A different perspective into clinical symptoms in CPT I deficiency

Abstract

Carnitine palmitoyltransferase I (CPT I) deficiency is an autosomal recessive disorder causing long-chain fatty acid oxidation defect, characterized by metabolic decompensation episodes accompanied by hypoketotic hypoglycemia, hepatomegaly, seizures, renal tubular acidosis, and hyperammonemia. The aim of this study was to investigate the neurological symptoms in CPT I deficiency and different outcomes with respect to predisposing factors for sequela and to draw attention to the neurological impairment that may develop during the course of the disease. The retrospective study reviewed clinical characteristics of 14 patients. Mean follow-up period was 10.3 ± 4.7 (range: 8 months-18.6 years; median: 10 years) years. Three patients were diagnosed with newborn screening. In the symptomatic group (n = 12) most common presenting symptoms were psychomotor retardation (n = 6), seizures (n = 5), encephalopathy (n = 5), dystonia (n = 1), Reye-like syndrome (n = 5), muscle weakness (n = 3), and autism (n = 1). Neurologic findings detected in the follow-up period included speech disorder (n = 9), abnormal cranial MRI findings (n = 5), neuropathy (n = 1), and attention deficit hyperactivity disorder (n = 1). Speech disorders collectively included delayed expressive language development, speech articulation disorder, speech delay, stuttering, and specific speech difficulties. After starting treatment for CPT I deficiency, speech disorders improved in 3 patients. Our findings confirmed that the clinical manifestations of CPT I deficiency is wider than previously thought, causing specific neurologic dysfunction, mainly speech disorders at a large scale, that were unexpected in a fatty acid oxidation disorder. We suggest that early diagnosis and treatment is the key factor to prevent neurologic sequelae while an extensive neurological evaluation is essential in patients with CPT I deficiency both at the time of diagnosis and during the follow-up period.

1. Introduction

Carnitine palmitoyltransferase I (CPT I) deficiency (OMIM 255120) is an autosomal recessive disorder of the carnitine shuttle causing a defect in long-chain fatty acid oxidation, characterized by episodes of metabolic decompensation accompanied by hypoketotic hypoglycemia, hepatomegaly, seizures, renal tubular acidosis, and hyperammonemia, which can present at initial admission or during the follow-up period [1]. It is an ultra-rare disease with an estimated prevalence of 1:500,000 to 1:1,000,000 newborn [2]. Disease prevalence is high among the Inuit and Hutterite communities, with recuring mutations p.P479 and p.G710E [3,4].

The CPT I enzyme is a critical enzyme of the carnitine cycle, located in the outer mitochondrial membrane, ensuring the transport of long-chain fatty acids in acyl-CoA esters to carnitine. The resulting acylcarnitines are transported to the carnitine palmitoyltransferase II (CPT II) located in the inner mitochondrial membrane through the enzyme carnitine-acylcarnitine translocase (CACT). The CPT II enzyme separates long-chain fatty acids from acylcarnitines, allowing them to undergo fatty acid oxidation [5].

Carnitine palmitoyltransferase Ia (CPT Ia or liver CPT I [L-CPT I]) is the only isoform expressed in human fibroblasts and is found in the liver and kidney [5]. CPT Ib or muscle CPT I (M-CPT I) is expressed in skeletal muscle and adipose tissue, and although both CPT Ia and CPT Ib are expressed in cardiac muscle, CPT Ib is the dominant isoform [6]. The third CPT isoform, CPT Ic, which has been recently discovered, is expressed in the brain, particularly in the regions closely related to food intake such as the hypothalamus, amygdala, and hippocampus [7,8].

Initial clinical manifestations of L-CPT I deficiency usually appear in infancy and the most common presenting symptom is Reye-like syndrome with hypoketotic hypoglycemia. Catabolic episodes, which are often triggered by conditions such as fasting and intervening infection, may cause acute liver failure and hepatic encephalopathy. Metabolic acidosis, elevated transaminases, and hyperammonemia may be observed at admission. Additionally, cases of L-CPT I deficiency accompanied by renal tubular acidosis have also been reported [9,10]. On the other hand, cardiomyopathy and myopathy have not generally been associated with CPT I deficiency [5].

Due to the function of the CPT I enzyme in the carnitine cycle, while free carnitine level is low in other mitochondrial fatty acid oxidation disorders (FAODs), it is high in CPT I deficiency, and it is the only FAOD characterized by increased total and free carnitine levels [11,12]. CPT I deficiency is often diagnosed based on the elevated ratio of free carnitine (C0) to the sum of long-chain acylcarnitines palmitoyl carnitine (C16) and stearyl carnitine (C18) [C0/(C16 + C18)] [13,14]. In patients with CPT I deficiency, disruption of fatty acid β-oxidation, which is the normal process of breaking down a long-chain acyl-CoA molecule to acetyl-CoA molecules, leads to moderately increased urinary excretion of dicarboxylic acids, the products of omega-oxidation [9].

The primary goal of the treatment of CPT I deficiency is to provide adequate glucose support to suppress adipose tissue lipolysis [15]. In addition, an optimal serum glucose level should be provided. In the long-term follow-up, prevention of long-term fasting with frequent nutrition and restriction of long-chain fatty acids in the diet constitute the basis of nutritional therapy. Adequate calories are provided by a low-fat, high-carbohydrate diet supplemented with medium-chain fatty acids (MCTs) [16]Supplementation of essential fatty acids is needed to prevent deficiency [17].

To our knowledge, among the deficiency of three isoforms, only L-CPT I deficiency has been reported to date, though the spectrum of clinical findings of CPT I deficiency appears to be wider than previously thought [1,5,13]. The aim of this study was to investigate the neurological symptoms in L-CPT I deficiency and to draw attention to the neurological symptoms that may develop during the course of the disease.

2. Materials and methods

The demographic and clinical characteristics of 14 patients that were diagnosed and followed with CPT I deficiency based on acylcarnitine profile by tandem mass spectrometry and/or molecular analysis in Istanbul University Medical Faculty, Division of Pediatric Nutrition and Metabolism, between 2006 and 2022 were retrospectively reviewed. An ethical approval (file number; 2021/805) was obtained from Istanbul Medical Faculty Ethics Committee before the study and written consent was obtained from each patient or guardian.

A metabolic episode was considered as mild if hospital admission was not required during the metabolic episode, moderate if hospital admission was indicated, or severe if it was a condition requiring intensive care intervention.

3. Results

The study group of 14 patients comprised 5 males and 9 females with a mean age of 16.1 ± 9.1 (median 15.6) years. Mean follow-up period was 10.3 ± 4.7 (range: 8 months-18.6 years; median: 10 years) years. Three patients were diagnosed by expanded newborn screening within the first month of life. In one case (P2) family history was also positive. Another patient diagnosed by expanded newborn screening was lost to follow-up after the first examination and admitted to hospital with hepatic dysfunction at 7 months of age. Eleven patients were diagnosed due to symptomatic presentation. Age at symptomatic presentation ranged from 7 months to 27 years (median 2.5 years). Age at diagnosis in this group ranged from 11 months to 30 years (median: 3.5 years). Eleven patients were born from consanguineous marriages. Most common presenting symptoms were psychomotor retardation (PMR) (n = 6), seizures (n = 5), encephalopathy (n = 5), dystonia (n = 1), Reye-like syndrome (n = 5), muscle weakness (n = 3), autism (n = 1), and recurrent pancreatitis (n = 1) (Table 1). The clinical findings of patients presenting with metabolic attacks are summarized in Table 2.

Table 1.

Presenting symptoms, neurological symptoms, follow-up findings, biochemical findings, and molecular analysis results of the CPT I patients

#	Gender	Presenting symptoms	Age at first symptom	Age at diagnosis	Follow-up duration	Neurological/pathologic findings			Metabolic episodes (n)				Additional findings	Acylcarnitines Initial Follow-up				Mutation analysis
						PMR	SD	Other	Before diagnosis	At diagnosis	After diagnosis	Treatment compliance		C0d	C0/ C16 + C18e	C0d	C0/ C16 + C18e
P1	F	Nonec	None	17 d	8 m	−	−	−	−	−	−	−	−	114	51	43	32	p.S106N (c.317 G > A)*
P2	M	Nonea	None	2 d	4.5 y	−	−	−	−	−	+(1) mild	+	Hepatosteatosis (G1)	137	370	268	3350	p.P247L (c.740C > T)*
P3	F	Hepatic dysfunctionb (7 m)	7 m	1 m	4 y	−	−	−	−	−	+(1) severe	−	Vit Kg Died during metabolic episode (4y)	140	411	223	2027	p.G197X (c.589G > T)f
P4	M	RLE, seizure	6.5 y	6.5 y	16 y	−	−	Left hemiparesis, EEG: Pathologic Cr MR: Intracranial hemorrhage in frontotemporoparietal area	−	+(1) severe	−	+	ECHO: Left ventricular dysfunction, hepatosteatosis (G1), VUR, vit Kg, neutropenia	263	1011	368	1187	p.H473Y (c.1417C > T)*
P5	F	Muscle weakness, fatigue	27 y	30 y	10 y	−	−	−	−	−	−	−	Horseshoe kidney	570	780	133	198	−
P6	F	Encephalopathy	5 y	5 y	12.7 y	−	+	−	−	+(1) moderate	−	+	ADHD, hepatosteatosis (G1)	147	1633	201	418	p.P247L (c.740C > T)*
P7	F	RLE, encephalopathy, seizure	22 m	22 m	9.7 y	−	+	−	−	+(1) severe	−	+	−	169	1056	178	988	p.A478V (c.1433C > A) c.1338–1 G>Ah
P8	F	Muscle weakness	1 y	1.5 y	12.6 y	−	+	−	−	−	+(3) mild (1) moderate (1) severe (1)	−	Hepatosteatosis (G2), elevated CK	318	1987	203	654	–
P9	F	Muscle weakness, dystonia, PMR	12 y	12 y	13.5 y	+ (mild)	+	Lower extremity dystonia, EMG: Neurogenic involvement Cr MR: Increased signal intensity in cerebral hemispheres, periventricular white matter, supratentorium	−	−	−	+	−	137	311	205	854	−
P10	M	Seizure, PMR	3.5 y	3.5 y	14.2 y	+ (mild)	+	Ataxia Cr MR: Symmetrical hyperintense lesions in bilateral parietal periventricular white matter	+(1) mild	−	−	−	−	310	2384	521	4341	−
P11	M	RLE, encephalopathy, status epilepticus, PMR	11 m	11 m	18.6 y	+ (mild)	+	Cr MR: Hyperintensity in left lateral periventricular area	−	+(1) severe	+(7) mild (3) moderate (4)	−	−	115	1437	265	1766	p.E525GfsX32 (c.1573)*
P12	M	Recurrent pancreatitis, PMR	18 m	4 y	7.7 y	+ (moderate)	+	−	+(2) mild (2)	−	+(1) mild	−	Autism, elevated CK, DM-I (11 y)	255	2318	258	3225	p.G446C (c.1336 G > T)*
P13	F	RLE, encephalopathy, status epilepticus, PMR	3 y	3.5 y	3.5 y	+ (moderate)	+	Cr MR: Wide sulci, and gliosis in the parieto-occipital area	+(1) mild	+(1) severe	+(1) severe	−	ECHO: Asymmetric VSH, hepatosteatosis (G2) Died during metabolic episode (8 y)	230	164	198	1237	−
P14	F	PMR	2 y	2 y	−	+ (severe)	+	No walking	−	−	−	+	−	175	324	120	387	−

P: Patient, F: female, M: male, y: years, m: months, d: days ADHD: Attention deficit hyperactivity disorder, CK: Creatine kinase, Cr MR: Cranial magnetic resonance imaging, DM: Diabetes mellitus, EEG: Electroencephalogram, EMG: Electromyography, G: Grade, PMR: Psychomotor retardation, SD: Speech disorder, VUR: Vesicoureteral reflux, VSH: Ventricular septal hypertrophy.

^aDiagnosed by expanded newborn screening based on family history *homozygous mutations

^bDiagnosed by expanded newborn screening, lost to follow-up until became symptomatic at 7 months of age

^cDiagnosed by expanded newborn screening

^d(10–90) (μM/L)

^e(<15)

^fHeterozygous mutation in both parents

^gVitamin K supplementation at follow-up.

^hCompound heterozygous mutations.

Table 2.

Characteristics of patients with CPT I admitted during a metabolic episode.

#	Encephalopathy	Hypoglycemia	Elevated transaminase	Hyperammonemia	Renal tubular acidosis	Status epilepticus	Pancreatitis
6	+		+
7	+	+	+	+
8			+				+
12	+	+	+	+	+	+
13	+		+	+
14	+		+			+

All the patients with PMR had speech disorder, whereas 3 patients had isolated speech disorders without concomitant PMR. Speech disorders including delayed expressive language development, speech articulation disorder, speech retardation, and stuttering were present in nine patients. In two patients, speech retardation was accompanied by global developmental delay.

The age at diagnosis varied between two days and 30 years in the five patients without speech disorder (median: 1 month; mean: 7.84 ± 12.90 years) and between 11 months and 12 years in the 9 patients with speech disorder (median: 3.5 years; mean: 3.80 ± 3.34 years). Only one patient (1/5) without speech disorder experienced a metabolic attack before or during diagnosis, whereas six patients (6/9) with speech impairment experienced a metabolic episode before or at diagnosis. Only one patient (1/5) without speech impairment experienced more than one metabolic attack, while four patients (4/9) with speech impairment had more than one metabolic attack. Of the three patients with speech impairment without psychomotor retardation, those whose speech impairment improved after CPT I treatment were compliant with treatment, whereas the patient whose speech impairment did not improve was non-compliant.

The diagnosis age of 8 patients who do not have PMR ranged between 1 month and 30 years, with a median of 1.75 years and a mean of 5.63 ± 10.13 years. On the other hand, 6 patients with PMR were diagnosed at an age ranging from 11 months to 12 years, with a median of 3.5 years and a mean of 4.31 ± 3.93 years. Before diagnosis, none of the patients (0/8) without PMR had a metabolic attack, however three patients (3/8) without PMR were diagnosed with a metabolic attack. Three patients with PMR (3/6) had a metabolic attack before diagnosis and one (1/6) was diagnosed during a severe metabolic attack. Of the 6 patients who had a severe metabolic attack at some time during their lives, only 2 had PMR and but 4 did not have PMR. Of the patients without PMR, four were compliant with CPT I treatment, as were two of those with PMR.

At diagnosis, free carnitine values ranged from 114 to 570 (normal range: 10–90; median, 172) μmol/L, while the C0/(C16 + C18) ratio ranged from 51 to 2384 (normal range:: <15; median, 895). The patient (P1) with an initial free carnitine level of 114 μmol/L and C0/(C16 + C18) ratio of 51 at diagnosis in expanded neonatal screening, had free carnitine and C0/(C16 + C18) ratio levels decreased to normal levels at follow-up, 43 μmol/L, and 32 respectively. No additional features for the differential diagnosis of metabolic diseases were found in the evaluation of the patients. The diagnosis was confirmed by the detection of CPT1A gene mutations in six patients and heterozygous CPT1A gene mutations were detected in the parents of one patient who died. For the rest of the patients, diagnosis was based on typical acylcarnitine profiles in the repetitive acylcarnitine analysis by tandem mass spectrometry.

Patients were treated with nutritional therapy that involved restriction of prolonged fasts, long-chain fatty acids and utilization of medium-chain fatty acids as a source of energy from fat. After starting treatment for CPT I deficiency, speech disorders improved in 3 patients, although none of these patients received speech therapy.

Eight patients did not have any metabolic episodes throughout the follow-up. Five of these eight patients had speech disorders and six patients had neurologic symptoms besides speech disorders. Only two of these eight patients did not have neurologic findings. In the group of six patients who had metabolic episodes, two did not have any neurological pathologic findings, while two had psychomotor retardation, and four patients had speech disorders.

Additionally, elevated creatine kinase (CK) levels ranging between 250 and 800 U/L (normal range: 20–180 U/L) were detected in seven patients both at the time of initial diagnosis and during the follow-up period. Cardiomyopathy was detected in two patients on echocardiography. Cardiomyopathy resolved in one patient after the first year of dietary treatment. The severity of PMR in six patients at diagnosis did not worsen during the follow-up period. Two patients died at the ages of 4 years and 8 years due to acute metabolic episodes triggered by acute infections. During the follow-up period, three patients had Reye-like episodes because of non-compliance in two. One other patient that had been diagnosed in the neonatal period due to the presence of a family history, had elevated CK (800 U/L) at the age of 4 months that resolved at follow-up.

4. Conclusions

To date, among the deficiency of three isoforms, only L-CPT I (CPT Ia) deficiency has been reported [18]. Since the CPT Ia isoenzyme is predominantly found in the liver tissue, patients are expected to have hypoketotic hypoglycemia and hepatomegaly accompanied by acute liver failure, elevated transaminases, and hyperammonemia, whereas skeletal and cardiac muscle involvement is not expected [5].

Neurological findings in patients with CPT I deficiency are considered to be associated with the central nervous system injury induced by hypoglycemia. However, accumulating evidence suggests that fatty acid metabolism is not only an energy pathway but has a critical role in storing energy in the body and maintaining homeostasis in the organism [[19], [20], [21]]. Central to these metabolic pathways is the carnitine palmitoyltransferase enzyme system. Disruption of these balances in CPT I deficiency may be a factor in the development of specific neurological symptoms including speech disorders. In our study, neurological symptoms were observed in a significant portion of our patients both at the time of diagnosis and persisted during the long-term follow-up period. Besides, it was of particular importance that specific speech disorders were detected in nine patients. All the patients with PMR had speech disorders whereas 3 patients had isolated speech disorder without concomitant PMR. Although other diseases that may cause speech disorders cannot be excluded without performing a whole genome analysis in patients, the increased frequency of observing a specific speech disorder in patients diagnosed with CPT I deficiency suggests that this may be a specific clinical finding associated with sequela of the disease. In this group of patients, the clinical and laboratory findings of the patients, including speech disorders, improved after CPT I treatment, therefore clinically we did not correlate this symptom with other disorders. We suggest that patients with CPT I deficiency undergo a neuropsychological assessment that includes speech evaluation.

The age of diagnosis in patients without PMR was 1.75 years, whereas the median age of diagnosis in patients with PMR was 3.50 years. All the patients with neurologic symptoms were late diagnosed patients, and none of them were diagnosed with expanded newborn screening. Moreover, three patients that were diagnosed during the screening performed due to family history (n = 1) and during expanded newborn screening did not develop any neurologic symptoms (n = 2) during the follow-up period. Of the two siblings, the elder one diagnosed by symptomatic presentation at 5 years of age had neurologic findings while the sibling diagnosed by expanded newborn screening had no neurologic symptoms (Table 1, P2, P6). We propose that early diagnosis, particularly in infancy, may be an important factor in the prevention of neurological findings including speech disorders.

None of the eight patients without PMR had experienced a metabolic attack prior to diagnosis. Meanwhile, three of the six patients with PMR had encountered an attack before diagnosis, indicating that the occurrence of a metabolic attack before diagnosis may increase the risk of developing PMR. Only one of the patients who had frequent metabolic episodes and neurological findings, had 7 metabolic episodes during the follow-up. On the other hand, one of the patients without neurological findings had 3 metabolic episodes during the follow-up. Rest of the patients with metabolic episodes had 2 or less episodes. These observations suggested that the number of metabolic episodes is not a determining factor for the presence of the neurological findings in patients. The free carnitine levels and C0/C16 + C18 ratios at admission and during follow-up was not associated with neurologic symptoms or frequency of metabolic episodes. It was challenging to reconcile the occurrence of a severe metabolic attack with the risk of PMR as only two out of the six patients who required intensive care admission due to a metabolic attack had psychomotor retardation but 4 patients did not have PMR. The fact that we could not correlate treatment non-compliance and PMR may be due to the small sample size of our study group.

The median age at diagnosis for patients with speech disorders was 3.5 years, compared to 1 month for patients without speech disorders. These findings imply that early diagnosis may prevent the onset of both speech impairment as well as psychomotor retardation. The proportion of patients without speech disorders who presented with metabolic attacks before or at diagnosis was lower than that of patients with speech disorders (1/5 vs. 6/9), indicating a potential association between metabolic episodes and the development of speech disorders.

Common variants other than the recurring p.P479L and p.G710E mutations, in the Inuit and Hutterite communities, are very rare in general population [3,4]. Apart from these mutations, majority of the variations identified are private mutations as in our cohort. In our group, except for the two siblings, the variants of CPT1 gene of the patients were different in each patient. Therefore, a relationship could not be established between genetic variants and the presence of neurological findings.

Hippocampal neurogenesis is essential for specific cognitive functions [19]. Continuation of the life cycle of these cells and of neurogenesis require CPT Ia-dependent fatty acid oxidation [19]. Energy deprivation in the specific regions of the brain might be the reason for neurologic sequela in CPT I patients, as in creatine deficiency syndromes [20,21]. Literature indicates that several mechanisms including CPT I enzyme deficiency play a role in numerous neurodegenerative and developmental disorders such as Parkinson's disease, Alzheimer's disease, autism, and ADHD [22]. Accordingly, we consider that the diagnosis of speech disorders in nine of our patients, autism in one patient and ADHD in another patient are significant findings in this respect.

One of the patients diagnosed by newborn screening, with free carnitine 114 μmol/L (n:10–90) and C0/(C16 + C18) ratio of 51 (n:<15) had C0 levels decreased to 43 μmol/L, while the C0/(C16 + C18) ratio decreased to 32 on follow-up, enabling the biochemical diagnosis non feasible. In this patient a missense mutation was detected in homozygous state in the CPT1 gene confirmed the diagnosis implying the importance of newborn screening and molecular analysis in the diagnosis of CPT I deficiency.

Cardiomyopathy was detected in two patients on echocardiography. Cardiomyopathy resolved in the first year of dietary treatment in the patient that was compliant with the dietary therapy. The patient who was noncompliant with dietary therapy died at 8 years of age. Compliance to the dietary treatment might be a factor in the resolution of cardiomyopathy as observed in other defects of the fatty acid oxidation disorders [23].

Given that the CPT Ia enzyme is predominantly found in the liver and kidney, it is considered that the findings of enzyme deficiency are mostly seen in these two organs. Moreover, since CPT Ib isoenzyme is predominantly located in the skeletal muscle tissue, and symptoms related to skeletal muscles are not expected in CPT I deficiency. A previous study reported on a patient homozygous for p.P479L and indicated that myopathic symptoms were the most salient symptoms at first presentation and the patient also had high residual enzyme activity [24]. Additionally, other cases with complaints of muscle cramps, muscle pain, and symptoms of skeletal muscle have also been reported in the literature [5,13,25]. Similarly, in our study, three patients presented with muscle weakness, fatigue, and muscle pain and almost half of the patients (n = 7) were found to have elevated creatine kinase levels during the metabolic episodes they experienced at the time of diagnosis and during the follow-up period. Literature suggests that cell membrane damage is induced by the detergent-like effects of non-esterified fatty acids on muscle cells [5]. In addition, based on the demonstration of fat accumulation in the skeletal muscle sample obtained from a patient diagnosed with CPT I deficiency during an episode of acute metabolic decompensation has led to the hypothesis that abnormal circulation of CPT Ia induced metabolites in the blood impairs fatty acid oxidation in the muscle tissue [26]. Likewise, the released metabolites may impair fatty acid oxidation in the central nervous system, leading to neurological symptoms.

CPT I deficiency is an ultra-rare disease, and the identification of neurological findings in the long-term follow-up of 14 patients will contribute significantly to the clinical spectrum of the disease. In line with the literature, our findings confirmed that the clinical manifestations of CPT I deficiency appear to be wider than previously thought especially when the high frequency of neurologic findings are considered. It is a life-threatening metabolic disorder, necessitating long-term nutritional therapy and intervention during the metabolic episodes. We observed a specific neurologic dysfunction, mainly speech disorders at a large scale, that were unexpected in a fatty acid oxidation disorder. Based on these findings, we suggest that early diagnosis and initiation of therapy is the key factor in the prevention of neurologic sequela. Since CPT I deficiency causes specific neurologic symptoms, an extensive neurological evaluation is essential in patients both at the time of diagnosis and during the follow-up period.

CRediT authorship contribution statement

Mehmet Cihan Balci: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Validation, Writing – original draft, Writing – review & editing. Meryem Karaca: Conceptualization, Investigation, Writing – original draft. Arzu Selamioglu: Data curation, Investigation. Huseyin Kutay Korbeyli: Data curation, Investigation. Asli Durmus: Data curation, Investigation. Belkis Ak: Data curation, Investigation. Tugba Kozanoglu: Data curation, Investigation. Gulden Fatma Gokcay: Conceptualization, Supervision, Writing – review & editing.

Declaration of Competing Interest

None.

Data availability

Data will be made available on request.