Professional Documents
Culture Documents
Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu
Original article
a r t i c l e i n f o s u m m a r y
Article history: Background & aims: Xylitol has been approved for parenteral nutrition and may be beneficial in catabolic
Received 11 April 2013 situations. The aim was to establish an easy method to monitor xylitol serum levels in patients receiving
Accepted 29 June 2013 xylitol and to determine whether xylitol is safe.
Methods: A commercially available xylitol test was validated and used to measure serum levels in
Keywords: 55 patients admitted to our intensive care unit with an indication for parenteral nutrition with xylitol for
Critically ill
at least 24 h. Controls consisted of the most recent 56 patients admitted to the intensive care unit who
Sepsis
received parenteral nutrition without xylitol for at least 2 days. Xylitol serum levels were determined
Antioxidants
Parenteral nutrition
using the test. Adverse events, liver enzymes, lactate, bilirubin, g-glutamyl transpeptidase, and insulin
Nutrient requirement were secondary endpoints.
Early feeding Results: Patients receiving xylitol received 32.6% less insulin than controls. The amount of energy they
received was comparable (xylitol: 810.1; controls: 789.8 kcal). Mean liver enzymes and lactate levels
were similar in both groups. Adverse events considered attributable to xylitol did not occur. Xylitol did
not accumulate in patients’ blood and returned to near baseline values one day after parenteral nutrition
was stopped.
Conclusions: Parenteral nutrition with xylitol appears to be safe for critical care patients. There were no
signs of hepatoxicity.
Trial registration DRKS: DRKS00004238.
Ó 2013 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism.
0261-5614/$ e see front matter Ó 2013 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism.
http://dx.doi.org/10.1016/j.clnu.2013.06.018
484 A.S. Schneider et al. / Clinical Nutrition 33 (2014) 483e488
expensive, and requires special expertise.10 Xylitol test kits exist, xylitol, residual serum samples taken as part of routine care were
but they have been established for use in the food industry. They used. Patients had to meet the following inclusion criteria:
have not been adapted to measure xylitol in human serum in admission to the intensive care unit at Hannover Medical School for
clinical situations. a maximum of 3 days before PN and PN containing xylitol for at
Under in vitro conditions, we validated a test to quantify xylitol least 24 h.
in human serum. This test is based on a first reaction in which a The retrospective control group consisted of the most recent 56
dehydrogenase oxidizes xylitol to D-xylulose. In a second reaction patients who had been admitted to the Hannover Medical School
NADH reduces iodonitrotetrazolium chloride with the help of intensive care unit for a maximum of 3 days before receiving PN
diaphorase to a 2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-phenyl-2H- prior to May 31, 2011 and who had received PN without xylitol for
tetrazolium chloride (INT)-formazan compound. The amount of at least 2 days. Until this date PN at the facility did not include
INT-formazan that is formed is proportional to the xylitol content xylitol on a regular basis. From May 31, 2011 on, there was a change
and can be measured at a wavelength of 492 nm using a spectro- of standard hospital procedure and all trauma patients received PN
photometer. The test’s accuracy and reliability was assessed by with xylitol unless indicated otherwise. Patients from the control
adding defined amounts of the sugar substitute to control serum group were admitted between December 7, 2008 and May 30, 2011.
that was definitely xylitol negative. After its validation, the test was Patients receiving PN with xylitol entered the trial between May 31,
used to assess xylitol levels in the serum of patients receiving PN 2011 and February 11, 2012.
with xylitol. To our knowledge, so far no good quality studies have The observation period was day 0 (before PN containing
been published in English that systematically compare PN with and xylitol was applied) up to day 7. According to our nutritional
without xylitol. Only a few older studies of poor quality are avail- standard, patients received PN with xylitol via a multi-chamber
able, mainly in German, which investigate PN with xylitol.5,7e9 The bag (Nutriflex Combi, B. Braun, Melsungen, Germany) plus an
relevance of the evaluations is limited. MCT/LCT lipid composite (Lipofundin 20%, B. Braun) and a
The aim of this case-control study was to determine whether preparation containing water-soluble and lipid-soluble vitamins
this method is reliable and safe under actual clinical conditions. and trace elements as a 24 h infusion via a central venous
Patients admitted to the intensive care unit who had an indication catheter. The ingredients of the solutions are shown in Table 1.
to receive PN with xylitol for at least 24 h were included in the The only difference between the PN of the two patient groups
study. Xylitol levels were measured in patients’ serum. In addition was that the group receiving xylitol (50 g) was given 100 g
to this primary objective, possible adverse events resulting from glucose in a 2-chamber bag while the controls received 150 g in
the application of PN containing xylitol and the subsequent xylitol a 3-chamber bag. Both groups received MCT/LCT lipids. If
serum levels were documented. Liver enzymes (aspartate amino- indicated, patients were administered enteral nutrition at a
transferase (AST), alanine aminotransferase (ALT)), lactate, g-glu- maximum of 500 kcal/d at the earliest on day 4. A maximum of
tamyl transpeptidase (g-GT), bilirubin, and daily insulin 120 g xylitol was given per day. Blood samples were taken once a
requirement were also assessed. day on day 0 up to 7 at about the same time as part of routine
care. Samples were stored at 4 C until they were tested. Xylitol
2. Material and methods serum levels (mg/dL) were determined within 14 d of blood
sampling. If necessary to maintain normoglycemia, patients
2.1. Validation of a test to determine xylitol in human serum received insulin to keep their plasma glucose level at a range of
6.5e9 mmol/L. The maximum insulin dose was set at 8 IU per
A commercially available test kit for xylitol (K-SORB, Mega- hour (concentration: 1 IU per mL). Adjustments of the insulin
zyme Deutschland, Gernsheim, Germany) was used according to dose were based on whole-blood glucose measurements at 2e
the manufacturer’s instructions. Before they were measured, 6 h intervals. Each sample was deproteinized and tested in
samples were deproteinized according to the instructions using duplicate against a standard xylitol solution according to the
Carrez solutions. Defined amounts of xylitol (SigmaeAldrich manufacturer’s instructions (Megazyme). In addition, the amount
Chemie, Munich, Germany) were added to control serum from a of infused xylitol (g/d) was calculated. Secondary parameters
healthy donor who was definitely negative for sugar substitutes. were liver enzymes (AST, ALT), bilirubin, g-GT, and lactate. More
Since measurement errors due to turbidity were possible than twice the initial value was defined as increased. The daily
(depending on the amount of sample in the cuvette), different total insulin requirement and acute kidney failure with the
sample volumes and concentrations were tested to assess the indication for kidney replacement therapy were documented.
test error and to determine the range in which the test is accu- Data were collected using the documentation and archive data-
rate. This volume was then used to measure sera from patients base of the Hannover Medical School (ALIDA, MHH, Hannover
who had received PN with xylitol. Samples were measured at
492 nm using a spectrophotometer (Hitachi U-2000, Mannheim,
Germany).
Table 1
Composition of parenteral nutrition given to each study group.
2.2. Patients and study design
Xylitol group Control group
This was a case-control study (registered at DRKS: NutriflexÒ Combi þ Lipofundin 20% NutriflexÒ lipid plus
DRKS00004238) performed at the trauma intensive care unit at the Volume (mL) 1250a 1250
Hannover Medical School. It conforms to the ethical guidelines of Energy (kcal) 1277 1265
the Declaration of Helsinki (1996) and was approved by the ethics Carbohydrate (g) 150 150
Glucose 100 150
committee of the Hannover Medical School. Consecutive adult
Xylitol 50 0
patients who had been admitted to the intensive care unit for a Fat (g) 50 50
maximum of 3 days before PN and had an indication to receive PN Amino acids (g) 50 48
with xylitol for at least 24 h were included in the study over a Electrolytes 239 198
period of about 8 months. Written informed consent was obtained (mmol)
25
20 The difference between the nominal and measured xylitol
15 concentrations was assessed. The mean values that were deter-
10 mined using the test were 0.815 0.011 (actual 0.8) and
2.079 0.073 (actual 2.0) mg/dL. The mean errors were 1.8 and
5
3.9%, respectively (n ¼ 5 each). Figure 1 demonstrates that the
0
actual and the measured xylitol concentrations were comparable at
0 4 10 20 30 40 a concentration ranging from 0 to 40 mg/dL. The mean xylitol level
Xylitol-concentration (mg/dL) in sera from patients who had received PN with xylitol was 3.4 mg/
dL (range 0.00e7.84; n ¼ 20).
Fig. 1. Comparison of actual vs. measured xylitol concentrations.
Excluded (n = 362)
received incorrect PN
Fig. 2. CONSORT flow diagram showing the number of patients screened, recruited, and dropped.
486 A.S. Schneider et al. / Clinical Nutrition 33 (2014) 483e488
Table 2
Patients’ baseline characteristics. 100 *
PN with xylitol Controls Difference between
groups (p-valuea) 90
Gender n(%) n(%)
Male 36(65.5) 42(75.0) 0.27
80
Female 19(34.5) 14(25.0)
Mean blood glucose values at baseline were similar in the This case-control study was done to determine whether a
groups receiving PN with and without xylitol (7.0 0.7 vs. 7.0 0.8 commercially available xylitol test kit that is normally used in the
(range 6.8e7.2 and 6.6e7.2), respectively, p ¼ 0.78). Both patient food industry is reliable and safe to assess xylitol in human serum
groups received the same mean amount of energy via PN per day as under actual clinical conditions. We first validated the xylitol test.
carbohydrates plus xylitol and carbohydrates (PN with xylitol: Our results showed that it was accurate (error 3.9%) in measuring
810.1 147.6 kcal (n ¼ 55) and controls: 789.8 110.7 kcal (n ¼ 56), xylitol levels found in serum at a range of 0e40 mg/dL. An earlier
p ¼ 0.41). One of the secondary endpoints of this study was the poor quality study that also measured xylitol in serum of 10 pa-
amount of insulin patients required to keep their serum glucose tients enzymatically did not describe the method used in detail.9 In
level at 6.5e9 mmol/L. Twenty percent of the patients receiving addition, their test concomitantly measured sorbitol which meant
xylitol had stable glucose levels and did not need insulin compared that these sugar substitutes could not be differentiated. Other
to only 7.1% of the controls (p ¼ 0.04). Figure 3 shows that subjects studies used gas chromatography to quantify xylitol.11,12 This is of
receiving PN with xylitol had a significantly lower insulin require- course valid. However, the method is tedious and time consuming
ment than control patients. and requires special skills.
A.S. Schneider et al. / Clinical Nutrition 33 (2014) 483e488 487
xylitol also determined AST. They did not mention whether there
was a significant difference between the groups. The values appear
similar since standard deviations depicted in the graph are high.
Also the results are based on few patients (n ¼ 10). In our study
mean ALT, g-GT, and bilirubin levels were comparable in both pa-
tient groups and also did not differ in their mean increase. Similar
to our study Leutenegger et al.9 found slight hyperbilirubinemia in
both patient groups, but they did not compare the groups directly.
Additionally, we did not observe metabolic decompensation due
to xylitol. Lactate was similar in both groups and within the normal
range. This is in agreement with Leutenegger et al.,9 but in contrast to
results from Georgieff et al.11 who described higher lactate values in
the xylitol group. However, these were also within the normal range.
Schricker et al.7,8 observed lower values in the sugar substitute group.
However, patients in this study as already mentioned did not receive
complete PN. Aside from xylitol or glucose, the infusion only included
amino acids and the authors did not give information on the exact
composition of the energy supply. Furthermore, the number of
patients included in all three studies was small (n ¼ 8e18).
We observed no adverse effects considered attributable to PN in
Fig. 4. Serum levels of xylitol vs. amount of xylitol infused in 24 h. the group receiving xylitol. Furthermore, there was no difference
with regard to adverse events between the two patient groups.
Therefore, PN with xylitol was safe for intensive care trauma pa-
Although there is still controversy about whether PN with tients who, based on experience, are metabolically more stable and
xylitol is safe and benefits critical care patients, to our knowledge usually do not have preexisting liver diseases.
there is no study that directly compares complete PN with and The mean blood glucose values at baseline for both groups were
without xylitol with the only difference being that one group comparable. However, the xylitol patient group received signifi-
receives less glucose and in exchange xylitol. Table 3 lists the cantly less insulin during the course of the study with the same
studies that have compared PN with and without xylitol. Leute- energy intake. The requirement was 32.6% lower and 20.0% did not
negger et al.9 compared different PN, but not all results were sta- need insulin at all. This is in accordance with results of López
tistically evaluated. Georgieff et al.11 also compared PN with and Martinez et al.13 who show that patients receiving PN with xylitol
without xylitol, but they gave no information on hepatoxicity. In also had a lower insulin requirement. This means that these
addition, their PN formula was not complete i.e. it contained only patients had a better and more constant metabolic adjustment and,
xylitol or glucose and amino acids. Our study is the first to sys- thus, a lower risk of steatosis hepatis.
tematically compare serum from patients receiving complete PN Mean xylitol levels (3.2 mg/dL) were lower in our study as
with and without xylitol. We compared serum from 55 patients compared to values found in the Georgieff study (about 7.5 mg/
receiving PN containing xylitol with serum from a retrospective dL).11 This was most likely due to differences in the amount of
group (n ¼ 56) who received the same amount of energy, amino xylitol that were given which was about twice the amount we gave,
acids, and lipids. The only difference was that control patients and perhaps to differences in the test method. We found no accu-
received 150 g glucose per bag while patients in the xylitol group mulation of xylitol in patient serum after they had stopped
received 100 g glucose plus xylitol per bag. The patients in both receiving PN with xylitol. Serum levels returned almost to baseline
groups were similar with regard to age, gender, diagnoses, diabetes, values within 24 h. Although there is no defined cut-off point
and SAPS score and, thus, comparable. known for xylitol, the levels were at a very low range (close to the
There were no clinical or laboratory signs of hepatotoxity due to detection level of 0.5 mg/dL).
xylitol. The groups showed no difference with respect to AST levels We conclude that PN with xylitol may be safe for critical care
or an increase of the level by at least double. Leutenegger et al.9 trauma patients. There were no signs of hepatoxicity due to the
who studied patients who were given PN with and without sugar substitute. In addition, the lower insulin requirement may be
Table 3
Studies that compare xylitol and glucose as energy sources in parenteral nutrition (PN).
Georgieff M et al. 198111 Lower glucose levels and insulin PN formula contained no lipids, n ¼ 11e13 per group;
requirement in the xylitol group gaschromatographic xylitol measurement; postoperative
gastric patients; PN: 5 d
Leutenegger A et al. 19849 Lower glucose levels and insulin Only xylitol group received lipids; n ¼ 10 per group;
requirement in the xylitol group enzymatic xylitol/sorbitol measurement which is not
described; trauma patients
Schricker T et al. 19937 Lower glucose levels in xylitol and Study 1: Compared glucose and xylitol PN in postoperative
xylitol/glucose groups in both studies pancreatitis or abdominal abscess patients; Study 2:
Compared glucose and xylitol/glucose (1:1) PN in sepsis
patients; n ¼ 8 per group; PN: 6 h; the exact formulations
of PN is not given; xylitol was not assessed
Schricker T et al. 19948 Lower glucose levels in the xylitol Compared glucose and xylitol/glucose (1:1); n ¼ 6 per
group, lower increase in serum insulin group; postoperative cardiac patients PN: 36 h; the exact
levels formulations of PN is not given; xylitol was not assessed
488 A.S. Schneider et al. / Clinical Nutrition 33 (2014) 483e488