In the first stage of the study, the response to a single 2.5g oral L-tyrosine load in two female adolescents with AN and two healthy peers, while on a low protein, low biogenic amine diet was tested. The second stage of the study involved tyrosine administration at 2.5g twice daily for a 12 week period in participants with AN. Testing occurred at baseline and at weeks one, six and 12. The study had approval from the health service and the University Human Research Ethics Committees. Participants and their parent provided signed informed consent.
Recruitment
Healthy female volunteers aged 12-17 years residing in New South Wales, Australia, were recruited through posters in a tertiary hospital and local community health centers from October to December 2006. Exclusion criteria included known significant medical or psychiatric illness, drug or alcohol abuse in the past six months, use of any amino acid supplement within past three months or current use of psychiatric medications. Healthy participants were screened for eligibility, including completion of psychological testing to exclude psychiatric illness. Height and weight measures were taken to ensure participants were not outside the acceptable body mass index (BMI) range based on BMI-for-Age Percentile Charts (14).
Criteria for participants with AN included 12-17 year old females admitted with AN to a tertiary hospital specialist child and adolescent mental health ward or pediatric ward in Newcastle, New South Wales, Australia. Diagnosis was confirmed by the Eating Disorders Examination (EDE) interview (child version), administered by a trained clinician (15). Exclusion criteria included use of any amino acid supplement within the previous three months, medical instability, concurrent severe medical or neurological illness, Phenylketonuria or drug or alcohol abuse within the previous six months. Participants requiring noradrenergic or combined noradrenergic medications or stimulant medication were excluded. Recruitment of participants with AN occurred from February 2007 to March 2010.
Intervention factors
All participants followed a low protein, low biogenic amine diet for the day prior to and initial day of testing, and fasted from 10pm the day prior to testing, until after the initial collection of study blood and urine (eight hours). Dietary protein intake was strictly limited to ≤7g per main meal and ≤1g per mid-meal on the day prior to testing and to ≤2.5g per main meal and ≤1g per mid-meal on the day of testing. Instructions in how to limit protein and amino acids and exclude foods known to be vasoactive or to influence urinary catecholamines or their metabolites (16, 17) were provided to participants and their parent by an experienced clinical dietitian. Parents supervised and supplied food for the diet for healthy participants. Healthy participants were required to keep a strict food and fluid diary during the special diet period (two days). For participants with AN, the diet was prepared, meals were supervised and a strict food and fluid diary was kept by inpatient pediatric nursing staff.
Tyrosine requirements for an average-weight female adolescent are estimated to be 1750-2050mg/day (18), with no observed side effects for intakes as high as 100-500mg/kg tyrosine in adults (19-23). Supplemental amounts of 100-300mg/kg have shown functional improvements in humans such as improved cognitive function (23-25). The tyrosine dosage deemed to be of clinical benefit for this study was therefore set at 100mg/kg/day expected body weight. Based on the average expected weight for a female adolescent (approximately 50kg) (14), the standard daily dose was set at 5g/day. Due to the short half-life of tyrosine (peak two-three hours post-ingestion) (19-21, 26) tyrosine was delivered as a split dose, at two 2.5g doses per day (morning and evening), administered orally in capsulated form.
Participants in the first stage of this study were required to ingest a single 2.5g tyrosine load after an overnight fast. Participants were reviewed by the hospital pediatrician four hours after tyrosine administration and monitored by nursing staff for eight hours after administering the initial tyrosine load. Participants with AN ingested twice daily 2.5g tyrosine doses for 12 weeks at pre-determined times (approximately 12 hours apart) at least 30 minutes before or one hour after meals to assist with absorption. The tyrosine was taken with a minimum of 100mls of fluid, preferably orange juice, as ascorbic acid is required as a cofactor in noradrenaline synthesis (27). Nursing staff administered supplements during hospital admissions, and parents supervised administration at home. Participants with AN were monitored in hospital by nursing and medical staff for the first four days of tyrosine supplementation. Participants were contacted fortnightly by the researcher to monitor compliance with the supplement regime. Supplements were provided by pharmacy twice throughout the study and unused supplements collected.
As phenylalanine is converted to tyrosine, dietary intake of tyrosine and phenylalanine in participants with AN were estimated using 24-hour recalls collected at four time points to coincide with blood and urine testing (28). To minimise respondent bias, participants were informed that they would be contacted four times throughout the study for dietary recalls and that the timing of recalls would not be disclosed in advance. A trained clinical dietitian administered the recalls using visual aides (plate, cup and ruler) and analysed data using a nutritional analysis program (29). Dietary sources high in protein, amino acids and aspartame (artificial sweetener), known to be high in people with an eating disorder and to influence plasma phenylalanine (30), were emphasised during collections.
Outcome measures
Blood tyrosine level was the main outcome measure. For the first stage of the study, blood samples were taken four hours prior to tyrosine ingestion (fasting), immediately prior to tyrosine administration and at 1, 2, 3, 4, 6 and 8 hours after supplement ingestion. Blood samples for the second stage of the study were taken at baseline (immediately prior to supplement administration) and two hours later at weeks one, six and 12 in participants with AN, allowing for estimation of trough and peak blood tyrosine levels. In those with AN, blood spot samples (whole blood dried on filter paper) were taken at every blood collection for quality monitoring purposes.
For healthy peers, heparinized plasma samples were analyzed by an independent National Association of Testing Authorities (NATA)-accredited laboratory using high performance liquid chromatography (HPLC) with electrochemical detection. Due to resource issues within the laboratory, plasma samples were unable to be analyzed for participants with AN at completion of this study and only blood spot analyses were used. Blood spot samples were analyzed at a NATA-accredited laboratory using electrospray tandem mass spectrometry in dried-blood-spots with underivatised samples (31). Within the laboratory, tyrosine levels in dried blood spots measured by tandem mass spectrometry and those in plasma, taken from the same blood samples and measured by ion-exchange chromatography, correlated well.
For the first stage of the study, urine samples were conducted as timed collections four hours before supplement ingestion (fasting sample), at baseline (supplement ingestion), then at four and eight hours after tyrosine administration. For the second stage, urine samples were collected 4 hours post-supplementation at weeks one, six and 12. Samples were analyzed by an independent NATA-accredited laboratory using HPLC with electrochemical detection adapted from Riggen and Kissinger (32). Urinary tyrosine and catecholamine activity was measured, including tyrosine, dopamine, noradrenaline, adrenaline, homovanillic acid, vanillylmandelic acid and 5-hydroxyindoleacetic acid.
Height and weight measurements were taken by pediatric nursing staff using the same calibrated digital scales and a stadiometer. Weight was measured to the nearest 0.1kg and height to the nearest 0.5cm, without shoes or heavy clothing. In participants with AN, height and weight were taken at baseline, six weeks and 12 weeks. Percent expected body weight calculations were based on the Centre for Disease Control Body Mass Index (BMI) Percentile Charts (14), with 100% weight for height being the 50th percentile BMI.
Participants and their parent or carer each completed a brief purpose-developed questionnaire in pencil and paper format regarding perceived study acceptability. To assess the possible effects of tyrosine administration and catecholamine synthesis in the brain, a range of psychological tests were administered to participants with AN. To measure eating disorders psychopathology, the EDE was administered at baseline and week 12 (15). Anxiety was measured at baseline and weeks one, six and 12 using the State-Trait Anxiety Inventory Form Y-1 (33). Depressive and obsessive compulsive symptoms were measured at baseline and weeks six and 12 using the Children’s Depression Inventory (34) and the Children’s Obsessive Compulsive Inventory (35) respectively. The Strengths and Difficulties Questionnaire was administered at baseline and weeks six and 12 as a general mental health measure (36).
A standardised battery of cognitive function tests were administered by experienced clinical psychologists at baseline and week 12. These were: Rey Complex Figure Test (Meyer and Meyer) initially copy and 30 minute recall (37), Verbal Fluency (FAS) Condition One (Baron) (38), Tower Task (Krikorian) (39), Stroop Color-Word task (Golden) (40), Verbal Paired Associate Learning (Wechsler Memory Scale, Revised) (41), Digit Symbol-Coding (Wechsler Intelligence Scale for Children, Fourth Edition) (42), Visual Learning (Wide Range Assessment of Memory and Learning) (43), Matching (Wide Range Assessment of Visual-Motor Abilities) (44), Trail Making (Reitan) (45) and Design Fluency (46). The Wide Range Achievement Test-3 Reading Test (47) was used as a measure of executive function. An experienced neuropsychologist converted participant test scores to normative data and applied ability ranges. In order to minimise systematic error (e.g. practice effects) or measurement error (e.g. test unreliability) in psychological tests, reliable change index (RCI) methods were used to examine change in psychological tests (48).
Participants and characteristics
Healthy peers were both female, aged 14 years and had a percentage expected body weight of 100-110%. Participant 1 with AN was 15 years of age, weighed 46kg at baseline and had been amenorrheic for three months. Participant 2 was 12 years of age, weighed 37kg at baseline and remained pre-menarchal. Both AN participants had a relatively short duration of illness (three months) and similar expected body weights at baseline (80-82%). Neither participant with AN had received treatment prior to the recent hospital admission and had no comorbid medical or psychiatric diagnoses. Both participants with AN were taking multivitamin, thiamine and phosphate supplements at baseline, though no other medications. Both participants were refed on the pediatric ward for the initial 2 weeks of the study, then attended CAMHS for weekly family therapy interventions. Participant 1 with AN had a secondary diagnosis of obsessive compulsive disorder at completion of the study. Both participants had consumed Olanzapine (an antipsychotic) at times during the study and Participant 1 also consumed Lorazepam (a benzodiazepine) towards the end of the study. Participant 1 reported several self-induced vomiting episodes before and during the study, though denied other purgative behavior. Binge eating and excessive exercise behavior were denied by both participants.
Dosage and dietary intakes
The initial 2.5g tyrosine load equated to a dosage of 45-47mg/kg actual body weight in the healthy participants. The dosage for Participant 1 with AN was 54mg/kg and for Participant 2 68mg/kg. Dietary intakes for all participants in the first stage remained within required protein restrictions for the day before and day of testing. With the ongoing 5g daily tyrosine dose for the participants with AN, Participant 1 received a slightly smaller average dose of tyrosine (110mg/kg/day) and had lower average dietary intakes (protein 55g, tyrosine 2.2g and phenylalanine 2.4g), though fewer unused supplements (<1%). The average tyrosine dose for Participant 2 was 121mg/kg/day with average daily dietary intakes of 94g protein, 3.8g tyrosine and 4.1g phenylalanine.
Outcome for blood, urine and expected body weight
On day one, baseline (time 0) plasma tyrosine concentrations were similar for all four participants (48-60µmol/L) (Figure 1). Peak tyrosine levels were observed at approximately two-three hours (132-240µmol/L) and approached baseline levels by eight hours (62-100µmol/L). Percentage change in plasma tyrosine (between trough and peak levels) ranged from 152-194% in healthy peers and 164-300% in participants with AN. Participant 1 with AN had a notably higher peak tyrosine response, nearly double that of other participants. Urinary tyrosine concentrations were too low for reliable quantification by the method used for all participants at all time points. There were no consistent changes over time in urinary catecholamines or metabolites, and no values exceeded the normal reference range.
By the end of stage two, percent expected body weight remained essentially unchanged (80%) in Participant 1, while Participant 2 was relatively weight-restored (96%). Table 1 details the blood tyrosine response to an oral tyrosine load in participants over 12 weeks. There was a chronic rise in blood tyrosine during the course of the study, and an acute rise two hours after supplement administration. In Participant 2, the morning trough level normalized by week 12. In Participant 1 this did not occur, though the magnitude of the acute rise diminished.
Psychological tests and side effects
Some improvements in participant psychological tests were evident, with no notable declines. While both participants remained within the clinically significant range for eating disorders psychopathology over the 12 weeks (Table 2), there was a statistically significant improvement in Eating Restraint (RCI=-2.50) and the Global Score (RCI=-1.65) for Participant 2. Non-significant improvements in both Weight and Shape Concern were found in both participants, along with a non-significant increase in Eating Concern. Clinically significant improvements in Trait Anxiety (Table 3) and depression scores for Anhedonia were found for both participants, and in Participant 2 for the depressive symptoms Total Score, Negative Mood, Ineffectiveness and Negative Self-Esteem (Table 4). For Participant 1, there were clinically significant increases in Interpersonal Problems and Ineffectiveness. For obsessive compulsive symptomology, improvements were evident in most symptom subscales and Total Impairment, aside from Obsessions Severity which increased in Participant 1. Total Impairment remained in the clinically significant range, aside from moving temporarily into the normal range at week six in Participant 2.