Introduction

Uric acid, the end product of purine metabolism, is actively reabsorbed by the kidney, suggesting a physiological role [9]. In humans, elevated UA may lead to systemic complications, including arthritic gout as solubility limit is exceeded, and is associated with cardiovascular, metabolic, and renal complications [37]. Uric acid and purines, including adenosine and adenosine triphosphate (ATP), have been implicated in the modulation of central nervous system functions such as convulsive threshold, memory, cognition, sleep, activity, appetite, mood, social interaction, drive, impulsivity, and intelligence [4, 33]. Lower serum UA has been reported in neurodegenerative disorders, including Parkinson’s disease [12], Alzheimer’s disease [17], and amyotrophic lateral sclerosis [1], suggesting neuroprotective effects. Previous studies reported elevated plasma UA levels among bipolar disorder patients [5] and lower levels among major depression and anxiety disorders patients, [8] with a tendency to normalize following treatment [6]. Umeki [41] reported that 11 of a sample of 27 chronic AN patients demonstrated abnormally high UA at admission [41]. Gupta and Kavanaugh-Danelon [20] reported 9% of a sample of consecutive ED patients presented with abnormally elevated UA at initial evaluation suggesting this may be related to increased ATP degradation due to strenuous physical activity or reduced renal UA clearance due to starvation [20]. A previous population-based retrospective cohort study reported a threefold higher frequency of renal calculi, with 16% coded as uric acid urolithiasis among AN, compared with 5% among control female participants [15]. Otherwise rare ammonium urate stones, have been repeatedly reported in AN, and associated with laxative abuse [16, 18]. Several cases were reported of hyperuricemia with chronic tophaceous gout complicating the course of long standing AN, and variably attributed to either pseudo-Barter syndrome secondary to laxative abuse [2], chronic diuretic ingestion [21, 30] or dehydration, hypokalemia, and chronic renal damage secondary to recurrent self-induced vomiting [25]. A previous commentary suggested altered UA in AN as an idea worth researching, based on its proposed evolutionary and pathophysiological roles [38]. Salivary UA demonstrates a linear correlation with serum UA and has been suggested as a non-invasive measure to replace blood draw in the clinical setting [39]. We compared salivary UA levels between a cohort of adolescents with ED presenting for treatment at our center and matched healthy controls.

Methods

Participants

All consenting female ED patients referred to the ambulatory clinic, day care facility, or inpatient unit of the Herman Dana Child Psychiatry Center, underwent structured clinical assessment using the Kiddie Schedule for Affective Disorders and Schizophrenia, or if over 18, the Structured Clinical Diagnostic Interview, administered by a trained psychiatrist to diagnose/exclude axis I psychiatric disorders, as well as a comprehensive medical evaluation. Participants meeting DSM-V criteria for Anorexia Nervosa (AN), Bulimia Nervosa (BN), or Eating Disorder Not Otherwise Specified (ED-NOS) were included if meeting inclusion and exclusion criteria. Co-morbid personality, anxiety, and unipolar depressive disorders were allowed. Psychotic and bipolar disorders or chronic medical or neurological illness were excluded. Psychiatrically and medically healthy age matched female subjects participated as control group. The study protocol was conducted in accordance with the Declaration of Helsinki, approved by the Institutional review board ethics committee, and subjects and their parents signed informed consent forms.

Salivary uric acid

Morning salivary assays were sampled at baseline and compared with healthy age matched controls. Saliva was sampled within half an hour after awakening from participants using SalivaBio oral swabs according to manufacturer’s instructions (Salimetrics Carlsbad CA). Participants were instructed to avoid food, liquids, and teeth brushing, and rinse their mouth with water 10 min prior to sampling. Samples were stored at 4 °C for 24 h, centrifuged and stored at − 80 °C until processing. Salivary UA concentration was assayed with duplicates applying a commercially available (high sensitivity) salivary UA enzyme immunoassay kit (Salimetrics Carlsbad CA). The analytical sensitivity was set to 0.07 mg/dl. Inter and average intra-assay co-efficient of variations were 10% and 5%, respectively.

Analyses

T test and none parametric Mann–Whitney test were performed using IBM SPSS Statistics 25.

Results

Salivary UA levels were compared between 39 female ED patients (34 Anorexia Nervosa, 5 Bulimia Nervosa) and 44 healthy female controls. The two groups did not differ in age (ED mean age 17 ± 3.6 years; CON mean age 18.7 ± 4.9 years; t = 1.97 df = 81; p = 0.06). BMI was significantly lower among ED (ED mean BMI 15.3 ± 3.3; CON mean BMI 20.9 ± 2.8; t = 8.43 df = 81; p = .000). BMI was not correlated with UA levels (r = − 0.1; p = 0.1). Salivary UA levels were higher among ED compared with healthy controls ED n = 39 mean 3.9 ± 1.9 mg/dl (range min: 1.5 mg/dl–max 8.5 mg/dl); CON n = 44; mean 2.9 ± 1.2 mg/dl (range: min 1.5 mg/dl–max 4.8 mg/dl); t = − 3.13 df = 81; p = 0.003 (Fig. 1). Creatinine values among ED subjects were all within normal limits (creatinine mean ± SD 62.9 ± 7.19 micromol/L; range: min 45 micromol/L–max 78 micromol/L). Estimated glomerular filtration rate values (eGFR mean ± SD 117.68 + 17.49 ml/min/1.73 m2; range: min 91 ml/min/1.73 m2–max 176 ml/min/1.73 m2) did not show a significant correlation with salivary UA values (r = 0.06; p = 0.75). The ED sample included 30 AN-restrictive type (AN-R), 4 AN-purging type (AN-P), and 5 BN subjects. To examine if purging practices affected salivary UA levels, we combined AN-P and BN subjects. Applying the Mann–Whitney non-parametric test, salivary UA levels were higher among the combined group of AN-P and BN subjects compared with AN-R subjects, but not reaching significance (AN-P + BN n = 9 mean 5 ± 2.5 mg/dl, AN-R n = 30 3.5 ± 1.5 mg/dl, U = 82 p = 0.077). Uric acid remained significantly elevated among AN-R compared with controls (AN-R n = 30 mean 3.5 ± 1.5 CON n = 44 2.9 ± 1.2 t = − 2.3 df = 72; p = 0.026).

Fig. 1
figure 1

Mean salivary uric acid concentration (mg/dl) was significantly elevated among eating disorders (ED) subjects compared with controls. Error bars represent standard deviation of the means

Full size image

In our sample, only 2 of 44 controls, but 35 of 39 ED adolescents reported amenorrhea. Four of 44 controls and 2 of ED subjects reported current oral contraceptive use.

Discussion

We found elevated salivary UA among a sample of adolescent ED patients compared with age matched controls. Previous case reports suggested hyperuricemia resulting in chronic tophaceous gout or renal calculi, may be a complication secondary to purging behaviors, including chronic diuretic or laxative abuse, or renal damage secondary to recurrent vomiting, complicating the course of long-standing AN. Patients admitted to treatment at our center undergo a comprehensive general medical examination and none had gout, arthritis, or renal failure. We found a trend for higher salivary UA levels among ED with purging behaviors despite their small number (n = 9) in our sample, suggesting these may contribute to elevate UA among the ED group. However, salivary UA remained significantly higher among the larger AN-R group (n = 30) compared with control values, suggesting additional factors underlying elevated UA among AN-R.

The potential impact of diuretic and laxative abuse on UA levels could not be assessed in the current sample of young AN subjects, all reporting none use. Older, more chronic samples would be required to assess the potential additional impact of such use. While eGFR did not demonstrate a significant correlation with salivary UA in our sample of adolescent ED patients all exhibiting creatinine levels within the normal range, kidney studies, including creatinine clearance and 24 h urinary levels of UA would be required to further evaluate the relationship of UA with renal clearance. None of the subjects reported clinical episodes of renal stones and renal ultrasound was not available for any of the study subjects, the possibility of silent renal stones, therefore cannot be ruled out.

During reproductive years, women generally manifest lower UA than men, in part due to a facilitative uricosuric effect of both estrogen and progesterone on renal UA clearance [3], and gout mostly afflicts men and is uncommon before menopause among women [24]. Adolescents and adults with AN exhibit lower serum levels of estradiol, which may manifest with a spectrum of menstrual dysfunction up to frank functional hypothalamic oligo-amenorrhea, as luteinizing hormone reverts to prepubertal secretion patterns [28]. Among healthy menstruating women, endogenous secretion of both estrogen and progesterone shows a dose-dependent correlation with UA level reductions during the menstrual cycle [29]. In our sample, only 2 of 44 controls, but 35 of 39 adolescents with ED reported amenorrhea, indirectly suggestive of a possible contribution of lower gonadotropin levels in elevating UA, among ED in our sample. Hormone therapy has been shown to reduce UA levels [24] and the incidence of gout [10], among postmenopausal women with combined estrogen–progestogen treatment, though not with estrogen therapy alone. Only 6 of 83 subjects in our entire sample reported oral contraceptive use. Further prospective studies correlating UA with gonadotropin levels among AN during active and remission phases, as well as UA changes in response to oral contraceptives, are required to shed light on such contribution. Differing dietary compositions have been associated with altered UA levels [36] and could potentially account for the UA differences observed. However, a high purine diet that would account for increased UA levels [13] appears less likely given dietary restriction habits among an ED sample composed predominantly of restrictive AN subjects. A correlation of BMI with UA has been previously reported with positive correlation among high BMI subjects, perhaps as part of a metabolic syndrome [26]. In our sample, no correlation was noted between BMI and salivary UA among the entire sample, ED, or controls. Serum UA may also increase during late stage extreme starvation, when protein is degraded following the depletion of fat, serving as a proposed alarm signal promoting foraging behavior [22], but this is less likely to represent a predominant mechanism in the BMI range observed in our ED sample. Salivary UA demonstrates a stable and strong linear correlation with serum UA (r = 0.69) [35] and has been suggested as a non-invasive alternative in various clinical settings [39, 40]. Salivary UA range in the control group (range min 1.5 mg/dl–max 4.8 mg/dl) is within the normal reference interval reported for a healthy sample of physically active males aged 18–20 years (1.18 mg/dl–5.38 mg/dl) [31], whereas the upper value in the clinical sample (range: min 1.5 mg/dl–max 8.5 mg/dl) is outside this range, with six AN subjects with salivary UA levels above 5.38 mg/dl. However, female values are generally somewhat lower than male values [3] and larger representative studies among healthy age and gender stratified populations are required to establish clinically accepted normative value ranges for salivary UA.

Uric acid is an end product of the degradation pathway of adenosine triphosphate (ATP), a ubiquitous coenzyme for energy transfer synthesized in the mitochondria during oxidative phosphorylation. ATP is degraded into adenosine that is further degraded into inosine and urate by adenosine deaminase and xanthine oxidase (XO), respectively [14]. UA circulates as monosodium urate in physiological PH, constituting the major antioxidant in human plasma [27]. Evolutionary mutations dated to the Miocene epoch rendering urate oxidase inactive among humans and higher primates, prevent its degradation to allantoin, resulting in a higher serum UA, proposed to increase blood pressure and antioxidant capacity in early hominoids, and confer historical evolutionary advantage [23]. However, more recently, increasing access to high purine, including high meat and fructose diet in developed countries, results in excessive UA levels, near the limit of its solubility, with consequent risk of gout and urolithiasis [13]. High serum uric acid levels have been further associated with an increased risk of hypertension, obesity, metabolic syndrome, insulin resistance, cardiovascular complications, and chronic kidney disease [23]. Peripherally generated UA may not readily cross the blood–brain barrier, as suggested by the tenfold gradient from the blood to cerebrospinal fluid (CSF) [34]. Accumulating evidence suggests that UA may serve beneficial roles as an endogenous CNS antioxidant and neuroprotectant [14], in line with observed association of lower plasma UA levels with several neurodegenerative disorders as noted above. The purinergic system—which comprises various receptor subtypes and ectoenzymes that degrade ATP into adenosine and inosine—is present in numerous brain areas, including the cerebral cortex, hypothalamus, basal ganglia, hippocampus, and other limbic areas [11]. Adenosine and ATP are stored in the cytoplasm of nerve terminals, acting as purinergic neurotransmitters and co-transmitters in neurons and glia and elevated UA levels accelerate purinergic transformation and decrease adenosinergic neurotransmission [11]. Lesch–Nyhan syndrome (LNS) is a rare inherited X-linked recessive disorder, resulting from mutations in the HPRT1 gene encoding hypoxanthine–guanine phosphoribosyltransferase (HPRT), causing enzymatic deficiency and accumulation of extreme UA levels with renal and neurological involvement and gouty arthritis [32]. While associated intellectual disability restricts discernment of specific behavioral symptoms, affected individuals typically exhibit a spectrum of compulsive self-injurious behaviors ranging from severe physical harm, to self-induced vomiting, spitting, and coprolalia, to more nuanced manifestations, including rejecting rewards, engaging in self-defeating practices, and provoking anger from caregivers, when affection is desired [19]. Early enzymatic deficiency interferes with brain development, and while treatment with the xanthine oxidase inhibitor allopurinol reduces UA levels, it does not affect the self-injurious behavioral manifestations [7]. Furthermore, LNS variant phenotypes that possess over 1% enzymatic HPRT activity do not demonstrate self-injurious behaviors, making it less likely these are a result of central action of extreme UA levels [7].

To conclude, our study is the first to report elevated levels of salivary UA, among adolescents diagnosed with ED. Further prospective studies extending to the premorbid and remission phases are required to shed light on a putative role for UA as a state or trait prognostic marker and studies comparing ED subtypes and correlating with complicating clinical course measures and behavioral and cognitive aspects, including structured quantitative measures such as physical activity levels, personality traits, and impulsivity/anger indices, are required to further elucidate its putative role in ED pathophysiology.