Variation of betaine,N,N- dimethylglycine,choline,glycerophosphorylcholine,taurine and trimethylamine- N -oxide in the plasma and urine of overweight people with type 2 diabetes over a two-year period

Abstract

Background

Plasma betaine concentrations and urinary betaine excretions have high test-retest reliability. Abnormal betaine excretion is common in diabetes. We aimed to confirm the individuality of plasma betaine and urinary betaine excretion in an overweight population with type 2 diabetes and compare this with the individuality of other osmolytes, one-carbon metabolites and trimethylamine-N-oxide (TMAO), thus assessing their potential usefulness as disease markers.

Methods

Urine and plasma were collected from overweight subjects with type 2 diabetes at four time points over a two-year period. We measured the concentrations of the osmolytes: betaine, glycerophosphorylcholine (GPC) and taurine, as well as TMAO, and the one-carbon metabolites, N,N-dimethylglycine (DMG) and free choline. Samples were measured using tandem mass spectrometry (LC-MS/MS).

Results

Betaine showed a high degree of individuality (or test-retest reliability) in the plasma (index of individuality = 0.52) and urine (index of individuality = 0.45). Betaine in the plasma had positive and negative log-normal reference change values (RCVs) of 54% and −35%, respectively. The other osmolytes, taurine and GPC were more variable in the plasma of individuals compared to the urine. DMG and choline showed high individuality in the plasma and urine. TMAO was highly variable in the plasma and urine (log-normal RCVs ranging from 403% to −80% in plasma).

Conclusions

Betaine is highly individual in overweight people with diabetes. Betaine, its metabolite DMG, and precursor choline showed more reliability than the osmolytes, GPC and taurine. The low reliability of TMAO suggests that a single TMAO measurement has low diagnostic value.

Keywords

Betaine choline coefficient of reliability diabetes glycerophosphorylcholine index of individuality reference change value taurine

Introduction

Betaine (N,N,N-trimethylglycine, glycine betaine, TMG) is an important osmolyte which is also involved in one-carbon metabolism as a major source of methyl groups in mammals.¹ Betaine can be obtained from the diet² or produced in the body by oxidation of choline. Choline is predominantly present in the diet and the body in the form of phosphatidylcholine.³ Urinary betaine excretion is often elevated in diabetes and renal failure.^4,5 Plasma betaine concentrations and urinary betaine excretions can predict secondary cardiovascular events in people with established vascular disease,⁶ and the plasma concentration of dimethylglycine, the product of betaine metabolism, is an especially strong prognostic marker.^6,7 These predictions are stronger for patients with diabetes. It is therefore important to know how much value can be attached to single measurements of these metabolites. Reference change values (RCVs) are useful to show a change over successive measurements that cannot be attributed to either analytical imprecision or biological variation. Betaine has been shown to have a high reliability in the plasma and urine of healthy young males over the period of a day, and over eight weeks.⁸ A high reliability (or individuality) of urine betaine excretion has been reported in subjects with stable angina over a few years,⁵ and plasma betaine concentrations were found to be consistent in a small sample of normal subjects over a similar time.⁹ It is not known if the high reliability which has been observed for betaine is related to its role in one-carbon metabolism, or its role as an osmolyte.

The objective of this study was to investigate the individuality (using the index of individuality) and test-retest reliability (using the coefficient of reliability) of plasma and urine betaine concentrations in a group of overweight people with type 2 diabetes over a two-year period and compare these data with some related compounds. The other one-carbon metabolites, N,N-dimethylglycine (DMG) and choline were measured. Trimethylamine-N-oxide (TMAO) is another methylamine which has been associated with cardiovascular risk.¹⁰ Wang et al.¹¹ showed that high plasma and choline concentrations were associated with a higher risk of major adverse cardiac events when plasma TMAO is also elevated. We therefore included TMAO in this study to investigate the usefulness of a single measurement as a diagnostic marker in type 2 diabetes. Since betaine is an important osmolyte, we also investigated the test-retest reliability of two other osmolytes, taurine and glycerophosphorylcholine (GPC).

Methods

Study design

Samples were collected during the diabetes excess weight loss (DEWL) study conducted by Krebs et al.¹² Briefly, overweight (body mass index [BMI] >27 kg/m²) subjects with type 2 diabetes were recruited, and the study compared the effects of a high protein versus a high carbohydrate diet on weight and metabolic variables in a randomised controlled trial over two years. Fasting EDTA plasma and a 24-h urine collection were obtained from subjects at four time points at six-month intervals over a two-year period.¹² Fasting samples were collected into EDTA tubes and immediately centrifuged with plasma aliquoted and stored at −80℃. For the present study, there was a complete set of plasma samples for 243 subjects over the four time points and 116 subjects with a complete set of urine samples for the four time points. Markers of glycaemic control and renal function, BMI and age for the subset of DEWL study subjects included in the present study are shown in Table 1. These data were measured during the original DEWL study by Krebs et al.¹² Participants were excluded if they were currently on weight-reducing medications, had weight loss of >5% in the past three months or had a psychiatric or eating disorder. Participants were also excluded if their glycated haemoglobin (HbA_1c) was >9.5% (80 mmol/mol) or they had had renal disease (estimated glomerular filtration rate [eGFR] <60 ml/min/1.73 m² or urine albumin:creatinine ratio [UACR] >30 mg/mmol), abnormal liver enzymes, heart failure, known active malignancy or myocardial infarction in the preceding six months. Details of variables such as weight, plasma glucose, HbA_1c and markers of renal failure throughout the study were previously reported by Krebs et al.¹² Because fibrates have been associated with increased urinary betaine excretion,^13–15 subjects who were on fibrate therapy were excluded from the analysis.

Table 1.

Characteristics of the study population.

		Interquartile range
	Median	25%	75%
Plasma glucose (mmol/L)	7.8	6.5	9.7
HbA_1c (%)	7.7	7.0	8.8
BMI (kg/m²)	34.6	31.3	39.4
Plasma creatinine (µmol/L)	74.2	64.3	88.2
eGFR (mL/min/1.73 m²)	82.2	69.8	95.8
Age (years)	60	53	67

Data collected during the DEWL study.¹² Median values were calculated for all subjects over all four time points.

HbA_1c: glycated haemoglobin; BMI: body mass index; eGFR: estimated glomerular filtration rate; DEWL: diabetes excess weight loss.

Reagents and chemicals

Glycine betaine (HCl), DMG (HCl), choline (iodide), GPC (cadmium chloride complex), TMAO (dihydrate) and taurine were obtained from Sigma (St Louis, MO, USA). N,N,N-Trimethyl-D₉-glycine (HCl) (D₉-betaine) and choline-trimethyl-D₉ (chloride) were obtained from Isotec (Miamisburg, OH, USA), D₉-TMAO was obtained from Cambridge Isotope Laboratories (Andover, MA, USA) and N,N-dimethyl-D₃-glycine (HCl) and 2-aminoethane-D₄-sulfonic acid (deuterated taurine) were obtained from CDN isotopes (Pointe-Claire, Quebec, Canada). Ammonium formate and formic acid were purchased from Sigma-Aldrich (St Louis, MO, USA). High-performance liquid chromatography (HPLC)-grade acetonitrile was obtained from Mallinckrodt (Paris, KY, USA), and methanol was obtained from Merck (Darmstadt, Germany).

Sample analysis

Samples were analysed using tandem mass spectrometry (LC-MS/MS) based on the method of Holm et al.¹⁶ Fifty microlitres of plasma or urine were extracted into 1.0 mL of 10% methanol, 90% acetonitrile containing 10 µmol/L of deuterated internal standard. Internal standards added include: D₉-betaine (mass transition 127 → 68), D₃-dimethylglycine (mass transition 107 → 61), D₉-choline (mass transition 113 → 69), D₄-taurine (mass transition 128 → 80) and D₉-TMAO (85 → 66). Plasma and urine samples were analysed for betaine, DMG, free choline, GPC, taurine and TMAO. The mass transitions and internal standards used to measure each analyte are shown in Table 2. An ABSciex (Mulgrave, VIC, Australia) API4000 triple quadrupole mass spectrometer was connected to a Shimadzu Prominence (Kyoto, Japan) HPLC with a binary pump system and used for analysis with a flow rate of 0.3 mL/min. The injection volume was 10 µL. Samples were separated using a Cogent 100 mm × 2.1 mm, 4 µm Diamond Hydride silica column (Microsolv Technologies, NJ, USA). The oven temperature was 40℃. Line A contained 10 mmol/L ammonium formate and 10 mmol/L formic acid in 50% distilled water and 50% acetonitrile. Line B contained 10 mmol/L ammonium formate, 10 mmol/L formic acid, 90% acetonitrile and 10% water. The gradient used for the analysis was from 50% A to 100% A over 8 min. An electrospray ion source was used with an ion source temperature of 350℃. The dwell time was 150 ms. Voltages and mass transitions used in the analysis are given in Table 2.

Table 2.

Mass transitions and voltages used in the mass spectrometry methods.

Analyte	Mass transition	DP (V)	CE (V)	CXP (V)	Internal standard
Betaine	118.2 → 59.0	56	27	4	D₉-betaine
Dimethylglycine	104.1 → 58.1	51	21	4	D₃-dimethylglycine
Choline	104.1 → 60.2	16	25	4	D₉-choline
GPC	258.2 → 104.0	66	23	8	D₉-betaine
Taurine	124.1 → 79.9	−55	−30	−13	D₄-taurine
TMAO	76.1 → 58.1	16	27	10	D₉-TMAO

DP: decoupling potential; CE: collision energy; CXP: collision cell exit cell potential; TMAO: trimethylamine-N-oxide; GPC: glycerophosphorylcholine.

Taurine was separated isocratically using an XBridge amide column (100 × 2.1 mm, 3.5 µm, Waters, USA). The mobile phase contained 20% water and 80% acetonitrile, the flow rate was of 0.3 mL/min, the injection volume was 10 µL and the oven temperature was 40℃. Taurine was measured in negative ion mode.

Urine creatinine was measured during the DEWL study on routine analysers at three locations in New Zealand using the Jaffè method. eGFR was calculated using the abbreviated modification of diet in renal disease (MDRD) equation.¹⁷

Statistical analysis

Statistical analyses were performed using SigmaPlot (Version 11.2). Differences between the subjects were investigated for the analytes in the urine and plasma using one-way repeated measures analysis of variance. Mann–Whitney rank sum tests were used to test for differences in metabolite concentrations between the dietary treatments, and between the genders. A P value <0.05 was considered to be statistically significant. The within-individual variation over the four time points was also investigated. The index of individuality was also calculated for each compound, where a lower number represents higher individuality of the data. The index of individuality was calculated as the ratio of within-subject coefficient of variation (CV_i) to between-subject coefficient of variation (CV_g). CV_i was calculated using the formula CV_i = (CV_t²–CV_a²)^½, where CV_t is the total imprecision and CV_a is the analytical imprecision.

The test-retest reliability of the results was also determined by calculating the coefficient of reliability (Cronbach’s α) in order to compare the data with those from previous studies. Coefficient of reliability = (N / (N−1)) ((total variance − sum of individual variance) / total variance), where N is the number of subjects. Because of the non-normal distribution in urine, the coefficient of reliability was calculated for log-transformed data. A reliability coefficient close to 1 represents a high degree of test-test reliability,¹⁸ whereas a coefficient of zero corresponds to random data (no consistency). A coefficient of reliability was calculated for each compound using only subjects where there was a complete set of four time points (baseline, 6 months, 12 months and 24 months).

Log-normal RCVs for a positive and negative change were calculated as described by Fraser¹⁹ using the equation RCV_pos = [exp(1.96 × 2^½× σ) − 1] × 100, and RCV_neg = [exp(−1.96 × 2^½× σ) − 1] × 100, where σ is the log-normal distribution defined as σ = [ln(CV_t²+ 1)]^½, and CV_t is the total imprecision. A Z-score of 1.96 was used to calculate RCVs, corresponding to P = 0.05, or a 95% chance that a change is significant.

Results

Quality control

The analytical coefficients of variation throughout the study were generally below 7% for betaine, DMG, choline, taurine and TMAO. However, GPC had a greater analytical CV of 21% in the plasma, which may be a consequence of lack of a suitable commercially available internal standard. The concentrations of GPC in the samples were also closer to the detection limits than the other analytes.

Effect of diet

No significant differences (P < 0.05) in the concentrations of betaine, DMG, choline, taurine, GPC or TMAO were observed between the diet treatments for both plasma and urine.

Gender differences

There were significant differences observed between the genders. Males had significantly higher (P < 0.01) plasma betaine, DMG and choline. Plasma concentrations of TMAO, GPC and acetylcarnitine were not significantly different between the genders. There were no significant gender differences detected for any metabolites in the urine.

Individuality

The coefficients of reliability and indices of individuality are shown in Table 3 for plasma and Table 4 for urine. Because differences were observed between males and females in betaine concentration, data are displayed for each gender.

Table 3.

The test-retest reliability and individuality in plasma.

	Median (µmol/L)	Interquartile range		Reliability coefficient	CV_a%	CV_i%	Index of individuality	Log-normal RCV%
	Median (µmol/L)	25%	75%	Reliability coefficient	CV_a%	CV_i%	Index of individuality	Positive	Negative
All subjects (N = 243)
Betaine	30.3	24.9	36.9	0.75	4.6	14.9	0.52	53.7	−34.9
DMG	2.34	1.89	2.96	0.79	5.1	17.2	0.30	63.8	−38.9
Choline	10.4	8.7	12.3	0.67	5.3	14.4	0.69	52.6	−34.5
GPC	1.54	1.17	2.12	0.42	20.9	21.5	0.60	125.5	−55.7
Taurine	49.3	41.4	60.5	0.09	7.7	28.7	1.45	124.0	−55.3
TMAO	6.31	4.09	10.4	0.17	5.7	63.3	0.84	402.6	−80.1
Males (n = 99)
Betaine	32.1	26.3	40.0	0.77	4.6	14.1	0.47	50.5	−33.6
DMG	2.50	2.10	3.12	0.83	5.1	16.9	0.29	62.5	−38.5
Choline	10.9	9.27	12.8	0.67	5.3	13.6	0.61	49.5	−33.1
GPC	1.52	1.18	2.15	0.48	20.9	18.2	0.61	112.6	−53.0
Taurine	50.3	42.3	60.9	0.13	7.7	27.1	1.62	115.1	−53.5
TMAO	6.64	4.04	11.0	0.18	5.7	66.9	0.79	442.4	−81.6
Females (n = 144)
Betaine	28.9	23.4	35.1	0.72	4.6	15.5	0.55	56.1	−35.9
DMG	2.20	1.78	2.77	0.76	5.1	17.3	0.30	64.2	−39.1
Choline	9.92	8.46	11.7	0.67	5.3	14.8	0.70	54.2	−35.1
GPC	1.55	1.14	2.09	0.30	20.9	23.2	0.63	132.9	−57.1
Taurine	48.8	40.8	60.4	0.08	7.7	29.6	1.48	129.1	−56.3
TMAO	6.08	4.14	10.2	0.17	5.7	60.5	0.83	373.0	−78.9

Median values were calculated from all four time points.

CV_a: analytical imprecision; CV_i: within-subject coefficient of variation; RCV: reference change values; DMG: N,N-dimethylglycine; GPC: glycerophosphorylcholine; TMAO: trimethylamine-N-oxide.

Table 4.

The test-retest reliability and individuality in urine.

	Median (µmol/L)	Interquartile range		Reliability coefficient	CV_a%	CV_i%	Index of individuality	Log-normal RCV%
	Median (µmol/L)	25%	75%	Reliability coefficient	CV_a%	CV_i%	Index of individuality	Positive	Negative
All subjects (N = 116)
Betaine	25.9	13.0	52.8	0.83	6.6	38.8	0.45	186.3	−65.1
DMG	6.75	4.10	9.76	0.81	6.8	32.0	0.36	142.0	−58.7
Choline	3.63	2.64	5.25	0.75	12.0	33.4	0.38	159.8	−61.5
GPC	0.56	0.41	0.77	0.59	17.6	28.5	0.39	146.9	−59.5
Taurine	29.0	10.3	54.9	0.73	5.8	56.1	0.22	329.2	−76.7
TMAO	70.7	46.3	115	0.31	5.3	70.2	0.35	480.4	−82.8
Males (n = 48)
Betaine	32.4	13.1	66.5	0.91	6.6	38.3	0.47	182.8	−64.6
DMG	6.92	4.27	9.46	0.89	6.8	25.3	0.50	104.2	−51.0
Choline	3.51	2.46	4.90	0.72	12.0	31.7	0.44	149.5	−59.9
GPC	0.49	0.37	0.72	0.83	17.6	10.5	0.21	75.5	−43.0
Taurine	36.9	19.0	58.1	0.68	5.8	49.7	0.80	271.0	−73.0
TMAO	65.7	45.8	94.7	0.30	5.3	53.2	0.79	301.4	−75.1
Females (n = 68)
Betaine	25.4	13.8	53.0	0.78	6.6	30.6	0.35	133.4	−57.2
DMG	6.99	4.39	10.8	0.79	6.8	25.1	0.25	103.2	−50.8
Choline	3.94	2.86	5.79	0.78	12.0	23.2	0.24	103.8	−50.9
GPC	0.62	0.46	0.87	0.52	17.6	22.7	0.28	118.2	−54.2
Taurine	19.6	6.44	51.5	0.75	5.8	41.6	0.14	205.7	−67.3
TMAO	86.8	49.4	130.0	0.32	5.3	57.6	0.28	343.3	−77.4

Median values were calculated from all four time points. The coefficients of reliability for the urine results were calculated on log-transformed data because the data were not normally distributed.

CV_a: analytical imprecision; CV_i: within-subject coefficient of variation; RCV: reference change values; DMG: N,N-dimethylglycine; GPC: glycerophosphorylcholine; TMAO: trimethylamine-N-oxide.

The results show that there is a high degree of reliability for betaine (coefficient of reliability (coefficient) = 0.75) and DMG (coefficient = 0.79) in the plasma. Of all the analytes investigated, betaine (coefficient = 0.83) and DMG (coefficient = 0.81) concentrations showed the highest degree of reliability in the urine.

While betaine showed a high level of individuality in the plasma and urine, the other osmolytes were more variable. GPC had a moderate degree of reliability in the plasma (coefficient = 0.42) and urine (coefficient = 0.59). Taurine was much more variable in the plasma (coefficient = 0.09) than in the urine (coefficient = 0.73).

Choline concentrations showed a moderate reliability in the plasma (coefficient = 0.67) and in the urine (coefficient = 0.75). TMAO had a low reliability in both the plasma (coefficient = 0.17) and the urine (coefficient = 0.31).

The indices of individuality show similar trends for the metabolites. A smaller number for the index of individuality indicates low within-individual variability (high individuality). Plasma taurine had a particularly high index, representing a low individuality. GPC and TMAO in the plasma also had high indices of individuality of 0.60 and 0.84, respectively.

The RCVs were much lower in the plasma than the urine for all analytes. A change in plasma betaine concentration of +51% (or −34%) in males and +56% (or −36%) in females can be considered significant (P < 0.05). A change in plasma DMG concentrations of +64% (−39%) is significant (P < 0.05). However, a change in plasma TMAO of +403% (or −80%) and a change of plasma taurine of +124% (or −55%) is required to be significant (P = 0.05). Larger log-normal RCV ranges were observed for urine metabolite excretions compared to the plasma. For example, urine betaine had a positive log-normal RCV of 186 and a negative log-normal RCV value of −65 over both genders. TMAO had the highest RCVs in the urine of +480% (or −83%) over both genders.

Discussion

Effect of diet

No significant differences in metabolite concentrations were observed between the high protein and high carbohydrate diets, which is surprising considering the high betaine content of many high carbohydrate foods, such as wheat-based cereal products.² However, this observation is likely to be at least partially a result of deviations the DEWL study participants may have made from their prescribed diets.¹²

Comparison of the DEWL study group with other studies

The concentrations of osmolytes and other metabolites in the plasma of this overweight group with diabetes are similar to those reported in the general population.^1,8,20,21 Comparisons of the DEWL study data with other studies are shown in Table 5. Published reference ranges for betaine and DMG are described in the review of Lever and Slow.¹ The overall median urine betaine excretion of 25.9 mmol/mol creatinine in this study population is higher than in a population without diabetes (median = 5.9 mmol/mol creatinine⁸). Some participants in the DEWL study were likely to be excreting more than the estimated New Zealand median dietary betaine intake (227 mg/day²), and thus dependent on choline oxidation to make betaine for osmoregulation and one-carbon metabolism. Elevated betaine excretion in diabetes has been previously reported.^4,5,22 We did not detect statistically significant trends for excretion to increase with time, possibly with a relatively small sample size (116 subjects) a longer time than two years may be required to show progressions in betaine excretions in diabetes. Schartum-Hansen et al.⁵ showed that urinary betaine excretion is consistent in people with diabetes over approximately three years. Lever et al.²³ showed that abnormal betaine excretion can persist for years in individuals. Plasma dimethylglycine concentrations are consistent with those reported in other populations,¹ and though the excretion of dimethylglycine is probably elevated, more data are needed to make this conclusive.

Table 5.

Comparison of DEWL data with other studies.

	DEWL study		Previous reports		Comments on previous
	Median	Reliability	Median	Reliability	Cohort type and reference
Plasma betaine	30.3	0.75	35.9		Stable angina⁷
			38.7	0.72	Healthy males, 24 h⁸
			41.1		Referred for CVD testing¹¹
			38.1	0.43	Healthy males, 8 weeks⁸
Plasma DMG	2.3	0.79	3.8	0.93	Stable angina⁷
			2.8	0.63	Healthy males, 24 h⁸
			2.1	0.15	Healthy males, 8 weeks⁸
Plasma choline	10.4	0.67	8.9		Stable angina⁷
Plasma choline	10.4	0.67	9.8		Referred for CVD testing¹¹
Plasma taurine	49.3	0.09	58^a		Healthy controls²¹
Plasma TMAO	6.3	0.17	3.7		Referred for CVD testing¹¹
Urine betaine	25.9	0.83	7.4	0.73	Stable angina⁵
			22.2	0.70	Stable angina + diabetes⁵
			5.9	0.59	Healthy males, 24 h⁸
			6.8	0.78	Healthy males, 8 weeks⁸
Urine DMG	6.75	0.81	1.1	0.61	Healthy males, 24 h⁸
Urine DMG	6.75	0.81	2.7	0.57	Healthy males, 8 weeks⁸

Plasma median concentrations are all in units of µmol/L; urinary excretions as mmole/mole creatinine. Reliability is the test-retest reliability coefficient.

DMG: N,N-dimethylglycine; TMAO: trimethylamine-N-oxide; DEWL: diabetes excess weight loss; CVD: cardiovascular disease.

Mean.

Plasma choline concentrations (overall median 10.4 µmol/L) in the overweight subjects with diabetes were consistent with data on the general population.²⁴ Cardiovascular events such as myocardial infarction have been associated with high plasma choline concentrations, though the event probably causes the elevation. Plasma choline concentrations ≥25 µmol/L have been reported to be predictive of major cardiac events.²⁴ The mean plasma taurine concentration of 53.5 µmol/L observed for the DEWL study samples is in good agreement with Laidlaw et al.,²¹ who published a mean value (±SD) of 58 ± 16 µmol/L in apparently healthy subjects. There are few reports published on the osmolyte GPC, especially in diabetes, so it is difficult to compare the results with other studies. However, the present study shows that the concentrations of GPC in plasma and urine of overweight people with diabetes are considerably lower than the other osmolytes, betaine and taurine. This could be expected because GPC is synthesised intracellularly in response to osmotic stress, rather than accumulated by active transport from the circulation.²⁵

Individuality

The results of the present study are consistent with the high coefficient of reliability of 0.73 which was reported for urine betaine concentrations by Schartum-Hansen et al.⁵ in a population with coronary artery disease; the coefficient in the subgroup with diabetes was similar (0.70) despite a much higher median excretion in this group.⁵ The present study found an even higher coefficient of reliability of 0.83 for urine betaine (0.91 in the males). The main difference in these populations is that the present study investigated people with diabetes who were overweight, whereas Schartum-Hansen et al.⁵ investigated a larger group of cardiovascular patients that included a significant minority with diabetes. A high reliability of betaine in the plasma and urine was also reported by Lever et al.^8,9 in healthy subjects. The coefficients observed for people with type 2 diabetes in the present study are generally consistent with those of Lever et al.⁸ Urinary DMG excretion may have slightly lower individuality than betaine, and free choline lower still.

Plasma dimethylglycine is much more individual in this population than in a small group of healthy young males,⁸ and these values can be expected to be population dependent. Svingen et al.⁷ also reported a high coefficient of reliability (0.93) for plasma DMG in people with stable angina. Plasma DMG was also shown to be predictive of secondary cardiovascular events,⁷ which is in agreement with the findings of Lever et al.⁶ The predictive value of single measurements of plasma DMG is not just a reflection of its correlation with homocysteine,⁷ and as noted before, it could be expected that changes in plasma DMG with time may be a useful additional tool in patient management, and this should be investigated.

The other osmolytes, GPC and taurine showed less test-test reliability in the plasma than betaine. The moderately high individuality and low RCVs suggest that plasma choline may be useful as an indicator of cardiovascular risk in this population. The high variability in plasma taurine (index of individuality = 1.45) observed here suggests that fluctuations in diabetic control and changes in the osmotic status in the body may be having an effect. However, taurine was more individual in the urine in this population (index of individuality = 0.22). Plasma GPC concentrations showed moderate individuality (index of individuality = 0.60) in this population.

The plasma TMAO concentrations were highly variable in this study group. Plasma TMAO has been promoted as a predictive marker of cardiovascular disease.¹⁰ However, in the DEWL study group, both plasma concentrations and urinary excretion of TMAO are highly variable. TMAO concentrations are especially affected by the dietary intake of marine foods. TMAO is an osmolyte in many marine animals, but not a mammalian osmolyte, and is often a marker of eating fish. TMAO concentrations are also raised in people on high choline diets or taking lecithin, betaine or carnitine supplements.²⁶ Elevated plasma concentration TMAO has been previously proposed as a maker of renal disease.^27,28 The high intraindividual variability for TMAO concentrations observed in this population may make it difficult to make clinical decisions on the basis of a single measurement.

A limitation on this study is that most subjects lost weight during it,¹² and the dietary changes may be expected to affect the handling of some metabolites. This makes the high level of consistencies reported here more remarkable, and we may have underestimated the individuality of some of those we studied. However, we have sought trends for changes with treatment and time and have found none that approach statistical significance.

Conclusions

The high test-retest reliability of betaine in this study group shows that the disruption of osmolytes by diabetes has little effect on within-individual reproducibility. This indicates that betaine concentrations are likely to be homeostatically controlled by one-carbon metabolism, even in diabetes. Betaine concentrations measured at any given time provide information about the betaine status of each individual. Other one-carbon metabolites (choline and DMG) also had high test-retest reliability in the urine and plasma. The other osmolytes: GPC and taurine were considerably more variable than betaine, meaning single measurements are likely to be of limited clinical use. The low individuality observed for plasma and urine TMAO in this study may cast some doubt on its value as a disease marker (at least in this population) despite evidence for its association with vascular disease,^10,11 and may be related to the consumption of seafood.

Footnotes

Acknowledgements

The authors thank the DEWL study team for allowing the use of the DEWL study samples in the present study.

Declaration of conflicting interests

None declared.

Funding

The DEWL study was funded by the Health Research Council of New Zealand (06/337). The authors also acknowledge support from the National Heart Foundation of New Zealand and the Maurice and Phyllis Paykel Trust and made use of equipment funded by New Zealand Lottery Health Research.

Ethical approval

Ethical approval was obtained to measure betaine and related metabolites in the DEWL study samples from the Multi-region Ethics Committee of New Zealand (MEC/06/08/081).

Guarantor

SC.

Contributorship

JDK and HL were involved with designing and carrying out the DEWL study. JDK obtained ethical approval. CJM carried out measurements for betaine and related compounds on the study samples. CMF advised on the statistical analysis. All authors contributed to and edited the manuscript.

References

Lever

Slow

. The clinical significance of betaine, an osmolyte with a key role in methyl group metabolism. Clin Biochem 2010; 43: 732–744.

Slow

Donaggio

Cressey

. The betaine content of New Zealand foods and estimated intake in the New Zealand diet. J Food Comp Anal 2005; 18: 473–485.

Magne Ueland

. Choline and betaine in health and disease. J Inherit Metab Dis 2011; 34: 3–15.

Lever

Sizeland

Bason

. Abnormal glycine betaine content of the blood and urine of diabetic and renal patients. Clin Chim Acta 1994; 230: 69–79.

Schartum-Hansen

Magne Ueland

Ringdall Pedersen

. Assessment of urinary betaine as a marker of diabetes mellitus in cardiovascular patients. PLoS One 2013; 8: e469454–e469454.

Lever

George

Elmslie

. Betaine and secondary events in an acute coronary syndrome cohort. PLoS One 2012; 7: e37883–e37883.

Svingen

GFT

Ueland

Pedersen

EKR

. Plasma dimethylglycine and risk of incident acute myocardial infarction in patients with stable angina pectoris. Arterioscler Thromb Vasc Biol 2013; 33: 2041–2048.

Lever

Atkinson

Slow

. Plasma and urine betaine and dimethylglycine variation in healthy young male subjects. Clin Biochem 2009; 42: 706–712.

Lever

Sizeland

Frampton

. Short and long-term variation of plasma glycine betaine concentrations in humans. Clin Biochem 2004; 37: 184–190.

10.

Wang

Kipfell

Bennet

. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature 2011; 472: 57–63.

11.

Wang

Tang

WHW

Buffa

. Prognostic value of choline and betaine depends on intestinal microbiota-generated metabolite trimethylamine-N-oxide. Eur Heart J 2014; 35: 904–910.

12.

Krebs

Elley

Parry-Strong

. The Diabetes Excess Weight Loss (DEWL) trial: a randomised controlled trial of high-protein versus high-carbohydrate diets over 2 years in type 2 diabetes. Diabetologia 2012; 55: 905–914.

13.

Lever

George

Slow

. Fibrates may cause an abnormal urinary betaine loss which is associated with elevations in plasma homocysteine. Cardiovasc Drugs Ther 2009; 23: 395–401.

14.

Lever

George

Atkinson

. Plasma lipids and betaine are related in an acute coronary syndrome cohort. PloS One 2011; 6: e21666–e21666.

15.

Lever M, McEntyre CJ, George PM, et al. Fenofibrate causes elevation of betaine excretion but not of other osmolytes by healthy adults. J Clin Lipidol 2014; doi: 10.1016/j.jacl.04.001.

16.

Holm

Magne Ueland

Kvalheim

. Determination of choline, betaine and dimethylglycine in plasma by a high-throughput method based on normal phase chromatography-tandem mass spectrometry. Clin Chem 2003; 49: 286–294.

17.

Jones

GRD

Lim

E-M

. The national kidney foundation guideline on estimation of the glomerular filtration rate. Clin Biochem Rev 2003; 24: 95–98.

18.

Cronbach

. Coefficient alpha and the internal structure of tests. Psychometrika 1951; 16: 297–334.

19.

Fraser

. Reference change values. Clin Chem Lab Med 2012; 50: 807–812.

20.

Konstantinova

Tell

Vollset

. Divergent associations of plasma choline and betaine with components of metabolic syndrome in middle age and elderly men and women. J Nutr 2008; 138: 914–920.

21.

Laidlaw

Schultz

Cecchino

. Plasma and urine taurine levels in vegans. Am J Clin Nutr 1988; 47: 660–663.

22.

Dellow

Chambers

Lever

. Elevated glycine betaine excretion in diabetes mellitus patients is associated with proximal tubular dysfunction and hyperglycemia. Diab Res Clin Pract 1999; 43: 91–99.

23.

Lever

Atkinson

George

. Abnormal urinary excretion of glycine betaine may persist for years. Clin Biochem 2007; 40: 798–801.

24.

Danne

Lueders

Storm

. Whole blood choline and plasma choline in acute coronary syndromes: prognostic and pathophysiological implications. Clin Chim Acta 2007; 383: 103–109.

25.

Burg

Ferraris

. Intracellular organic osmolytes: function and regulation. J Biological Chem 2008; 283: 7309–7313.

26.

Mackay

McEntyre

Henderson

. Trimethylaminuria: causes and diagnosis of a socially distressing condition. Clin Biochem Rev 2011; 31: 3–13.

27.

Bell

Lee

. Nuclear magnetic resonance studies of blood plasma and urine from subjects with chronic renal failure: identification of trimethylamine-N-oxide. Biochim Biophys Acta 1991; 1096: 101–107.

28.

Bain

Faull

Fornasini

. Accumulation of trimethylamine and trimethylamine-N-oxide in end-stage renal disease patients undergoing haemodialysis. Nephrol Dial Transplant 2006; 21: 1300–1304.