aimed to (a) provide a naturalistic long-term follow-up of adults with ADHD treated with micronutrients to document how they were functioning both as compared with baseline and with the end of the trial and (b) determine whether dominant treatment status from the end of openlabel (OL) to follow-up influenced psychiatric functioning. Measures included standardized clinician-rated scales
1
University of Canterbury, Christchurch, New Zealand University of Otago, Christchurch, New Zealand 3 Canterbury District Health Board, Christchurch, New Zealand 2
Corresponding Author: Julia J. Rucklidge, Department of Psychology, University of Canterbury, Private Bag 4800, Christchurch, 8140, New Zealand. Email: [email protected]
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
2
Journal of Attention Disorders
capturing attention, hyperactivity/impulsivity, mood, quality of life, and overall psychiatric functioning.
Method The study, including the follow-up, received ethical approval from the Human Ethics Committee, University of Canterbury and the Upper South A Regional Ethics Committee. Written informed consent was obtained from all participants. The trial was prospectively registered: www.anzctr.org.au:ACTRN12609000308291
Participants and Entry Criteria Participants were recruited in Canterbury, New Zealand, from May 2009 to May 2012 via referrals from public services, private clinicians, and self-referrals based on advertisements. From 136 referrals, 80 adults (age ≥ 16 years) with ADHD were assigned in a 1:1 ratio to 8 weeks of treatment with either micronutrients or placebo (Figure 1). After 8 weeks, all participants were offered the micronutrients for a further 8 weeks as OL. Participants had to meet criteria for ADHD based on the Conners’ Adult ADHD Diagnostic Interview for Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994; CAADID; Epstein, Johnson, & Conners, 2002) or, for those under 18 years (n = 7), the Kiddie-Schedule for Affective Disorders and Schizophrenia (K-SADS-PL; Kaufman et al., 1997). Only individuals not taking psychiatric medications (medication-free for at least 4 weeks) were considered. See Rucklidge et al. (2014) for other exclusions and study details. Comorbidity was assessed using the Structured Clinical Interview for DSM (4th ed., text rev.; DSM-IV-TR; APA, 2000) Axis I Disorders, Research Version (SCID-I; First, Spitzer, Gibbon, & Williams, 2002). Clinical interviews were conducted by clinical psychologists or senior clinical psychology graduate students. Once eligibility was confirmed and baseline assessment completed, participants were allocated to either the micronutrients (EMPowerplus [EMP]—a formula consisting of 14 vitamins, 16 minerals, 3 antioxidants, and 3 amino acids) or placebo. The ingredients of EMP are available in Rucklidge et al. (2014). The placebo consisted of Fiber Acacia Gum, Maltodextrin, Cocoa Powder, and Riboflavin Powder. Participants titrated their dose over a week from 5 capsules per day to 15 capsules per day, in 3 doses of 5 capsules, taken with food and water. Safety of EMP is now well established and reported elsewhere (Rucklidge et al., 2014; Simpson et al., 2011).
Psychiatric Assessment All participants were monitored with face-to-face or telephone meetings at screening, baseline, and Weeks 1, 2, 4, 6,
and 8 (end of RCT), 10, 12, 14, 16 (end of OL), and 52 weeks (follow-up). At baseline, 8, 16, and 52 weeks, the following measures were completed: (a) the three DSM-IV subscales (inattentive, hyperactivity/impulsivity, and combined) of the Conners’ Adult ADHD Rating Scale–Clinician form (CAARS-O:SV; Conners, Erhardt, & Sparrow, 1999); (b) The Clinical Global Impressions-Improvement (CGI-I) Scales (Guy, 1976) producing two scores: CGI-I-Global (capturing global change in functioning from baseline, acknowledging that individuals with ADHD typically present with many psychiatric symptoms beyond ADHD), and CGI-I-ADHD (reflecting change in ADHD symptoms); (c) The Montgomery–Åsberg Depression Rating Scale (MADRS; Montgomery & Åsberg, 1979); (d) The GAF (APA, 2000); and (e) The Range of Impaired Functioning Tool (LIFE-RIFT; Leon et al., 2000). Treatment response at 52 weeks was defined in two ways: (a) a final CGI-I-Global Impression of either “much” or “very much” improved as compared with baseline and (b) ≥30% decrease as compared with baseline, which is a standard percentage change in ADHD ratings used in the ADHD literature (Wilens et al., 2005), on the CAARS clinician-rated DSM-IV combined subscale. We also identified “remitters” as those who obtained a T-score of 60 or below at 52 weeks on the clinician CAARS DSM-IV combined subscale.
Other Areas Assessed A brief diet intake questionnaire (modified from Baker et al., 2003) was used to assess dietary patterns. At follow-up, participants were asked whether they were still taking micronutrients and at what dose, taking psychotropic medications, or receiving other forms of treatment such as counseling, omega-3s, or other dietary supplements. This information was used to determine the dominant form of therapy over the period from the end of OL to follow-up. Also, if they had discontinued the treatment, they were presented with a list of items and were asked to rate them from 1 (not at all) to 7 (very much) how much that specific item affected their decision to stop taking the micronutrients post-trial. Areas covered included: experienced no symptom benefit, side effects, could not remember to take them, number of pills to take, the cost, and found another treatment.
Procedures Approximately 1-year post-baseline, all participants were contacted by phone and/or email, irrespective of whether they completed the trial or the OL phase, and the clinicianrated measures were completed either over the phone (n = 55) or in person (n = 17). Seven people could not be contacted and one declined to do the follow-up interview.
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
3
Rucklidge et al.
Assessed for eligibility (n = 136)
Enrollment
Not enrolled (n = 56) ♦ Declined to participate (n = 15) ♦ Did not meet inclusion criteria [e.g. not ADHD] (n= 12) ♦ On medications (n = 18) ♦ Abnormal blood work (e.g. sickle cell anemia) (n = 3) ♦ Moved away (n = 8)
1 Year Follow-up
Open Label
Allocation - RCT
Randomized (n = 80)
Allocated to EMP+ (n = 42) ♦ Completed full treatment (n = 38) ♦ Discontinued intervention (n = 4) Participant choice (n = 1, week 2) Perceived lack of efficacy (n = 2, weeks 4,6) Adverse event (n = 1, week 4)
♦ ♦ ♦
Allocated to placebo (n = 38) ♦ Completed full treatment (n = 36) ♦ Discontinued intervention (n = 2) Participant choice (n = 1, week 2) Perceived lack of efficacy (n=1, week 2)
♦ ♦ ♦
Enrolled in open label phase (n = 33) Completed open-label phase (n = 31) Discontinued intervention (n = 2) Adverse event (n = 2)
Enrolled in open label phase (n = 36) Completed open-label phase (n = 34) Discontinued treatment (n = 2) Adverse event (n = 1) Lost to follow-up (n = 1)
Contacted (n = 36) ♦ Lost to follow-up (n = 1) ♦ Refused (n = 1)
Contacted (n = 36) ♦ lost to follow up (n = 6) Dominant therapy: ♦ Micronutrients (n = 6) ♦ Medications (n = 8) ♦ No treatment (n = 22)
Dominant therapy: ♦ Micronutrients (n = 8) ♦ Medications (n = 9) ♦ No treatment (n = 19)
Figure 1. CONSORT flow diagram for participants through the trial. Note. EMP = EMPowerplus.
Statistical Analyses The changes from baseline to follow-up and the end of OL treatment to follow-up were independently compared using paired-sampled t tests (two-tailed). The sample was divided into three groups based on the dominant treatment through to follow-up and the changes were compared between these
groups using ANCOVA with the pre-change level as a covariate and summarizing the treatment effects as mean differences and 95% confidence intervals. Where the ANCOVA model indicated significant differences among groups these were further explored with pairwise comparisons using Fisher’s least significance difference (LSD) tests. Categorical outcomes were compared between groups
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
4
Journal of Attention Disorders
using chi-square tests. The p values less than .05 were considered statistically significant.
Table 1. Demographic and Clinical Features of Those Participants From the RCT Included and Not Included in the12Month Assessment.
Results
Followed up (72)
Not followed up (8)
M ± SD or n (%)
M ± SD or n (%)
Those Followed Versus Those Lost to Follow-Up Of the 80 participants who entered the trial, 72 (90%) were successfully followed up approximately 1 year from baseline (M = 11.0 months; range = 7-19 months). Figure 1 illustrates the retention of participants at each phase. Table 1 shows the baseline demographics and clinical features of those followed versus not followed up. Those not followed up were more likely to be from a lower socio-economic background (28.3 ± 3.0) than those followed up (45.4 ± 18.0), F(1, 79) = 7.2, p = .009, and tended to be older although this was not statistically significant.
Psychiatric Status at 52 Weeks Across most clinician-rated variables (DSM inattention, DSM hyperactivity/impulsivity, DSM combined, MADRS, GAF, CGI-I-ADHD, and CGI-I-Global), there was a significant regression from the end of OL to 52 weeks, although the participants were still significantly improved on all measures compared with baseline (Table 2). Compared with baseline, 45 (62.5%) were identified as “much” to “very much” improved and 23 (31.9%) showed a decrease of at least 30% based on the clinician-rated ADHD scales. Dietary patterns did not change significantly from 16 to 52 weeks as assessed in a subsample, t(32) = 0.70, ns.
Dominant Therapy From 16 to 52 Weeks All participants were asked whether they had stayed on the micronutrients, switched to medications, discontinued the micronutrients, and had not pursued other treatment options, seen a therapist or switched to other nutrients. Based on responses, participants were classified according to the dominant therapy over the time period from 16 to 52 weeks—three groups were identified as follows: (a) those who stayed on micronutrients, (b) those who switched to medications, and (c) those who stopped all forms of therapy. Only 14 (19%) participants had consistently stayed on the micronutrients, with a daily dose ranging from 4 to 15 pills. Seventeen (24%) had switched to medications; 12 were taking ADHD medications (e.g., atomoxetine, methylphenidate, dexamphetamine), 4 were taking medications for mood (e.g., lamotrigine, fluoxetine), and 1 was taking a medication for anxiety. All of those on medications had been taking them for a minimum of 2 months (16 had switched within 1 month of the OL phase). The remainder (n = 41; 57%) had stopped taking the micronutrients and had not switched to medications and the majority (37/41)
Characteristic
Demographics Age 34.2 ± 12.7 41.2 ± 19.0 Male, n (%) 48 (66.7) 5 (62.5) Estimated IQa 113.0 ± 13.9 105.1 ± 13.1 Incomeb $70,000 24 (33.3) 2 (25) Socio-economic statusc 45.4 ± 18.0 28.3 ± 3.0** Ethnic origin, n (%) New Zealanders of 57 (79.2) 7 (87.5) European descent Mäori 3 (4.2) 0 Other 12 (16.7) 1 (12.5) Married/common-law 22 (30.6) 1 (12.5) Clinical, n (%) ADHD type Inattentive 26 (36.1) 3 (37.5) Hyperactive/impulsive 5 (6.9) 1 (12.5) Combined 41 (56.9) 4 (50) Mood disorder (Major Depressive Disorder, dysthymia, or Bipolar Disorder) Current 17 (23.6) 1 (12.5) Past 43 (59.7) 4 (50) At least one anxiety disorder Current 30 (41.7) 3 (37.5) Past 54 (75) 5 (62.5) Alcohol/substance abuse or dependence Current 11 (15.3) 0 Past 23 (31.9) 2 (25) Learning disabilityd 12 (16.7) 3 (37.5) Any co-occurring disorder Current 41 (56.9) 4 (50) Past 54 (75) 5 (62.5) Clinician CAARS (T-scores) DSM inattention 72.2 ± 8.3 71.9 ± 7.5 DSM hyperactivity/ 65.0 ± 12.2 71.3 ± 9.0 impulsivity DSM combined 71.3 ± 9.0 71.5 ± 8.1 MADRS 15.9 ± 6.8 14.2 ± 9.3 Note. RCT = randomized controlled trial; CAARS = Conners’ Adult ADHD Rating Scale; DSM = Diagnostic and Statistical Manual; MADRS = Montgomery-Åsberg Depression Rating Scale. a Assessed using Block Design and Vocabulary subtests of the WAIS-III (Wechsler, 1997). b In New Zealand dollars. c Based on the NZSEI-96 which ranks occupations from 10 to 90 (Davis, McLeod, Ransom, & Ongley, 1997). d Defined as having at least one standard score below 85 on either reading or spelling of the WRAT3 (Wilkinson, 1993). **p < .01, significantly different between those included and not included.
were not taking other non-pharmacological treatments. Four of those who had stopped EMP had switched to other
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
5
Rucklidge et al. Table 2. Baseline, 16-, and 52-Week Data on Outcome Measures for Those Who Completed 52-Week Follow-Up. Total sample (n = 72)
Variable Clinician CAARS (T-scores) DSM inattention DSM hyperactivity/impulsivity DSM combined CGI-I-ADHDb CGI-I-Globalb MADRS total GAF Life-rift total
Baseline
16 weeks
52 weeks
M
SEM
M
SEM
M
SEM
72.2 65.0 71.3
1.0 1.4 1.1
15.9 60.1 11.5
0.8 0.9 0.3
60.9 54.1 58.9 2.5 2.5 9.4 68.4 9.8
1.3 1.3 1.2 0.2 0.2 0.9 1.2 0.3
64.1 60.1 64.1 3.0 3.1 12.6 64.1 10.4
1.3 1.4 1.4 0.1 0.2 1.0 1.2 0.4
Comparing baseline with 52 weeks Change from baseline
Confidence interval
Effect sizea
−8.3*** −4.7*** −7.2***
[−5.6, −10.9] [−2.2, −7.2] [−4.6, −9.7]
0.74 0.44 0.67
−3.4*** 4.0** −1.1**
[−1.5, −5.3] [1.5, 6.5] [−0.3, −1.8]
0.42 0.39 0.35
Comparing 16 weeks with 52 weeks Change from 16 weeks
Confidence interval
Effect sizea
3.2* 6.1** 5.2** 0.5** 0.6*** 3.2** −4.3** 0.6
[0.5, 5.9] [3.8, 8.5] [2.7, 7.6] [0.1, 0.9] [0.2, 1.0] [1.1, 5.3] [−7.0, −1.7] [0.2, 1.3]
0.28 0.62 0.50 0.32 0.41 0.36 0.39 0.19
Note. SEM = Standard Error of the Mean; CAARS = Conners’ Adult ADHD Rating Scale; DSM = Diagnostic and Statistical Manual; CGI-I = Clinical Global Impression Improvement; MADRS = Montgomery-Åsberg Depression Rating Scale; GAF = Global Assessment of Functioning. a Cohen’s d (effect size) measured as the mean difference in the change divided by the within-group SD of the change. b Assesses change so not measured at baseline. *p < .05. **p < .01. ***p < .001.
supplements such as fish oils, spirulina, evening primrose, and/or probiotics but were not taking them consistently such that these treatments were not considered as a dominant therapy. Six of the 72 (8%) had had a few sessions of psychotherapy intervention since stopping the trial (under 6 sessions); 3 (18%) from the medication group, and 3 (7%) in the group who had stopped taking micronutrients. There were no between-group differences in the demographic variables, age: F(2, 69) = 1.2, ns; IQ: F(2, 69) = 0.7, ns; socio-economic status (SES): F(2, 69) = 0.7, ns; gender: χ2(72) = 1.5, ns; or ethnicity: χ2(72) = 0.4, ns, indicating that it was not those who were more educated or had higher incomes who chose to stay on the micronutrients. Also, there was no group difference in time to follow-up, F(2, 69) = 0.8, ns. Table 3 shows the comparisons across these three groups for change in the clinician-rated measures from 16 to 52 weeks, illustrating that with the exception of two comparisons, there were significant group differences in changes in psychiatric functioning from the end of OL to 52 weeks. The pairwise analyses showed that for the CGI-I, MADRS, and GAF, those remaining on micronutrients did not show the decline in functioning seen in those who had discontinued or had switched to medications. It appears that those who switched to medications ended the OL phase more depressed and with more ADHD symptoms than those who stayed on EMP; however, those on medication still showed a worsening in MADRS scores at 52 weeks and showed no improvement in ADHD symptoms even after switching to medications. Figures 2 to 6 illustrate the pattern of change over time for these three groups.
Treatment Response Based on Treatment From the End of OL Therapy to Follow-Up Using the CGI-I-Global as an indicator of a responder (“much” to “very much” improved) compared with baseline functioning, a chi-square revealed a significant group difference, χ2(72) = 18.150, p < .001, with 12 (86%) of those staying on the micronutrients being identified as a responder, 6 (35%) of those on medications identified as a responder, and 27 (38%) of those who stopped the micronutrients being identified as a responder. Using a change of ≥30% on the clinician CAARS as an indicator of clinical improvement from baseline, there was a significant group difference, χ2(72) = 9.4, p = .009. Nine (64.3%) of those on EMP, 6 (35.3%) of those on medications, and 8 (20%) of those who stopped were identified as responders. We also determined how many participants now had a score within the normal nonclinical range (remitters) on the CAARS: 9 (64.3%) of those taking EMP, 5 (29.4%) of those taking medications, and 11 (27.5%) of those who stopped, χ2(72) = 6.5, p = .039.
Reasons for Discontinuing the Micronutrients Those who had stopped the micronutrients after the trial ended (n = 58) were asked to rate the extent to which different reasons affected their decision to stop. The highest mean was associated with the cost of taking the pills (4.3, SD = 2.3), followed by the number of pills (3.7, SD = 2.4) and then not experiencing symptoms benefit (3.6, SD = 2.5).
Downloaded from jad.sagepub.com at LAKEHEAD UNIV on October 28, 2014
6
Journal of Attention Disorders
Table 3. Comparison of Change From 16 to 52 Weeks on Outcome Measures Based on Dominant Treatment at Follow-Up: (a) Stayed on EMP, (b) Switched to Medications, or (c) Stopped All Treatment. End of open-label: 16 weeks
Variable
Follow-up: 52 weeks
Stayed on Switched to Stopped all Stayed on Switched to EMP medications treatments EMP medications (n = 14); A (n = 17); B (n = 41); C (n = 14); A (n = 17); B
M (SEM)
Clinician CAARS (T-scores) DSM 55.9 (2.8) inattention 52.5 (2.9) DSM hyperactivity/ impulsivity DSM combined 55.1 (3.1) CGI-I-ADHD 2.1 (0.4) CGI-I-global 2.1 (0.4) MADRS total 8.9 (2.6) GAF 71.4 (3.2) Life-rift total 9.1 (0.9)
Change from 16 to 52 weeks
Stopped all treatments (n = 41); C
Stayed on EMP (n = 14); A
Switched to medications (n = 17); B
Stopped all treatments (n = 41); C
Difference (confidence interval)a
Difference (confidence interval)a
Difference (confidence interval)a
Fc
Post hoc analyses A, B < C
M (SEM)
M (SEM)
M (SEM)
M (SEM)
M (SEM)
65.5 (2.3)
60.7 (1.7)
55.6 (2.9)
63.8 (1.9)
67.2 (1.8)
−3.0 [−8.1, 2.1]
0.7 [−3.9, 5.3]
6.4 [3.5, 9.4]
6.09**
58.4 (2.8)
52.9 (1.6)
54.2 (3.1)
63.1 (2.7)
61.0 (1.8)
1.2 [−3.7, 6.0]
6.3 [1.8, 10.7]
7.8 [5.0, 10.7]
2.79
63.9 (1.8) 2.9 (0.2) 2.8 (0.2) 11.7 (1.6) 65.4 (1.9) 11.2 (0.6)
58.2 (1.6) 2.4 (0.2) 2.5 (0.2) 8.6 (1.2) 68.7 (1.5) 9.4 (0.4)
56.0 (3.2) 1.8 (0.2) 1.7 (0.2) 6.6 (1.1) 71.0 (2.5) 8.2 (0.8)
65.5 (2.3) 2.8 (0.3) 3.1 (0.3) 14.2 (2.2) 61.4 (2.2) 11.7 (0.8)
66.3 (1.7) 3.5 (0.2) 3.6 (0.2) 13.9 (1.2) 62.8 (1.5) 10.6 (0.6)
−0.6 [−5.6, 4.3] −0.6 [−1.1, −0.1] −0.7 [−1.3, −0.1] −2.6 [−6.3, 1.1] 1.7 [−3.1, 6.4] −1.1 [−2.7, 0.4]
3.5 [−1.0, 8.1] 7.9 [5.0, 10.8] 0.3 [−0.2, 0.7] 1.0 [0.7, 1.3] 0.5 [−0.0, 1.0] 1.1 [0.8, 1.4] 4.0 [0.6, 7.4] 4.9 [2.7, 7.0] −6.1 [−10.4, −1.8] −5.7 [−8.5, −2.9] 1.2 [−0.2, 1.7] 0.9 [0.0, 1.7]
4.92** 14.74*** 14.69*** 6.23** 3.95* 2.90
A
