Improving the specificity of the [13C]mixed triacylglycerol breath test by estimating carbon dioxide production from heart rate

Abstract

Background:

The [13C]mixed triacylglycerol (MTG) breath test is a non-invasive measure of fat digestion and can be used to assess the need for enzyme replacement therapy in children with cystic fibrosis (CF). However, it lacks specificity. Quantitation of cumulative percent dose recovered (cPDR) requires knowledge of carbon dioxide production rate (VCO2). A resting value is assumed, but children are unlikely to remain at rest during the test.

Objective:

To improve the specificity and therefore the positive predictive value (PPV) of the MTG breath test using calibrated heart rate monitors to estimate non-resting VCO2.

Design:

Proof of concept study.

Subjects:

Six children with CF, 10 healthy children and eight healthy adults performed [13C]MTG breath tests.

Methods:

Heart rate monitors were worn throughout the test. Non-resting VCO2 was estimated continuously from heart rate. Percentage dose recovered was calculated using predicted resting VCO2, measured resting VCO2 and non-resting VCO2 estimated from heart rate. Physical activity level (PAL) was taken as cPDR calculated using non-resting VCO2 divided by cPDR calculated using measured resting VCO2. The cutoff point was determined using two graph-receiver operator characteristics.

Results:

Use of calibrated heart rate monitors to estimate non-resting VCO2 improved the specificity of the test. The PPV increased from 0.67 to 0.99. PAL was 1.3 in adults and children who performed the test in hospital, and 1.7 in children who performed the test at home.

Conclusion:

Individually calibrated heart rate monitors are useful tools to estimate non-resting VCO2 during the [13C]MTG breath test.

Introduction

There is a well-recognized need for a safe and simple method of assessing fat digestion, especially in children with cystic fibrosis (CF) (Amarri and Weaver, 1995). CF is the most common cause of exocrine pancreatic insufficiency in children. The mixed triacylglycerol (MTG) 13C-breath test is a non-invasive measure of intraluminal fat digestion (Ghoos et al., 1981; Vantrappen et al., 1989). 13C-breath tests are particularly attractive for use in children because they are non-invasive, and pose no radiation hazard. The 13C-MTG breath test has been used to assess the need for pancreatic enzyme replacement therapy (PERT) in children with CF and to assess the ideal dose (Amarri et al., 1997; de Boeck et al., 1998), but it has not been widely adopted because it lacks specificity.

MTG is a synthetic triacylglycerol (1,3 distearyl 2-[1-13C]octanoyl-glycerol). Pancreatic lipases preferentially hydrolyse triacylglycerol at the sn-1 and -3 positions, releasing 13C-labelled octanoate and monoacylglycerol, which are rapidly absorbed independently of bile salts, and oxidized in the liver with the generation of 13CO2. Quantitation of tracer recovery requires knowledge of carbon dioxide production rate (VCO2) because labelled CO2 originating from oxidation of the tracer is diluted by endogenously produced CO2 from metabolism. A resting value based on basal metabolic rate (BMR) is often assumed (Swart et al., 1997; van Dijk-van Aalst et al., 2001), but even if resting VCO2 is measured (Kalivianakis et al., 1997; Amarri et al., 1998), this may not be appropriate unless subjects remain seated and fast for the duration of the test. If subjects are not at rest and not in the post-absorptive state, for example, after an overnight fast, when all food from the last meal has been absorbed from the small intestine, use of resting VCO2 will underestimate the true CO2 production rate resulting in an underestimation of the quantity of tracer recovered. We have previously shown that negligible amounts of 13C are excreted in stool following ingestion of [13C]MTG by healthy children and children with CF, who took their normal dose of PERT (Slater et al., 2002a).

The hypothesis tested in this study is that the poor specificity of the [13C]MTG breath test is owing, in part, to lack of knowledge of the true CO2 production rate during the test. The aims of the study were to use calibrated heart rate monitors to estimate non-resting VCO2 during the [13C]MTG breath test in healthy adults, healthy children and children with CF, to determine the reference range of 13C excretion in healthy subjects, and to determine the physical activity level (PAL) during the test so that cumulative percent dose recovered (cPDR) in children with CF can be corrected for non-resting VCO2 without the need for individual heart rate calibration.

Materials and methods

[13C]MTG breath test

Subjects were asked to avoid foods that are naturally enriched with 13C in the days preceding the test (Schoeller et al., 1980; Morrison et al., 2000). After an overnight fast, six children with CF,10 healthy control children and eight healthy adults ingested 10 mg kg−1 body weight [13C]MTG (Cambridge Isotopes Laboratory Inc., Andover, MA, USA) baked in a biscuit composed of 90 g rolled oats, 40 g butter and 40 g honey (weighed to within 1 g). The honey and butter were melted over a bain-marie. The tracer (weighed to 0.001 g) was stirred into the melted ingredients until dissolved and distributed evenly. The oats were added and the mixture and stirred thoroughly until all the liquid was absorbed. The mixture was spread over the bottom of an aluminium foil-lined tray (160 mm2, surface area 0.0256 m2 or 170 mm diameter, surface area 0.0227 m2), and baked at 130°C for 15 min in a fan-assisted oven. The bowls, spoons, baking tray and foil were weighed before and after use, so that the tracer weight could be corrected for losses during preparation of the test meal (Slater et al., 2002b). The nutritional composition of the test meal was calculated from food composition tables (Holland et al., 1991). This recipe provides 3.13 MJ and is sufficient for two 75 kg adults or 3–4 children, depending on their age and size. Subjects were given a portion equivalent to 17–20% of their daily energy requirements (Department of Health, 1991) together with a drink of unsweetened orange juice or water, as they preferred. In vitro experiments had confirmed that the tracer was homogeneously distributed through the biscuit and there was no loss during preparation (Slater, 2004). Children with CF did not take their PERT with the test meal. A light lunch, composed of foods with low 13C natural abundance (Morrison et al., 2000) was taken 4 h after the test meal. Children with CF took their usual dose of PERT with this meal.

Heart rate monitors (Polar Vantage NV, Polar Electro Oy, Kempele, Finland) were worn throughout the test. Alveolar breath was sampled by exhalation into an Exetainer breath-sampling vial (Labco, High Wycombe, UK) through a straw until condensation appeared on the inside wall of the vial (Slater et al., 2004). The cap was replaced immediately. Breath samples were collected at baseline and every 20 min for 6 h. Resting VCO2 was measured 3 h after the test meal using a ventilated hood indirect calorimeter (GEM, Nutren Technologies, Manchester, UK). Body weight was measured (to 0.1 kg) using Seca scales (model 707, Seca Ltd, Birmingham, UK). Height was measured (to within 1 mm) using a Holtain stadiometer (Holtain, Crymych, Dyfed, UK). Body mass index (kg m−2) was calculated in adults. In children, body mass index was expressed as standard deviation scores relative to the UK 1990 reference data (Cole et al., 1995) using software from the Child Growth Foundation (London, UK).

Eight healthy control children and one child with CF performed the test in the familiar surroundings of their own homes. Two healthy controls and five children with CF performed the test at the day care ward at the Royal Hospital for Sick Children, Glasgow. Resting metabolic rate was measured and the heart rate calibration procedure described below was performed within 1 week of the breath test in children, who performed the test at home. Breath sample collection was supervised by the same researcher in all tests to avoid systematic differences in sample collection.

Calibration of heart rate monitors

At the end of the test or within 1 week, heart rate was calibrated against VCO2 in all subjects (adults, healthy children and children with CF) by making simultaneous measurements with the subject lying on a bed, sitting, standing and during a continuous series of increasing workloads on a bicycle ergometer (Slater et al., 2006). Data (averaged over 1-min periods) were downloaded from the heart rate monitor and the indirect calorimeter to a personal computer (PC) and read into a proprietary spreadsheet (Microsoft Excel 97 SR2). Resting data collected earlier in the day were also included. Time data were matched to give simultaneous readings of both heart rate and VCO2. Heart rate was smoothed by averaging data two points forward of any particular time point. Smoothing heart rate in this way gives a more accurate estimate of resting VCO2 (Slater et al., 2006). VCO2 (mmol min−1 m−2) was plotted against heart rate with heart rate the independent variable plotted on the x axis and VCO2 the dependent variable plotted on the y axis. A sigmoid function was fitted to the data using the ‘Solver’ function of Microsoft Excel (Walsh and Diamond, 1995). Solver used non-linear regression to determine, by iteration, the value of the constants that gave the line of best fit through the data (Motulsky and Ransnas, 1987).

Determination of percentage dose recovered in breath CO2

Breath 13CO2 abundance was measured by continuous-flow isotope ratio mass spectrometry as previously described (Prosser et al., 1991; Slater et al., 2004). The system comprised a 20–20 isotope ratio mass spectrometer (IRMS) interfaced to an automated breath carbon analyser (Hydra, PDZ Europa, Crewe, UK). Measurements were made against a reference gas traceable to international standards (Coplen, 1996). Variability in the amount of CO2 in the breath-sampling vial can be tolerated as the measured isotope ratio does not vary with sample size over the expected concentration range. This phenomenon is known as linearity. The IRMS used in this study is linear over the range from 0.5 to 7% CO2 in air (Slater, 2004). The amount of CO2 in exhaled breath is usually between 2 and 5%. Data from vials containing less than 0.5% CO2 were rejected.

Enrichment was calculated by subtracting the abundance of the baseline sample (p.p.m. 13C) from that of the post-dose sample (Schoenheimer and Rittenberg, 1939; Slater et al., 2001). The percentage dose recovered (PDR) in each breath sample was calculated using the following equation:

and using each of the following values of VCO2:

  1. 1

    A constant value of resting VCO2 predicted from body surface area (Shreeve et al., 1970; Haycock et al., 1978).

  2. 2

    A constant value of resting VCO2 predicted from BMR using the Schofield (1985) equations based on height and weight and assuming a respiratory quotient of 0.85 and therefore the energy equivalent of CO2 is 23.76 kJ l−1 (IDECG, 1990).

  3. 3

    A constant resting value of VCO2 measured using a ventilated hood indirect calorimeter.

  4. 4

    Continuously measured non-resting VCO2 estimated from heart rate using the parameters from the sigmoid function fitted to each individual's calibration data. When VCO2 was estimated from heart rate, a different value was used to calculate PDR in each breath sample. The VCO2 value was the mean within ±10 min of the time of the breath sample.

Cumulative excretion of 13C in breath CO2 (cPDR) was calculated using the trapezoidal rule.

Physical activity during the [13C]MTG breath test

PAL during the [13C]MTG breath test was taken as cPDR calculated using VCO2 estimated from heart rate divided by cPDR calculated using measured resting VCO2 (Method 4 above divided by Method 3). This is mathematically equivalent to non-resting VCO2 divided by resting VCO2, which equates to total energy expenditure (TEE)/resting energy expenditure.

Statistics

Differences between groups were assessed using a Mann–Whitney U-test at the 95% confidence level using standard statistical software (Minitab for Windows, Release 11.2).

Determining the cutoff of the [13C]MTG breath test

The appropriate cutoff for the test using each value of VCO2 as described above was determined using two graph-receiver operator characteristics (TG-ROC; Greiner et al., 1995) at the 95% accuracy level, using a freely available Excel spreadsheet template downloaded from http://city.vetmed.fu-berlin.de/~mgreiner/TG-ROC/tgroc.htm. The spreadsheet contains a test for normality. TG-ROC is a plot of the test sensitivity (Se) and specificity (Sp) against the threshold (cutoff) value assuming the latter to be an independent variable. A cutoff (d0) that realises equal test parameters (Se=Sp=θ0, theta-zero) can be obtained as the intersection point of the two graphs at the chosen level of accuracy (95 or 90%). As the value for θ0 (theta-zero) is below the chosen accuracy level, two cutoff values are selected that represent the bounds of an intermediate range that can be considered as the equivocal range for the clinical interpretation of test results (Greiner et al., 1995). The proportion of the measurement range that gives unambiguous test results is the valid range proportion (VRP). VRP and θ0 (theta-zero) can be used to compare tests, as they do not depend upon the selection of a single cutoff point. The ideal test has both equal to unity and an intermediate range of zero.

The positive and negative predictive values for each MTG test were calculated (Bland, 1995) using a prevalence of pancreatic insufficiency in the study population of 0.21 (five out of 24 subjects, who had CF and had been prescribed PERT). These were compared with the values obtained by Vantrappen et al. (1989) in their validation study of the MTG test in adult patients with pancreatic disease.

Ethical considerations

The study was carried out in accordance with the principles outlined in the Declaration of Helsinki. The work described here posed no risk to the subjects taking part in the study. Subjects were free to drop out of the study at any time. Children with CF were recruited from those attending the Royal Hospital for Sick Children, Yorkhill, Glasgow, but were free of pulmonary infection at the time of the test. Ethical approval was obtained from the Yorkhill Research Ethics Committee for the study involving children and from the University of Glasgow Ethics Committee for Non Clinical Research Involving Human Subjects for the study involving adults. The purpose of the study was carefully explained and written informed consent was obtained from all adult subjects taking part in the study. In the case of children, their written informed consent and also that of their parents was obtained.

Results

The characteristics of the subjects studied are shown in Table 1. There was no significant difference in age, height or weight between healthy control children and children with CF, but the median body mass index standard deviation score was significantly lower in children with CF (P=0.04). One child with CF was not pancreatic insufficient as assessed using the MTG breath test. This child had not been prescribed PERT.

Table 1 Characteristics of subjects studied

The heart rate monitors and calibration procedure were well tolerated by all subjects including children with CF, who were free from pulmonary infection at the time of the test. Children with CF preferred the breath test to alternative non-invasive tests of exocrine pancreatic function such as 3-day faecal fat estimations. There was no problem in sampling breath by blowing through a straw in children with CF.

Figure 1 and Table 2 show the cPDR in 6 h following a test meal containing [13C]MTG. The median (range) PAL in adults, who remained seated as much as possible during the test, was 1.3 (1.0–1.6). The PAL was 1.7 (1.0–2.2) in children who performed the test at home and 1.3 (1.2–1.5) in children who came to the hospital, where there was less choice of activity to occupy them between breath-sampling times.

Figure 1
figure1

Cumulative PDR in 6 h following ingestion of 10 mg kg−1 [13C]MTG. Open symbols are the median cPDR from pancreatic-sufficient subjects (n=19, shown with range). Closed symbols are the median cPDR from children with CF, who have been prescribed PERT (n=5, shown with range). Use of a variable value of non-resting CO2 production rate (VCO2) estimated using individually calibrated heart rate monitors (ICHRM, ♦) improved the specificity of the test compared to using resting VCO2, whether predicted from body surface area ( •) (Shreeve et al., 1970) or BMR ( ) (Schofield, 1985), or measured using a ventilated hood indirect calorimeter (□ ▪). Also shown is the median (x), with range, cPDR from pancreatic-insufficient children with CF calculated using measured resting VCO2 with a constant PAL of 1.3 (n=44, this study plus data from Amarri et al., 1997). - - - - - - Cutoff value of 25.9 determined by TG-ROC (n=69, 24 subjects from this study plus 45 additional children with CF).

Table 2 Cumulative PDR at 6 h during the [13C]MTG breath test

Table 3 shows data from individual children with CF. One patient was taking high-dose PERT. This patient had lowest recovery of 13C in breath CO2, confirming that the patient with poorest pancreatic function had the lowest recovery of 13C in breath, and the patient who was not taking PERT had the highest.

Table 3 Cumulative PDR at 6 h during the [13C]MTG breath test in children with cystic fibrosis

There was no significant difference in cPDR between healthy adults and healthy children when PDR was calculated using VCO2 predicted using the Schofield equations, using resting VCO2 measured by indirect calorimetry and non-resting VCO2 predicted from heart rate. Therefore, data from all healthy subjects, plus the child with CF, who was not taking PERT were combined to determine the reference range (mean±2 s.d.) in pancreatic-sufficient subjects (n=19). The reference range of cPDR in 6 h following ingestion of [13C]MTG was 24–58% when VCO2 was estimated continuously from heart rate. Figure 2 shows the TG-ROC analysis of the cPDR in 6 h, calculated using non-resting VCO2 estimated from heart rate. The cutoff point is the threshold value where the sensitivity is equal to the specificity. The cutoff point was determined using non-parametric statistics, as data from pancreatic-insufficient subjects were not normally distributed. The cutoff point was 25%, with sensitivity and specificity equal to 0.99. There was no intermediate range and therefore the VRP was 1.0. The positive and negative predictive values were 0.99. For comparison, the reference range calculated using a constant value of VCO2 of 5 mmol min−1 m−2 (Shreeve et al., 1970) was 18–43% and that using a constant value of measured resting VCO2 was 17–40%. In both cases, the cutoff point was 20% with sensitivity and specificity=0.83, giving an equivocal range of 17–20% using predicted resting VCO2 and 18–20% using measured resting VCO2. The VRP was 0.92 using predicted resting VCO2 and 0.95 using measured resting VCO2. The positive predictive value (PPV) was 0.67 and the negative predictive value was 0.94.

Figure 2
figure2

TG-ROC analysis to determine the cutoff point in the [13C]MTG breath test. Sensitivity and specificity are shown as a function of the threshold value of cumulative PDR (cPDR) in 6 h, calculated using non-resting CO2 production rate. The cutoff is the point where sensitivity is equal to specificity. In this example, sensitivity and specificity equal 0.99 and the cutoff is 25.4%.

Use of individually calibrated heart rate monitors to estimate VCO2 during the [13C]MTG test improved the PPV of the test from 0.67 using resting VCO2 to 0.99 using non-resting VCO2 estimated from heart rate.

Discussion

In 13C-breath tests, an increase in VCO2 will result in dilution of labelled CO2 from the tracer by unlabelled CO2 leading to low 13C enrichment in breath CO2. VCO2 is usually expressed relative to body surface area to take into account differences with body size. Total CO2 production includes resting VCO2 plus post-prandial or diet-induced thermogenesis (DIT) plus CO2 produced during physical activity. There may also be an increase above BMR owing to infection or inflammation in some subjects. DIT accounts for approximately 10% TEE over a 24-h period. DIT was measured in the adults taking part in this study. VCO2 in the post-prandial period (2 h after the test meal) was 1.16 times resting VCO2 (Slater, 2004). Subjects taking part in 13C-breath tests are instructed to sit quietly in a chair to minimise the increase in VCO2 caused by physical activity. In addition, children with CF may have elevated resting VCO2 compared to healthy children (Amarri et al., 1998). Heart rate monitors were used in preference to motion sensors (Schulz et al., 1989; Montgomery et al., 2004) because motion sensors do not take account of increases in metabolic rate owing to infection or inflammation. It was expected that elevated VCO2 would be accompanied by elevated heart rate in children with CF. However, in this study, children with CF had elevated VCO2 only when lying, not when sitting or standing (Slater, 2004). The most variable component of total CO2 production in children is owing to physical activity.

Use of a resting value of VCO2 will lead to an underestimation of the true amount of tracer recovered in breath, even in compliant adults. Use of non-resting VCO2 to calculate PDR improved the specificity of the 13C[MTG] breath test. The heart rate monitors and calibration procedure were well tolerated by all subjects including children with CF, who were free from pulmonary infection at the time of the test. Children with CF preferred the breath test to alternative non-invasive tests of exocrine pancreatic function such as 3-day faecal fat estimations (personal communication). There was no problem in sampling breath by blowing through a straw in children with CF.

Most of the control children in this study performed the breath test in their own homes, where there was a greater choice of activity than at the hospital. This was performed to demonstrate the problem of physical activity during 13C-breath tests, that is, apparently low recovery of 13C in breath CO2 in some healthy subjects, and that use of non-resting VCO2 could resolve the problem. There was a wider range of physical activity in children (1.0–2.2) than in adults (1.0–1.6), as expected. This leads to bigger errors in PDR calculation when a resting value of VCO2 is assumed in children, than in adults. Children undertaking the test in their own home had a more variable PAL than those coming to the hospital (Table 2). The median PAL in tests that were performed under more clinical conditions at the hospital, with fewer choices of activity, was 1.3, the same as that in compliant adults who remained seated during the test. This PAL is consistent with the observed behaviour and the energy cost of the activities undertaken in terms of metabolic intensity (MET) quoted by Ainsworth et al. (2000), although a detailed activity diary was not recorded. One MET is the ratio of the metabolic rate associated with a particular activity to resting metabolic rate in adults (Ainsworth et al., 1993). MET tables were not available for children's activities.

The problem of physical activity during 13C-breath tests is more obvious in children, than in adults, and in pancreatic-sufficient subjects than pancreatic-insufficient subjects. A 50–100% error in VCO2 (PAL 1.5–2.0) will result in a greater absolute error in healthy subjects than in subjects with poor fat digestion, who have low 13C excretion in breath owing to pancreatic insufficiency. If the mean cPDR in the MTG breath test in healthy subjects were approximately 30% (assuming PAL=1.0), a PAL of 2 would result in a calculated cPDR of 15% in a non-compliant subject, resulting in a false-positive result. In the MTG test, a positive result is low recovery of 13C in breath. A PAL of 2 in a child with CF and pancreatic insufficiency would result in a calculated cPDR of 5% instead of 10%, which would not make a difference to the clinical diagnosis.

van Dijk-van Aalst et al. (2001) used the [13MTG] test to assess lipase activity in healthy children and found significantly lower recovery (after 6 h) of 13C in the breath of teenagers aged 11–17 (range 15–35%), than of children aged 3–10 (range 25–44%). Full-term infants aged 33–121 days had similar recovery (26–44%) as to the adults in this study, when PDR was calculated using a predicted value of resting VCO2 (29–43%), except for one infant with a low recovery of 14%. This could have been owing to poor lipase activity in this child, but it could equally be owing to increased VCO2, owing to physical activity during the test. The study of van Dijk-van Aalst et al. (2001) used different test meals for the infants and older children, and yet another test meal was used in this study. Different test meals may affect the rate of gastric emptying of the meal, but if breath is sampled for sufficient time, it should not affect the area under the curve (cPDR) in healthy subjects, as long as all the meals are composed of ingredients with low 13C natural abundance (Morrison et al., 2000). Kalivianakis et al. (1997) found that the nature of the test meal did not affect recovery of 13C in the breath of adults. We can think of no physiological reason why teenagers would have poorer lipase activity than infants, children and adults, but have shown that increased physical activity can lead to low recovery of 13C in the breath of healthy individuals at any age (Slater, 2004).

The recovery of 13C, when PDR was calculated using resting VCO2, was in the same range as has been reported previously (Amarri et al., 1997; Swart et al., 1997; Wutzke et al., 1999; van Dijk-van Aalst et al., 2001). There was overlap between pancreatic-sufficient and pancreatic-insufficient subjects in the cumulative amount of 13C excreted in breath CO2, no matter whether the resting VCO2 was measured or predicted. However, use of calibrated heart rate monitors to estimate non-resting VCO2 eliminated false-negative results in healthy children. When a variable non-resting value of VCO2 was used, the populations of pancreatic-sufficient and pancreatic-insufficient subjects were completely separated (Figure 1).

The sensitivity and specificity of the [13C]MTG test, when cumulative excretion was calculated using resting VCO2 (86%), was similar to that calculated by Vantrappen et al. (1989) in their validation study (sensitivity 89%, specificity 81%). This suggests that erroneous use of resting VCO2 in the calculation of PDR may have been the cause of the poor PPV of the [13C]MTG test observed by Löser et al. (1998). In this study, the control group had lower cumulative 13C excretion than the group with gastrointestinal diseases of non-pancreatic origin, although the upper limit was similar in both groups. One explanation for this is that patients are more compliant subjects than control subjects, especially students. If some members of the control group failed to remain seated during the test and therefore their energy expenditure and CO2 production rate were not at resting levels throughout, the higher CO2 production rate would result in lower breath 13CO2 enrichment, and therefore the PDR calculated using resting VCO2 would be lower than the actual excretion, resulting in false-positive results and a poor PPV.

Because of the small number of pancreatic-insufficient subjects in this study, data were obtained from a previous study (Amarri et al., 1997) performed under similar conditions to this study. A factor for physical activity (1.3) was applied to the 6 h cPDR, which had been calculated using measured resting VCO2 in children with CF (n=45). The results were compared with the reference range established in this study with no loss of specificity (Figure 1). The cutoff, determined by TG-ROC, was 25.9% compared to the cutoff of 25.4% previously determined. Calibrated heart rate monitors can be used to establish VCO2 and PAL under the conditions that usually occur at each centre and to establish the reference range of cPDR in healthy individuals under these conditions. The PAL measured under the conditions of the test can then be used to correct PDR for non-resting VCO2. This is a more practical approach for routine use of [13C]MTG breath tests to assess the adequacy of PERT, and conversion to pancreatic insufficiency in children with CF, as it is not possible to routinely perform a heart rate calibration during the test.

Use of non-resting VCO2 could lead to improvements in other 13C-breath tests that report results in terms of PDR, especially in children. These include tests of starch digestion and fermentation (Dewit et al., 1992; Weaver et al., 1995; Christian et al., 2003) and liver function tests (Parker et al., 1994, 1997, 1998).

In conclusion, use of non-resting VCO2 improved the specificity and therefore the PPV of the [13C]MTG breath test, confirming our hypothesis that the poor specificity of the breath test is due, in part, to inappropriate use of resting VCO2 to calculate PDR in breath. Calibrated heart rate monitors can be used to determine the reference range for cPDR during 13C-breath tests in healthy children, and the PAL during the [13C]MTG breath test as it is performed at each clinical centre so that cPDR in children with CF can be corrected for non-resting VCO2.

References

  1. Ainsworth BE, Haskell WL, Leon AS, Jacobs Jr DR, Montoye HJ, Sallis JF et al. (1993). Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc 25, 71–80.

  2. Ainsworth BE, Haskell WL, Whitt MC, Irwin MA, Swartz AM, Strath SJ et al. (2000). Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 32, S498–S516.

  3. Amarri S, Coward WA, Harding M, Weaver LT (1998). Importance of measuring CO2-production rate when using 13C-breath tests to measure fat digestion. Br J Nutr 79, 541–545.

  4. Amarri S, Harding M, Coward WA, Evans TJ, Weaver LT (1997). 13Carbon mixed triglyceride breath test and pancreatic enzyme supplementation in CF. Arch Dis Child 76, 349–351.

  5. Amarri S, Weaver LT (1995). 13C-breath tests to measure fat and carbohydrate digestion in clinical practice. Clin Nutr 14, 149–154.

  6. Bland M (1995). An Introduction to Medical Statistics. Oxford University Press: Oxford, p 275.

  7. Christian MT, Edwards CA, Preston T, Johnston L, Varley R, Weaver LT (2003). Starch fermentation by faecal bacteria of infants, toddlers and adults: importance for energy salvage. Eur J Clin Nutr 57, 1486–1491.

  8. Cole TJ, Freeman JV, Preece MA (1995). Body mass index reference curves for the UK, 1990. Arch Dis Child 73, 25–29.

  9. Coplen TB (1996). New guidelines for reporting stable hydrogen, carbon, and oxygen isotope-ratio data. Geochim Cosmochim Acta 60, 3359–3360.

  10. de Boeck K, Delbeke I, Eggermont E, Veereman-Wauters G, Ghoos Y (1998). Lipid digestion in cystic fibrosis: comparison of conventional and high lipase enzyme therapy using the mixed-triglyceride breath test. J Pediatr Gastroenterol Nutr 26, 408–411.

  11. Department of Health (1991). Dietary Reference Values for Food Energy and Nutrients for the United Kingdom. Report of the Panel on Dietary Reference Values of the Committee on Medical Aspects of Food Policy, HMSO: London.

  12. Dewit O, Prentice A, Coward WA, Weaver LT (1992). Starch digestion in young children with cystic fibrosis measured using a 13C breath test. Pediatr Res 32, 45–49.

  13. Ghoos YF, Vantrappen G, Rutgeerts PJ, Schurmans PC (1981). A mixed-triglyceride breath test for intraluminal fat digestive activity. Digestion 22, 239–247.

  14. Greiner M, Sohr D, Göbel P (1995). A modified ROC analysis for the selection of cut-off values and the definition of intermediate results of serodiagnostic tests. J Immunol Methods 185, 123–132.

  15. Haycock G, Schwartz G, Wisotsky G (1978). Geometric method for measuring body surface area: a height–weight formula validated in infants, children and adults. J Pediatr 93, 62–66.

  16. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT (1991). McCance and Widdowson's, The Composition of Foods, 5th edn. The Royal Society of Chemistry: Cambridge.

  17. IDECG (1990). Converting carbon dioxide production to energy expenditure. In: Prentice AM (ed). Report by the International Dietary Energy Consultancy Group (IDECG). NAHRES-4. The Doubly-Labelled Water Method for Measuring Energy Expenditure: Technical Recommendations for human use. IAEA: Vienna, p 201.

  18. Kalivianakis M, Verkade HJ, Stellaard F, van der Werf M, Elzinga H, Vonk RJ (1997). The 13C-mixed triglyceride breath test in healthy adults: determinants of the 13CO2 response. Eur J Clin Invest 27, 434–442.

  19. Löser C, Brauer C, Aygen S, Hennemann O, Fölsch UR (1998). Comparative clinical evaluation of the 13C-mixed triglyceride breath test as an indirect pancreatic function test. Scand J Gastroenterol 33, 327–334.

  20. Montgomery C, Reilly JJ, Jackson DM, Kelly LA, Slater C, Paton JY et al. (2004). Relation between physical activity and energy expenditure in a representative sample of young children. Am J Clin Nutr 80, 591–596.

  21. Morrison DJ, Dodson B, Slater C, Preston T (2000). 13C natural abundance in the British diet: implications for 13C breath tests. Rapid Commun Mass Spectrom 14, 1321–1324.

  22. Motulsky HJ, Ransnas LA (1987). Fitting curves to data using nonlinear regression: a practical and nonmathematical review. FASEB J 1, 365–374.

  23. Parker AC, Preston T, Heaf D, Kitteringham NR, Choonara I (1994). Inhibition of caffeine metabolism by ciprofloxacin in children with cystic fibrosis as measured by the caffeine breath test. Br J Clin Pharmacol 38, 573–576.

  24. Parker AC, Pritchard P, Preston T, Choonara I (1998). Induction of CYP1A2 activity by carbamezepine in children using the caffeine breath test. Br J Clin Pharmacol 45, 176–178.

  25. Parker AC, Pritchard P, Preston T, Dalzell AM, Choonara I (1997). Lack of inhibitory effect of cimetidine on caffeine metabolism in children using the caffeine breath test. Br J Clin Pharmacol 45, 467–470.

  26. Prosser SJ, Brookes ST, Linton A, Preston T (1991). Rapid, automated analysis of 13C and 18O of CO2 in gas samples by continuous-flow, isotope ratio mass spectrometry. Biol Mass Spectrom 20, 724–730.

  27. Schoeller DA, Klein PD, Watkins JB, Heim T, MacLean WCJ (1980). 13C abundance of nutrients and the effect of variations in 13C isotopic abundances of test meals formulated for 13CO2 breath tests. Am J Clin Nutr 33, 2375–2385.

  28. Schoenheimer R, Rittenberg D (1939). Studies in protein metabolism I. General considerations in the application of isotopes to the study of protein metabolism. The normal abundance of nitrogen isotopes in amino acids. J Biol Chem 127, 285–289.

  29. Schofield WN (1985). Predicting basal metabolic rate, new standards and review of previous work. Hum Nutr: Clin Nutr 39C, 5–41.

  30. Schulz S, Westerterp KR, Bruck K (1989). Comparison of energy expenditure by the doubly labeled water technique with energy intake, heart rate, and activity recording in man. Am J Clin Nutr 49, 1146–1154.

  31. Shreeve VW, Cerasi E, Luft R (1970). Metabolism of [2-14C]pyruvate in normal, acromegalic and HGH-treated human subjects. Acta Endocrinol 65, 155–169.

  32. Slater C (2004). Improvements to the [13C]MTG breath test for measuring fat digestion. PhD thesis, University of Glasgow, https://dspace.gla.ac.uk/handle/1905/449, pp 97, 117, 144, 159, 161, 162.

  33. Slater C, Ling SC, Preston T, Weaver LT (2002a). Bulk and compound specific analysis of stool lipid confirm that the ‘missing’ 13C in mixed triacylglycerol breath test is not in the stool. Food Nutr Bul 23 (Suppl), 48–52.

  34. Slater C, Preston T, Morrison DJ, Weaver LT (2002b). A shelf-stable, palatable test meal suitable for use with hydrophilic and lipophilic tracers in 13C breath tests. Proc Nutr Soc 61, 66A (abstract).

  35. Slater C, Preston T, Weaver LT (2001). Stable isotopes and the International System of Units. Rapid Commun Mass Spectrom 15, 1270–1273.

  36. Slater C, Preston T, Weaver LT (2004). Is there an advantage in normalising the results of the Helicobacter pylori [13C]urea breath test for CO2 production rate in children? Isotopes Environ Health Stud 40, 89–98.

  37. Slater C, Preston T, Weaver LT (2006). Comparison of accuracy and precision of heart rate calibration methods to estimate CO2 production during breath tests. Eur J Clin Nutr 60, 69–76.

  38. Swart GR, Baartman EA, Wattimena JL, Rietveld T, Overbeek SE, Van den Berg JW (1997). Evaluation studies of the 13C-mixed triglyceride breath test in healthy controls and adult cystic fibrosis patients with exocrine pancreatic insufficiency. Digestion 58, 415–420.

  39. van Dijk-van Aalst K, Van Den Driessche M, van der Schoor S, Schiffelers S, van't Westeinde T, Ghoos Y et al. (2001). 13C mixed triglyceride breath test: a noninvasive method to assess lipase activity in children. J Pediatr Gastroenterol Nutr 32, 579–585.

  40. Vantrappen GR, Rutgeerts PJ, Ghoos YF, Schurmans PC (1989). Mixed triglyceride breath test: a noninvasive test of pancreatic lipase activity in the duodenum. Gastroenterology 96, 1126–1134.

  41. Walsh S, Diamond D (1995). Non-linear curve fitting using Microsoft Excel Solver. Talanta 42, 561–572.

  42. Weaver LT, Dibba B, Sonko B, Bohane TD, Hoare S (1995). Measurement of starch digestion of naturally 13C-enriched weaning foods, before and after partial digestion with amylase-rich flour, using a 13C breath test. Br J Nutr 74, 531–537.

  43. Wutzke KD, Radke M, Breuel K, Gurk S, Lafrenz J-D, Heine WE (1999). Triglyceride oxidation in cystic fibrosis: a comparison between different 13C-labeled tracer substances. J Pediatr Gastroenterol Nutr 29, 148–154.

Download references

Acknowledgements

We acknowledge financial support from the UK Medical Research Council Joint Research Equipment Initiative and the University of Glasgow, and thank Cambridge Isotopes Laboratory Inc, Andover, MA, USA for donating the [13C]MTG used in this study. We thank Dr Simon Ling, formerly consultant gastroenterologist, Royal Hospital for Sick Children (RHSC), Glasgow, for assistance in recruiting children with CF.

Author information

Correspondence to C Slater.

Additional information

Guarantor: C Slater.

Contributors: CS performed the study, analysed the data, wrote the first draft and refined the manuscript. TP conceived the original idea, advised on study design and data analysis and critically appraised the manuscript. LTW cosupervised the project and critically appraised the manuscript.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Slater, C., Preston, T. & Weaver, L. Improving the specificity of the [13C]mixed triacylglycerol breath test by estimating carbon dioxide production from heart rate. Eur J Clin Nutr 60, 1245–1252 (2006). https://doi.org/10.1038/sj.ejcn.1602444

Download citation

Keywords

  • breath test
  • pulmonary gas exchange
  • cystic fibrosis
  • mixed triacylglycerol breath test

Further reading