44 Arch Dis Child 1998;79:44–47 Metabolic handling of 13C labelled tripalmitin in Arch Dis Child: first published as 10.1136/adc.79.1.44 on 1 July 1998. Downloaded from healthy controls and patients with cystic fibrosis

J L Murphy, K M Laiho, A E Jones, S A Wootton

Abstract uring total excretion of the 13C label in stools.3 Aim—To examine the gastrointestinal Despite raised total stool lipid losses, more of handling and metabolic disposal of emul- the 13C labelled palmitic acid was absorbed in sified [1-13C]palmitic acid esterified into a this group of patients with cystic fibrosis com- triglyceride in nine healthy children and pared with healthy children (94% v 76%, seven patients with cystic fibrosis on respectively), suggesting that there is not a spe- enzyme replacement treatment. cific malabsorptive defect of labelled palmitic Methods—After an overnight fast, each acid in cystic fibrosis when ingested as the free child was given 10 mg/kg body weight acid. We also showed that palmitic acid was the [1,1,1-13C]tripalmitin with a standardised major stool fatty acid in those stools with the test meal of low natural 13C abundance. highest 13C enrichment and that, primarily, the The total enrichment of 13C was measured 13C label was restricted to the species con- using isotope ratio mass spectrometry in sumed by the subjects (that is, palmitic acid). stool collected for a period of up to five Using diVerent 13C labelled triglycerides and days and in breath samples collected over fatty acids, maldigestion and malabsorption a 24 hour period. have been assessed by measuring the excretion 4–6 Results—The mean proportion of admin- of label on the breath alone. Such studies istered 13C label excreted in stool was 6% have not directly determined the excretion of (range, 1–12.7%) in healthy children and label in the stool, assuming that diVerences in 24.6% (range, 0–64%) in patients with the excretion of label on the breath are cystic fibrosis. Healthy children excreted determined wholly by diVerences in digestion 31.3% of the administered label on their within the gastrointestinal tract. A further breath (range, 14.2–42.9%). Correcting theoretical limitation of using breath tests is the the excretion of administered 13C label on implicit assumption that alterations in lipid the breath for diVerences in digestion and maldigestion and/or malabsorption do not 7 absorption in patients with cystic fibrosis aVect fatty acid oxidation. We are not aware of increased the diVerence between indi- any studies in cystic fibrosis that have meas- http://adc.bmj.com/ viduals from 0–31.3% of administered ured directly the extent of digestion and dose (mean, 17.9%) to 0–49.1% of ab- absorption in combination with “breath tests” 13 sorbed dose (mean, 23.2%) and was poorly using C labelled lipids. To address this issue, related to the amount of 13C label in stool. we have examined both the gastrointestinal handling and the subsequent metabolism of Conclusion—Measurements of breath 13 13 emulsified [1- C]palmitic acid esterified into a CO2 do not consistently reflect the gas- trointestinal handling of emulsified 13C triglyceride in healthy children and patients with cystic fibrosis by measuring excretion of labelled tripalmitin because of diVerences on September 29, 2021 by guest. Protected copyright. the label in stools and on the breath as 13CO . in digestion and absorption in cystic 2 fibrosis. Further studies need to examine whether “breath tests” alone can predict Subjects and methods with confidence the gastrointestinal han- 13 Seven patients with cystic fibrosis (three boys dling of other C labelled triglycerides and four girls) aged 4–11 years (mean, 8.0 and fatty acids. years) from the cystic fibrosis clinic at South- (Arch Dis Child 1998;79:44–47) ampton University Hospitals NHS Trust were Keywords: fatty acids; absorption; metabolism; stable studied while on their normal habitual pancre- Institute of Human atic enzyme replacement treatment. Patients Nutrition, University isotopes; cystic fibrosis of Southampton, with small bowel resection or other known fac- Southampton SO16 tors that might limit absorption (for example 6YD, UK Measurements of total stool lipid made during Crohn’s disease) were excluded from the study. J L Murphy gross balance studies provide a gross measure The cystic fibrosis patients were taking be- K M Laiho of the availability of dietary lipid in cystic tween 13 462 and 34 000 IU lipase/kg body A E Jones 1 S A Wootton fibrosis. Modern developments in mass spec- weight/day (mean, 18 877); an enzyme dosage trometry in combination with stable isotopes achieved by self-titration against gastrointesti- Correspondence to: provide the opportunity to examine the func- nal symptoms and bowel habit. No attempt was Dr J L Murphy, Clinical tional capacity of the gastrointestinal tract to made to alter or intervene with the manage- Nutrition and Metabolism Unit, Mailpoint 113, Level C handle specific triglycerides found in the diet. ment of the pancreatic enzyme replacement West Wing, Southampton Because palmitic acid is the predominant satu- treatment. Nine healthy children (six boys and General Hospital, Tremona rated fatty acid in the UK diet,2 previously we three girls) aged 5–8 years (mean, 7.2 years) Road, Southampton SO16 examined the absorptive capacity of the gut from local schools also participated in the 6YD, UK. using 13C labelled palmitic acid in healthy chil- study. Informed consent was obtained from all Accepted 5 February 1998 dren and patients with cystic fibrosis by meas- of the subjects and the study protocol was Metabolic handling of 13C labelled tripalmitin in cystic fibrosis 45

approved by the Ethical Committee of South- before and during the stool collection period.

ampton and South West Hampshire Health None of the animal source protein foods eaten Arch Dis Child: first published as 10.1136/adc.79.1.44 on 1 July 1998. Downloaded from Commission. (such as eggs and meat) were naturally After an overnight fast, the [1,1,1- enriched with 13C. 13C]tripalmitin was administered orally to the subjects at a dose of 10 mg/kg body weight (99 BREATH AND STOOL ANALYSES atom % excess; Masstrace, Woburn, USA) as The methodology for processing stools has part of a controlled standard test meal consist- been described previously.8 Enrichment of 13C 13 ing of 120 g white bread, 100 g orange juice, in the stool and as CO2 on the breath was and 10 g butter of low 13C abundance analysed by continuous flow isotope ratio mass (−25.5‰).8 The patients with cystic fibrosis spectrometry (ANCA-NT GSL; Europa Sci- took the same amount of enzymes with the test entific Ltd). Total stool lipid was extracted meal that they would usually take with a snack from stools by a modification of the method of meal and this ranged from 1458 to 6849 IU Folch and colleagues.11 The 13C enrichment of 13 lipase/kg body weight/day (mean, 3750). The samples of breath CO2 and stool was 13C labelled tripalmitin was given as a glucose- expressed as the “per ml relative diVerence sucrose-casein emulsion, using a modification from the reference standard Pee Dee Belemnite 9 13 12 of Emken et al, because emulsification before (PDB)” as defined by Craig ( CPDB,‰). The administration overcomes the problems associ- proportion of administered 13C label in the ated with the physicochemical properties of the stools was calculated according to the formula crystalline form of this label, which might oth- presented by Schoeller and colleagues.13 The erwise result in poor digestion and absorption apparent absorption of the 13C label was deter- if given in a free, unemulsified form.10 mined from the diVerence between the amount Breath samples were collected using breath of label administered and that excreted in the collection bags (Quintron) before consuming stool. the test meal to provide a measure of baseline The proportion of 13C label excreted on the 13 13 C excretion on the breath, and were then col- breath as CO2 was expressed as a percentage lected at hourly intervals for a period of at least of absorbed 13C label each hour and as the six hours and then after eight hours, 10 hours, cumulative percentage dose excreted over 24 and 24 hours. Specimen breath samples were hours, according to the formula presented by transferred into evacuated gas sample contain- Watkins et al.4 ers (Exetainers; Isochem, Finchampstead, UK) for analysis in duplicate. All subjects were STATISTICS rested for the duration of the tests and were The results are reported as mean (SD). Statis- given constant supervision at the Clinical tical comparisons between the data were Nutrition and Metabolism Unit at Southamp- performed using the unpaired Student’s t test. ton General Hospital. No additional food or DiVerences between means were considered

drink was permitted during the initial six hour significant when p < 0.05. Associations be- http://adc.bmj.com/ period except for bottled mineral water. After tween variables were tested by the Pearson six hours, the subjects ate a meal without any product moment correlation coeYcient (R). foods naturally enriched for 13C. Whole body

CO2 excretion was measured by indirect calor- Results imetry (GEM; Europa Scientific Ltd, Crewe, In healthy children, 6.0% (3.7%) of the UK) for a period of 15 minutes at the same administered 13C label was excreted in the stool time points as breath sampling over the first six (range, 1–12.7%) (table 1). In contrast, hours to calculate the amount of label excreted patients with cystic fibrosis excreted more 13C on September 29, 2021 by guest. Protected copyright.

each hour. It was not possible to measure CO2 label in the stool: 24.6% (22.3%) of the excretion at eight,10, and 24 hours. The value administered 13C label was found in the stool,

for CO2 excretion at six hours was used to cal- although there was a wide range (range, culate the amount of label excreted at eight and 0–64%; p > 0.05; table 1). Usually, the stools 10 hours. The proportion of label excreted on with the greatest 13C enrichment were found on the breath at 24 hours was calculated using the the day after the test meal and enrichment

baseline value for CO2 excretion. returned to baseline values by day 3 or 4. In A stool sample was collected on the day patients with cystic fibrosis, excretion of total before the labelled test meal to measure lipid in the stool was four times greater than baseline 13C excretion. Thereafter, all stools that seen in healthy children (13.2 g/day (3.1) v passed were collected and processed individu- 2.95 g/day (1.0); p < 0.001). Analysing both ally for a period of up to five days. All the stool groups together, there was only a weak associ- collections were carried out at home and ation between the amount of 13C label and the monitored daily throughout the study period. total lipid in the stool (R = 0.53; p > 0.09). All of the subjects and parents of children were In both groups, excretion of the label on the 13 cooperative, willing, and capable of following breath as CO2 peaked between one and six the details of the procedure. At no time was hours after the test meal, returning to baseline diYculty encountered nor was there any need values in almost all subjects by 24 hours. In one for modification of the procedure to accommo- of the healthy children and one cystic fibrosis date any of the subjects. All the stool patient, 13C enrichment on the breath was only collections were complete. The subjects were 2 SD units greater by 24 hours than that also instructed to avoid eating naturally 13C observed at baseline. In healthy children, the enriched foods (such as corn or maize based proportion of administered 13C label excreted products and cane sugar) for at least two days on the breath was 31.3% (10.6%) (range, 46 Murphy, Laiho, Jones, Wooton

13 13 Table 1 Excretion of C in stool and on breath as CO2 in healthy children and patients could have been malabsorbed because of a fail- with cystic fibrosis ure in the absorptive capacity of the gastroin- Arch Dis Child: first published as 10.1136/adc.79.1.44 on 1 July 1998. Downloaded from

13 13 13 testinal tract. These observations are in con- Stool C excretion Breath CO2 Breath CO2 Subject (% admin dose) (% admin dose) (% absorbed dose) trast to those shown previously by our group using [1-13C]palmitic acid presented as the free Healthy children 1 6.6 31.4 33.7 acid, where more of the label was absorbed in 2 8.1 14.2 15.5 patients with cystic fibrosis compared with 3 1.0 42.9 43.3 healthy children. It might be that these 4 4.8 40.9 42.9 5 2.1 24.0 24.5 diVerences reflect the physicochemical proper- 6 4.8 41.7 42.9 ties of the crystalline form of saturated fatty 7 9.5 18.1 20.0 8 12.7 37.8 43.3 acids, which appear to favour enhanced 9 4.4 30.6 32.0 absorption in patients with cystic fibrosis com- Mean (SD) 6.0 (3.7) 31.3 (10.6) 33.1 (10.9) pared with healthy subjects when the fatty acid Cystic fibrosis is given as the free acid. 1 64.4 0 0 2 10.0 17.8 20.4 A previous study has shown that doubling 13 3 36.5 31.3 49.1 the dose of enzymes could increase C 4 0 16.5 16.5 recovery on the breath after oral administration 5 29.3 23.7 33.6 6 27.8 16.5 22.8 of a mixed triglyceride, which is assumed to 7 4.2 19.4 20.3 reflect a reduction in excretion of the label in Mean (SD) 24.6 (22.3) 17.9 (9.5) 23.2 (15.2) the stool.14 In our study, the habitual dose of Significance p < 0.05 p < 0.05 p > 0.05 enzymes was taken with the test meal, a dose that might not have been optimal for this 14.2–42.9%), whereas patients with cystic 13 particular meal, and this could have contrib- fibrosis excreted significantly less C label on uted to poor digestion of the labelled tripalmi- the breath (17.9% (9.5%); range, 0–31.3%; tin. Owing to concern over the adverse eVects p < 0.05). In healthy children, the proportion 13 of high doses of pancreatic enzyme on the large of C label excreted on the breath was not 15 13 intestine, any increase in the intake of pancre- altered if expressed directly as C label admin- atic enzymes would be supported by further istered rather than as label made available studies examining the nature of species bearing through digestion and absorption (33.1% the 13C label to discriminate between maldiges- (10.9%); range 15.5–43.3%; table 1). How- tion and malabsorption. The appearance of the ever, correcting excretion of the label on the label in the stool predominantly as a trigly- breath for diVerences in digestion and absorp- ceride would indicate a failure of digestion tion in patients with cystic fibrosis increased (impaired lipase activity arising from pancre- the diVerences seen between individuals atic insuYciency, or solubilisation, or poor (23.2% (15.2%) of absorbed label; range, pancreatic enzyme replacement treatment 0–49.1%; p > 0.05). There was no association 13 management), while appearance of the label as

between the amount of C label in the stool http://adc.bmj.com/ 13 free fatty acid, 2-monoacylglycerol, or as soaps and proportion of absorbed C label excreted would indicate a failure in absorption of on the breath (R = −0.15; p > 0.75). hydrolysed trigylceride. Our study revealed that 13C labelled tripalmi- Discussion tin was not oxidised to the same extent in both Our study aimed to examine both the gastro- groups, whether it was expressed as a intestinal handling and metabolism of emulsi- proportion of the 13C label administered or the fied [1-13C]palmitic acid esterified into a proportion of that absorbed. This could be triglyceride in healthy children and patients examined in two ways. First, the value of char- on September 29, 2021 by guest. Protected copyright. with cystic fibrosis by measuring excretion of acterising the extent of digestion and absorp- the label in the stool and on the breath as tion in cystic fibrosis using emulsified 13C 13 13 CO2. This is the first time that C labelled labelled tripalmitin in breath tests needs to be tripalmitin, orally administered as an emulsion, considered. While patients with cystic fibrosis has been used to measure the digestion and excreted more 13C label in their stools, recovery absorption of dietary triglycerides in healthy of the label on their breath was lower than in children and patients with cystic fibrosis on healthy children, although there was some habitual pancreatic enzyme replacement treat- overlap between the two groups. For example, ment by measuring directly the excretion of 13C in patient 1 (table 1), although excretion of the label in the stool. In healthy children, only a 13C label in the stool was raised, no label was small proportion of the 13C label was excreted detected in the breath; other patients not only in the stool, which is consistent with total stool had raised stool losses but also a greater recov- lipid losses within the normal range of ery of the label on the breath, comparable to < 5 g/day.1 Almost a quarter of the 13C label healthy children (patients 3 and 5; table 1). appeared in stools from patients with cystic This suggests that in some individuals emulsi- fibrosis, with as much as 64% in one child, and fied 13C labelled tripalmitin does not suY- there were substantially greater amounts of ciently reflect the processes involved in the lipid in the stool when compared with healthy digestion and absorption of dietary long chain children. The diVerences seen in excretion of saturated triglycerides. However, it should be the 13C label in the stool between the patients noted that measurements of 13C label in the with cystic fibrosis could be a result of poor stool have not been reported for other 13C digestion of the labelled triglyceride, which labelled mixed triglycerides and medium chain limits subsequent absorption. Alternatively, or triglycerides used in breath tests. The other in addition, the labelled products of digestion consideration is the extent of recovery of the Metabolic handling of 13C labelled tripalmitin in cystic fibrosis 47

absorbed label. In both groups, there was a functional capacity of the gastrointestinal tract

large variability between subjects in the to handle saturated fatty acids when esterified Arch Dis Child: first published as 10.1136/adc.79.1.44 on 1 July 1998. Downloaded from proportion of absorbed 13C label recovered on into triglycerides, despite pancreatic enzyme the breath. Having corrected for diVerences in replacement treatment. It should not be digestion and absorption, it is not clear what assumed that excretion of the label on the other factors might account for this observa- breath reflects the gastrointestinal handling of tion. The metabolic competence of an indi- emulsified 13C labelled tripalmitin. Further vidual influenced by either genetic, nutritional, studies are needed to examine whether breath or lifestyle factors would be important determi- tests alone can predict with confidence the nants of the partitioning and subsequent extent of digestion and absorption in cystic oxidation of 13C labelled fatty acids. Other fac- fibrosis of other 13C labelled triglycerides and tors such as physical activity might also fatty acids. 13 influence the extent of breath CO2 recovery, particularly when performed throughout the We thank the children and parents whose cooperation was 16 essential for the completion of this study and Mrs A Hounslow, postprandial period, although it is unlikely Mrs J Gavin, Dr C J Rolles, and Dr G Connett for their assist- that they contributed much towards the diVer- ance. The support of The Cystic Fibrosis Trust, Scientific Hos- pital Supplies UK Ltd, and the Wessex Medical Trust is grate- ences between the subjects in our study. All fully acknowledged. subjects were measured under similar resting conditions and avoided eating and drinking for 1 Murphy JL, Wootton SA, Bond SA, Jackson AA. Energy at least 12 hours before label administration content of stools in normal healthy controls and patients with cystic fibrosis. Arch Dis Child 1991;66:495–500. and for six hours during the period of the 2 Gregory J, Foster K, Tyler H, Wiseman M. The dietary and nutritional status of British adults. London: HMSO, breath test. The subjects also avoided any foods 1990:87–122. naturally enriched for 13C over the period of the 3 Murphy JL, Jones AE, Stolinski M, Wootton SA. Gastroin- study. testinal handling of [1-13C]palmitic acid in healthy controls and patients with cystic fibrosis. Arch Dis Child 1997;76: Previous studies have shown that not all of 425–7. 4 Watkins JB, Klein PD, Schoeller DA, Kirschner BS, Park R, the CO2 generated from the label would have Perman JA. Diagnosis and diVerentiation of fat malabsorp- 13 13 been excreted as breath CO2. A proportion of tion in children using C-labeled lipids; trioctanoin, 13 triolein, and palmitic acid breath tests. Gastroenterology the CO2 formed from the oxidation of labelled 1982;82:911–17. substrate is not excreted directly on the breath 5 Vantrappen GR, Rutgeerts PJ, Ghoos YF, Hiele MI. Mixed but might be retained in the body pools or triglyceride breath test: a noninvasive test of pancreatic lipase activity in the duodenum. Gastroenterology 1989;96: excreted via the urine, skin,17 or stool.18 1126–34. 6 McClean P, Harding M, Coward WA, Green MR, Weaver Previous studies have either presented uncor- LT. Measurement of fat digestion in early life using a stable 4–6 14 rected data or have used correction factors isotope breath test. Arch Dis Child 1993;69:366–70. 7 Klein S, Wolfe RR. The use of isotopic tracers in studying to predict more accurately the extent of oxida- lipid metabolism in human subjects. 819 Clinical endocrinology tion of labelled substrates. Precisely how and metabolism. International practice and research. Techniques 13 for metabolic investigation in man. London: Balliere Tindall much of the CO2 is retained in the body pools 1987;4:797–816. remains unresolved, with estimates ranging 8 Murphy JL, Jones A, Brookes S, Wootton SA. The gastroin- 13 from 10% to 50% in normal individuals.20 testinal handling and metabolism of [1- C]palmitic acid in http://adc.bmj.com/ healthy women. Lipids 1995;30:291–8. Because there is no specific information on this 9 Emken EO, Adlof RO, Rohwedder WK, Gulley RM. Influ- ence of linoleic acid on desaturation and uptake of issue in patients with cystic fibrosis, we have deuterium-labeled palmitic and stearic acids in humans. not attempted to correct the oxidation data. It Biochim Biophys Acta 1993;1170:173–81. 13 10 Jones AE, Smith RD, Nurmi JA, Hameen-Antilla P, Murphy is possible that the C labelled products of JL, Wootton SA. The gastrointestinal handling and maldigestion and malabsorption provide a postprandial metabolism of [1-13C]palmitic and [1-13C]oleic acids in healthy women. Proc Nutr Soc substrate for colonic bacterial metabolism. 1997;56:197A. However, little is known about the extent to 11 Folch JL, Lee M, Sloane Stanley GH. A simple method for which fermentation of fatty acids by colonic the isolation and purification of total lipids from animal tis- on September 29, 2021 by guest. Protected copyright. sues. J Biol Chem 1957;226:497–509. bacteria influences measurements of 13C recov- 12 Craig H. Isotopic standards for carbon and oxygen and cor- ery on the breath. Fermentation of these rective factors for mass-spectrometric analysis of carbon dioxide. Geochim Cosmochim Acta 1957;12:133–49. labelled products might result in the excretion 13 Schoeller DA, Klein PD, Van Santen E. Fecal 13C analysis for the detection and quantitation of intestinal malabsorp- of the label as flatus gases (such as methane or tion. Limits of detection and application to disorders of carbon dioxide), which would result in an intestinal cholylglycine metabolism. J Lab Clin Med 1981;97:439–48. underestimation of oxidation. Alternatively, 14 Amarri S, Harding M, Coward WA, Evans TJ, Weaver LT. any such 13C labelled carbon dioxide from fer- 13Carbon mixed triglyceride breath test and pancreatic enzyme supplementation in cystic fibrosis. Arch Dis Child mentation might be absorbed across the 1997;76:349–51. colonic mucosa and could contribute to the 15 Smyth RL, van Velzen D, Smyth AR, Lloyd DA, Heaf DP. amount of 13CO excreted on the breath. The Strictures of the ascending colon in cystic fibrosis and 2 high-strength pancreatic enzymes. Lancet 1994;343:85–6. potential contribution made by colonic bacte- 16 Kalivianakis M, Verkade HJ, Stellaard F, Van der Werf M, 13 Elzinga H, Vonk RJ. The 13C-mixed triglyceride breath test rial fermentation of C labelled lipids to diVer- in healthy adults: determinants of the 13CO response. 13 2 Eur ences in the recovery of C on the breath J Clin Invest 1997;27:434–42. 17 Elia M, Fuller NJ, Murgatroyd PR. Measurement of requires further study. bicarbonate turnover in humans: applicability to estimation In conclusion, after oral administration of of energy expenditure. Am J Physiol 1992;263:E676–87. 13 18 Caprilli R, Frieri G, Latella G, Vernia P, Santoro ML. Fae- emulsified C tripalmitin, patients with cystic cal excretion of bicarbonate in ulcerative colitis. 13 Digestion fibrosis excreted more of the C label in the 1986;35:136–42. 19 Jones PJH, Pencharz PB, Clandinin MT. Whole body oxida- stool when compared with healthy children, tion of dietary fatty acids: implications for energy although a wide diVerence in excretion was utilization. Am J Clin Nutr 1985;42:769–77. 20 Irving CS, Wong WW, Schuman RJ, O’Brian Smith E, Klein observed between these patients. This suggests PD. 13C bicarbonate kinetics in humans: intra- verses inter- that in cystic fibrosis there are diVerences in the individual variations. Am J Physiol 1983;245:R190–202.