Nutrition and Cancer, 66(6), 1070–1076 C 2014, Taylor & Francis Group, LLC Copyright ISSN: 0163-5581 print / 1532-7914 online DOI: 10.1080/01635581.2014.916316
Short-Term Intravenous Fish-Oil Emulsions in Pediatric Oncologic Patients—Effect on Liver Parameters K. Martin Hoffmann and Michael Grabowski Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Austria
Siegfried R¨odl Pediatric Intensive Care Unit, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Andrea Deutschmann Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Gerold Schwantzer Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Graz, Austria
Petra Sovinz, Volker Strenger, and Christian Urban Division of Pediatric Hematology/Oncology, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Wolfgang Muntean and Almuth C. Hauer Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Pediatric oncologic patients often need parenteral nutrition (PN) during chemotherapy. Long-term use of soybean-based lipid emulsions is associated with progressive liver disease and cholestasis, whereas fish-oil based emulsions have anticholestatic effects. We studied the potentially hepato-protective effects of short-term use of SMOF lipids in children undergoing chemotherapy. Fifteen pediatric oncologic patients treated with SMOF lipids were retrospectively analyzed in respect to bilirubin and liver parameters and compared to matched-controls who had received soybeanbased fat emulsions. For statistics the time-points baseline, Day 14 of PN (PN14), and post (Day+7) were chosen. None of the study patients developed cholestasis. Within the SMOF-lipid group there were no differences in the laboratory parameters between baseline, PN14, and post. In the control group, gamma glutamyltransferase (γ GT) levels increased during PN (baseline vs. PN14, 26.43 vs.
Submitted 23 February 2013; accepted in final form 10 March 2014. Address correspondence to Karl Martin Hoffmann, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Auenbruggerplatz 34/2, A-8036, Graz, Austria. Tel: +43 316 385 83729. Fax: +43 316 385 14961. E-mail: [email protected]
63.00 U/l, P < 0.05). Lactate dehydrogenase (LDH) levels showed a significantly different behavior in the 2 groups: In the SMOF lipids group, LDH decreased whereas it increased in the controls (−32.75 U/l vs. + 29.57 U/l, P < 0.05). An advantage of fish oilbased fat emulsions can be shown even after short-term PN. In children undergoing chemotherapy the use of soybean-based fat emulsions but not SMOF lipids led to increased γ GT levels.
INTRODUCTION Parenteral nutrition related liver disease (PNLD) is a potentially life-limiting disease that occurs in patients who depend on parenteral nutrition (PN) for a prolonged period of time (1, 2). In children PNLD affects mainly preterm babies who often need PN in the first weeks of life and children with intestinal failure who cannot be sufficiently nutritionally sustained via the enteral route. One of the main clinical features is the development of cholestasis [i.e., elevated bilirubin, and/or gammaglutamyl transpeptidase (γ GT) and alkaline phosphatase (AP)] (1,3,4). The patho-physiological mechanisms of the development of PNLD are still not fully understood but are believed
1070
1071
SMOF LIPIDS IN PEDIATRIC ONCOLOGIC PATIENTS
to be multifactorial (e.g., prematurity of the liver, sepsis, lack of enteral feeding, impaired bile acid circulation, overload of parenteral nutrients, etc.). Ultimately the sum of these mechanisms lead to reduced bile flow and cholestasis. In recent years it has become apparent that the composition of the lipid solution used for PN plays a critical role in the natural course of the disease (1,3,4). Recent clinical studies have shown that increasing the omega-3 content of fatty acids [e.g., eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA)] in PN solutions by adding fish oil can potentially reverse the cholestatic effects of long-term PN in children with intestinal failure (5–8). The mechanisms by which these novel omega-3-rich fat emulsions assert their beneficial effect remain to be elucidated. A number of possible explanations have been presented in the literature: Omega-3 fatty acids like EPA and DHA have been shown to cause a shift to antiinflammatory eicosanoid production, which in turn could possibly have a cyto-protective effect on liver and bile-duct cells. (9) Novel omega-3 rich fat preparations have a higher alpha-tocopherol content, which prevents radical oxygen species to react with polyunsaturated fatty acids and has potent antioxidative properties (10). The accumulation of alpha-tocopherol in liver cells via tocopherol-transfer protein could explain part of the liver-protective effects of newer alphatocopherol enriched lipid solutions (11,12). Also, the higher phytosterol content of exclusively plant-based lipid solutions (i.e., based on soybean oil) has led to the implication that plant phytosterols have a direct effect on the development of cholestasis, a theory that is backed by studies on hepatic cell lines (13) and animal experiments (14,15). In reviewing clinical studies comparing fish-oil based to standard soybean-based emulsions it has to be concluded that a clear beneficial effect on liver function has to date only been demonstrated in patients receiving long-term PN (>4 wk) (5,7,8,16). Two studies in preterm babies have analyzed the effects and safety of the short-term use of fish-oil based emulsions and have shown conflicting results in regard to the effect on liver parameters (17,18). It is currently not clear whether fish-oil based emulsions have a clear advantage over standard soybean oil-based products in regard to their hepato-protective and anticholestatic effects if used only for short periods of time. Pediatric hemato-oncologic patients often require PN for prolonged periods of time due to the side effects of chemotherapy (19, 20). Because most chemotherapeutic regimens have unspecific cytotoxic effects (21), a hepato-protective effect of fish-oil based PN lipid emulsions could be of clinical relevance in these patients. In the current study, we therefore attempted to analyze if a novel fish-oil enriched fat emulsion has a hepato-protective (anticholestatic) effect compared to standard soybean formula in pediatric hemato-oncologic patients who received PN during their chemotherapy. METHODS In June 2010, as a result of the growing evidence in the literature of the benefits of adding fish oil to parenteral fat emul-
TABLE 1 Comparison of fatty acid content of Intralipid (100% soy oil) and SMOF lipid (30% soy oil, 30% MCT, 25% olive oil, 15% fish oil) Intralipid 0% 0% 15.8% 22.8% 54.4% 6.8%
Fatty acid (FA) Short chain FA MCT Long-chain FA Monosaturated FA Omega-6 FA Omega-3 FA
SMOF lipid
Difference
0% 30% 9.8% 26.2% 19% 15%
−38 % +14,9 % −65,1 % +120 %
sions, our pediatric oncology department changed their parenteral nutritive regimen from a PN with a standard soybeanR , Fresenius Kabi, Bad Hombased lipid emulsion (Intralipid burg, Germany) to a formula containing soybean oil, MCT, olive R , Fresenius Kabi, Bad Homburg, oil, and fish oil (SMOF lipids Germany). A comparison of the lipid formulas used is provided in Table 1. We performed a retrospective cohort study of pediatric hemato-oncologic patients who required PN during chemotherapy. We compared patients who received SMOF lipids to a historical age-matched patient group who had been treated with soybean-based fat emulsion during their chemotherapy. The hemato-oncologic ward uses a standardized parenteral regime for all patients: In the SMOF lipids group, 1000 ml of parenteral solution contained 15% carbohydrates, 3% aminoacids, and 2.0% fat. In the control group, 1000 ml of parenteral solution contained 14.5% carbohydrates, 2.7% aminoacids and 1.8% fat. Daily total parenteral nutrition (TPN) doses were calculated for 20-h infusion periods to minimize TPN infusion rates. Data on enteral nutrition and especially amount and quality of food consumed orally were not documented in a way sufficient for analysis and comparison between groups and were therefore not included in our study. We included patients who were treated at our pediatric hemato-oncologic ward between 2004 and 2011. Inclusion criteria were patients suffering from hematooncologic or oncologic disease (age 1 to 18 yr) requiring chemotherapy. To reduce possible bias (change in treatment protocols and chemotherapeutics used) we included only patients who underwent stem cell transplantation (induction chemotherapy) or chemotherapy for acute myeloid leukemia (AML), both conditions often requiring longer periods of PN. PN during chemotherapy including either Intralipid or SMOF lipid lasted for at least 14 consecutive days. Patients with elevation of liver enzymes and/or bilirubin at start of PN where not included. Laboratory data including total bilirubin, γ GT, aspartate transaminase (AST), alanine transaminase (ALT), AP, and lactate dehydrogenase (LDH) were retrospectively extracted from our hospitals medical database. For statistical analysis the following time points before, during and after PN were defined:
1072
K. M. HOFFMANN ET AL.
baseline (mean of 7 days before start of PN), PN14 (Day 14 of PN), post (Day 7 after discontinuation of PN, full enteral feeding). Statistical analysis was performed using PASW (PASW Statistics for Windows, SPSS, Chicago, IL.). Data are displayed as median with interquartile range (IQR, Q1-Q3). Two-tailed t-test was used for group comparisons (patients’ characteristics). To assess the changes of measured serum values we calculated the differences PN14-baseline, post-PN14, and postbaseline. For within group comparisons, we used the Wilcoxon signed rank test. Group differences in the serum values and in the changes of the serum values were tested with the MannWithney U test. We used exact tests and report uncorrected P values, which were considered significant if less than 0.05. Our local ethical committee at the Medical University of Graz approved our study (24-095 ex 11/12).
RESULTS Thirty-two patients were eligible, 30 patients (15 treated with standard soybean-based emulsion, 15 with SMOF lipids) met our inclusion criteria. Two patients who were eligible were excluded because of preexisting liver disease (i.e., marked bilirubin and liver enzyme elevation at start of chemotherapy and/or PN). Patients in both groups were comparable regarding gender and mean age. There were no statistical differences in the amount of PN (49.02 vs. 52.93 ml/kg body weight/day, n.s.) or the dose of total fat given (0.92 vs. 0.98 g/kg body weight/day, mean difference 0.06, n.s.; Table 2). All included patients had their PN started after the beginning of their chemotherapy (Table 2). The oncologic diagnoses were heterogenous: 21 patients had hemato-oncologic diagnoses (9 AML, 2 T-ALL, 10 other hematologic disease), 4 had brain tumors, and 5 had solid tumors. Table 2 gives an overview of the chemotherapeutic drugs used in the study groups. Twenty-three of the 30 patients (11 in the soybean-based group, 12 in the SMOF lipids group) received their PN during stem cell transplantation. The remaining 7 patients had conventional chemotherapy for AML (chemotherapeutic drugs used in these patients included cytarabine, daunorubicine, etoposide, mitoxanthron, idarubicine, and thioguanine). In the group who had received standard soybean-based fat emulsion (within-group comparison, Table 3), there was no significant increase in serum values of ALT, AP, and LDH between the chosen time points (baseline, PN14, post). There was a significant decrease of total bilirubin between the time points baseline and post (0.500 vs. 0.385; P = 0.030), whereas there was no statistical difference between baseline and PN14 or between PN14 and post. γ GT increased significantly from baseline to PN14 and from baseline to post (26.4 at baseline vs. 63.0 at PN14 and 63.0 at post; P = 0.0497 and 0.048, respectively; Table 3). In the SMOF lipid group there was no significant difference in serum values of total bilirubin, γ GT, AST, ALT, AP, and LDH (Table 4). When comparing both groups there was
no significant difference in total bilirubin, γ GT, AST, ALT, AP, and LDH (data not shown). When analyzing the changes of values (PN14-baseline, postbaseline) we found no significant difference in change of serum levels of total bilirubin, γ GT, AST, ALT, and AP (data not shown). There was a significant difference in the change in serum LDH levels between the 2 groups (Fig. 1): LDH levels in the SMOF lipid group changed to lower levels, whereas in the soybean-based group there was a change to higher levels (−32.75 vs. 29.57; P = 0.016). DISCUSSION Our study showed that when used for short periods of time for PN both lipid solutions used were safe in regard to the development of PNLD, because none of our patients developed cholestasis during the study period. A fish-oil containing fat emulsion (SMOF lipids) seems to have an advantage over standard soybean-based formulas: While none of the patients showed elevated bilirubin levels above levels considered abnormal, we found a significant increase in γ GT levels in the patient group treated with conventional soybean-based fat emulsion, which can be interpreted as an early sign of cholestasis(21–23), but not in patients treated with SMOF lipids. Recent clinical studies have demonstrated, that in children who require long-term PN changing to novel PN regimens with fat emulsions that are rich in omega-3 fatty acids (fish oil) and ∝-tocopherol, and have low phytosterol content can positively influence the natural course of cholestasis in PNLD: Gura et al. first demonstrated in two children that adding fish-oil to PN fat emulsions led to improvement of cholestasis in children with PNLD. (6) The same authors later showed that changing soybean-based formula to a regimen containing fish oil (SMOF lipids) led to resolution of already existing cholestasis caused by long-term PN in most patients.(7) These findings were confirmed by Muhammed et al. (8), who switched children with short bowel syndrome to a fish oil-based regimen once patients showed signs of cholestasis during PN containing a conventional soybean-based formula. Seven of 8 patients showed improvement or resolution of bilirubin levels. Compared to a historical cohort the intervention group had significantly lower bilirubin levels in the change of values analysis. A recent double blind randomized controlled study by Goulet et al. (5) prospectively enrolled patients requiring home parenteral nutrition to receive either fish oil containing formula (SMOF lipids) or standard soybean-based emulsion for more than 4 wk. Apart from demonstrating that fish oil-based formula was well tolerated; they found a significant decrease in bilirubin levels (within levels considered normal) in the group receiving fish oil. With respect to the short-term application of fish oil-based regimens compared to a standard soybean-based emulsion 2 randomized controlled studies in preterm neonates demonstrated a beneficial effect on liver function and development of cholestasis. Although Tomsits et al. (18) showed a significant increase
1073
SMOF LIPIDS IN PEDIATRIC ONCOLOGIC PATIENTS
TABLE 2 Descriptive data on patient characteristics, received parenteral nutrition, chemotherapy, and chemotherapeutic agents used
Female Mean age, years Mean weight, KG Parenteral nutrition dose parenteral nutritionf ml/kg BW/d Mean dose fatf g/kg BW/d Mean dose carbohydrates (glucose)f g/kg BW/d Chemotherapy Chemotherapy without PN, days Chemotherapeutics used for SCT induction§
Total 30#
SMOF 15 (50)
Soybean-based 15 (50)
P value
17 (56.7) 10.69 ± 7.11 35.97 ± 21.39
8 (53.3) 10.70 ± 7.2 33.41 ± 17.5
9 (60) 10.69 ± 7.27 38.54 ± 25.08
0.71 0.26 0.87
59.98 ± 24.49 0.95 ± 0.45 7.65 ± 3.67
49.02 ± 16.85 0.92 ± 0.33 7.35 ± 2.53
52.93 ± 30.81 0.98 ± 0.55 7.94 ± 4.62
0.67 0.76 0.67
6.03 ± 5.70
5.3 ± 4.90 Cyclophosphamide (8) ATG (5) Thiopeta (5) Treosulfan (4) Fludarabin (3) Etoposide (1) Melphalan (2) Busulfan (1) Carboplatin (0) Cytarabin (0)
6.6 ± 6.17 Cyclophosphamide (3) ATG (5) Thiopeta (5) Treosulfan (4) Fludarabin (3) Etoposide (3) Melphalan (5) Busulfan (2) Carboplatin (3) Cytarabin (1)
0.54
There were no statistical differences between the groups in the clinical parameters shown. PN = parenteral nutrition; BW = body weight; SCT = stem cell transplantation; ATG = anti-hymocyte globulin. Two-tailed student’s t-test was used for group comparisons. # Numbers are expressed in n (%) and mean ± SD. § Chemotherapeutic is followed by the number (in parenthesis) of patients who received the drug in this group. f Daily total parenteral nutrition (TPN) doses were calculated for 20-hour infusion periods in order to minimize TPN infusion rates.
TABLE 3 Within-group comparisons of bilirubin, liver enzymes, and LDH levels in patients receiving soybean-based fat emulsion Soybean-based fat emulsion Within-groups∗ P =
Bili-tot GGT AST ALT AP LDH
mg/dl U/l U/l U/l U/l U/l
Standard 0.10–1.20 −38 −30 −35 35–105 120–240
Median (IQR)
Median (IQR)
Median (IQR)
PN14 vs. baseline
Post vs. baseline
Post vs. PN14
Baseline1 0.500 (0.615) 26.43 (32.90) 47.83 (53.88) 45.86 (63.00) 123.4 (128.8) 244.6 (137.7)
PN142 0.410 (0.390) 63.00 (49.00) 31.00 (36.00) 32.00 (44.50) 154.5 (128.0) 239.0 (103.0)
Post3 0.385 (0.200) 63.00 (61.00) 46.00 (27.00) 49.00 (48.00) 141.0 (76.0) 272.0 (100.0)
0.121 0.0497 0.041 0.252 0.639 0.804
0.030 0.048 0.421 0.762 0.847 0.208
0.169 0.795 0.095 0.492 0.879 0.397
IQR = interquartile range Q3–Q1;Bili-tot = total bilirubin; GGT = gamma-glutamyl transferase; AST = asparatate transaminase; ALT = alanine transaminase; AP = alkaline phosphatase; LDH = lactate dehydrogenase. ∗ Exact P-values of Wilcoxon Signed Rank Test. 1 Baseline = mean value of Day 7-1 pre parenteral nutrition (15 children observed). 2 PN14 = Day 14 on parenteral nutrition (15 children observed; offset −4 for 2, −3, −2, and −1 for 1 patient, respectively). 3 Post = Day +7 post PN14 phase (15 children observed; offset −5 for one patient, −1 and +1 for 1 patient, respectively).
1074
K. M. HOFFMANN ET AL.
TABLE 4 Within group comparisons of bilirubin, liver enzymes and LDH levels in patients receiving SMOF lipids SMOF lipids Within groups∗ P =
Bili-tot GGT AST ALT AP LDH
mg/dl U/l U/l U/l U/l U/l
Standard 0.10–1.20 −38 −30 −35 35–105 120–240
Median (IQR)
Median (IQR)
Median (IQR)
PN14 vs. baseline
Post vs. PN14
Post vs. baseline
Baseline1 0.363 (0.312) 27.45 (30.71) 52.93 (71.71) 53.40 (57.98) 95.0 (104.6) 237.6 (129.0)
PN142 0.370 (0.340) 50.00 (97.00) 44.00 (23.00) 60.00 (65.00) 126.0 (70.0) 286.0 (149.0)
Post3 0.500 (0.380) 56.00 (66.00) 45.00 (41.00) 46.00 (66.50) 137.0 (39.0) 251.0 (103.0)
0.235 0.083 0.188 0.639 0.890 0.454
0.814 0.636 0.987 0.247 0.890 0.638
0.720 0.107 0.203 < 0.999 0.169 0.075
IQR = interquartile range Q3–Q1;Bili-tot = total bilirubin; GGT = gamma-glutamyl transferase; AST = asparatate transaminase; ALT = alanine transaminase; AP = alkaline phosphatase; LDH = lactate dehydrogenase. ∗ Exact P-values of Wilcoxon Signed Rank Test. 1 Baseline = mean value of Day 7-1 pre parenteral nutrition (15 children observed). 2 PN14 = Day 14 on parenteral nutrition (15 children observed; offset −4 for 2, −3, −2, and −1 for 1 patient, respectively). 3 Post = Day +7 post PN14 phase (15 children observed; offset −5 for one patient, −1 and +1 for 1 patient, respectively).
of γ GT in the group receiving standard soybean lipid emulsion, there was no such change in the group treated with the fish oil containing fat emulsion. Rayyen et al. (17), who used a similar study design, did not observe this change in γ GT levels, however, they found significant increases and decreases of bilirubin levels (within ranges considered normal at this age) in the groups receiving soybean-based and fish oil-based lipids, respectively. In our study we found a significant increase of γ GT levels only in the patient group receiving standard soybean-based lipids (PN14 and post). We observed a decrease of bilirubin (well within the normal ranges) in the soybean-based group, which was an unexpected finding. The study of Gura et al. provides data, which allow the interpretation that PNLD potentially can be reversed when continuing parenteral soybean-based lipid solutions. Furthermore, animal studies have shown reversal of cholestasis using soybean oil (6,24). However, our observation of increased γ GT levels in the group receiving soybean-based lipids, and the growing evidence of clinical trials showing reversal of cholastasis (bilirubin and γ GT) support the conclusion of the beneficial effects of fish oil or fish oil-containing lipid solutions on PNLD. Although studies investigating short-time effects of fish oil-based PN (14 days), including our study, show changes that could be interpreted as early signs of cholestasis (i.e., rise in γ GT) (22,23), it has to be stated that most of these observed changes were within or near ranges considered normal. Whether this rise in γ GT is of clinical relevance might be investigated in a long-time follow-up cohort study in patients who have been treated with standard soybean-based emulsions for short periods of time (e.g.
Short-Term Intravenous Fish-Oil Emulsions in Pediatric Oncologic Patients—Effect on Liver Parameters K. Martin Hoffmann and Michael Grabowski Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Austria
Siegfried R¨odl Pediatric Intensive Care Unit, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Andrea Deutschmann Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Gerold Schwantzer Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Graz, Austria
Petra Sovinz, Volker Strenger, and Christian Urban Division of Pediatric Hematology/Oncology, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Wolfgang Muntean and Almuth C. Hauer Division of General Pediatrics, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Graz, Austria
Pediatric oncologic patients often need parenteral nutrition (PN) during chemotherapy. Long-term use of soybean-based lipid emulsions is associated with progressive liver disease and cholestasis, whereas fish-oil based emulsions have anticholestatic effects. We studied the potentially hepato-protective effects of short-term use of SMOF lipids in children undergoing chemotherapy. Fifteen pediatric oncologic patients treated with SMOF lipids were retrospectively analyzed in respect to bilirubin and liver parameters and compared to matched-controls who had received soybeanbased fat emulsions. For statistics the time-points baseline, Day 14 of PN (PN14), and post (Day+7) were chosen. None of the study patients developed cholestasis. Within the SMOF-lipid group there were no differences in the laboratory parameters between baseline, PN14, and post. In the control group, gamma glutamyltransferase (γ GT) levels increased during PN (baseline vs. PN14, 26.43 vs.
Submitted 23 February 2013; accepted in final form 10 March 2014. Address correspondence to Karl Martin Hoffmann, Department of Pediatrics and Adolescent Medicine, Medical University Graz, Auenbruggerplatz 34/2, A-8036, Graz, Austria. Tel: +43 316 385 83729. Fax: +43 316 385 14961. E-mail: [email protected]
63.00 U/l, P < 0.05). Lactate dehydrogenase (LDH) levels showed a significantly different behavior in the 2 groups: In the SMOF lipids group, LDH decreased whereas it increased in the controls (−32.75 U/l vs. + 29.57 U/l, P < 0.05). An advantage of fish oilbased fat emulsions can be shown even after short-term PN. In children undergoing chemotherapy the use of soybean-based fat emulsions but not SMOF lipids led to increased γ GT levels.
INTRODUCTION Parenteral nutrition related liver disease (PNLD) is a potentially life-limiting disease that occurs in patients who depend on parenteral nutrition (PN) for a prolonged period of time (1, 2). In children PNLD affects mainly preterm babies who often need PN in the first weeks of life and children with intestinal failure who cannot be sufficiently nutritionally sustained via the enteral route. One of the main clinical features is the development of cholestasis [i.e., elevated bilirubin, and/or gammaglutamyl transpeptidase (γ GT) and alkaline phosphatase (AP)] (1,3,4). The patho-physiological mechanisms of the development of PNLD are still not fully understood but are believed
1070
1071
SMOF LIPIDS IN PEDIATRIC ONCOLOGIC PATIENTS
to be multifactorial (e.g., prematurity of the liver, sepsis, lack of enteral feeding, impaired bile acid circulation, overload of parenteral nutrients, etc.). Ultimately the sum of these mechanisms lead to reduced bile flow and cholestasis. In recent years it has become apparent that the composition of the lipid solution used for PN plays a critical role in the natural course of the disease (1,3,4). Recent clinical studies have shown that increasing the omega-3 content of fatty acids [e.g., eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA)] in PN solutions by adding fish oil can potentially reverse the cholestatic effects of long-term PN in children with intestinal failure (5–8). The mechanisms by which these novel omega-3-rich fat emulsions assert their beneficial effect remain to be elucidated. A number of possible explanations have been presented in the literature: Omega-3 fatty acids like EPA and DHA have been shown to cause a shift to antiinflammatory eicosanoid production, which in turn could possibly have a cyto-protective effect on liver and bile-duct cells. (9) Novel omega-3 rich fat preparations have a higher alpha-tocopherol content, which prevents radical oxygen species to react with polyunsaturated fatty acids and has potent antioxidative properties (10). The accumulation of alpha-tocopherol in liver cells via tocopherol-transfer protein could explain part of the liver-protective effects of newer alphatocopherol enriched lipid solutions (11,12). Also, the higher phytosterol content of exclusively plant-based lipid solutions (i.e., based on soybean oil) has led to the implication that plant phytosterols have a direct effect on the development of cholestasis, a theory that is backed by studies on hepatic cell lines (13) and animal experiments (14,15). In reviewing clinical studies comparing fish-oil based to standard soybean-based emulsions it has to be concluded that a clear beneficial effect on liver function has to date only been demonstrated in patients receiving long-term PN (>4 wk) (5,7,8,16). Two studies in preterm babies have analyzed the effects and safety of the short-term use of fish-oil based emulsions and have shown conflicting results in regard to the effect on liver parameters (17,18). It is currently not clear whether fish-oil based emulsions have a clear advantage over standard soybean oil-based products in regard to their hepato-protective and anticholestatic effects if used only for short periods of time. Pediatric hemato-oncologic patients often require PN for prolonged periods of time due to the side effects of chemotherapy (19, 20). Because most chemotherapeutic regimens have unspecific cytotoxic effects (21), a hepato-protective effect of fish-oil based PN lipid emulsions could be of clinical relevance in these patients. In the current study, we therefore attempted to analyze if a novel fish-oil enriched fat emulsion has a hepato-protective (anticholestatic) effect compared to standard soybean formula in pediatric hemato-oncologic patients who received PN during their chemotherapy. METHODS In June 2010, as a result of the growing evidence in the literature of the benefits of adding fish oil to parenteral fat emul-
TABLE 1 Comparison of fatty acid content of Intralipid (100% soy oil) and SMOF lipid (30% soy oil, 30% MCT, 25% olive oil, 15% fish oil) Intralipid 0% 0% 15.8% 22.8% 54.4% 6.8%
Fatty acid (FA) Short chain FA MCT Long-chain FA Monosaturated FA Omega-6 FA Omega-3 FA
SMOF lipid
Difference
0% 30% 9.8% 26.2% 19% 15%
−38 % +14,9 % −65,1 % +120 %
sions, our pediatric oncology department changed their parenteral nutritive regimen from a PN with a standard soybeanR , Fresenius Kabi, Bad Hombased lipid emulsion (Intralipid burg, Germany) to a formula containing soybean oil, MCT, olive R , Fresenius Kabi, Bad Homburg, oil, and fish oil (SMOF lipids Germany). A comparison of the lipid formulas used is provided in Table 1. We performed a retrospective cohort study of pediatric hemato-oncologic patients who required PN during chemotherapy. We compared patients who received SMOF lipids to a historical age-matched patient group who had been treated with soybean-based fat emulsion during their chemotherapy. The hemato-oncologic ward uses a standardized parenteral regime for all patients: In the SMOF lipids group, 1000 ml of parenteral solution contained 15% carbohydrates, 3% aminoacids, and 2.0% fat. In the control group, 1000 ml of parenteral solution contained 14.5% carbohydrates, 2.7% aminoacids and 1.8% fat. Daily total parenteral nutrition (TPN) doses were calculated for 20-h infusion periods to minimize TPN infusion rates. Data on enteral nutrition and especially amount and quality of food consumed orally were not documented in a way sufficient for analysis and comparison between groups and were therefore not included in our study. We included patients who were treated at our pediatric hemato-oncologic ward between 2004 and 2011. Inclusion criteria were patients suffering from hematooncologic or oncologic disease (age 1 to 18 yr) requiring chemotherapy. To reduce possible bias (change in treatment protocols and chemotherapeutics used) we included only patients who underwent stem cell transplantation (induction chemotherapy) or chemotherapy for acute myeloid leukemia (AML), both conditions often requiring longer periods of PN. PN during chemotherapy including either Intralipid or SMOF lipid lasted for at least 14 consecutive days. Patients with elevation of liver enzymes and/or bilirubin at start of PN where not included. Laboratory data including total bilirubin, γ GT, aspartate transaminase (AST), alanine transaminase (ALT), AP, and lactate dehydrogenase (LDH) were retrospectively extracted from our hospitals medical database. For statistical analysis the following time points before, during and after PN were defined:
1072
K. M. HOFFMANN ET AL.
baseline (mean of 7 days before start of PN), PN14 (Day 14 of PN), post (Day 7 after discontinuation of PN, full enteral feeding). Statistical analysis was performed using PASW (PASW Statistics for Windows, SPSS, Chicago, IL.). Data are displayed as median with interquartile range (IQR, Q1-Q3). Two-tailed t-test was used for group comparisons (patients’ characteristics). To assess the changes of measured serum values we calculated the differences PN14-baseline, post-PN14, and postbaseline. For within group comparisons, we used the Wilcoxon signed rank test. Group differences in the serum values and in the changes of the serum values were tested with the MannWithney U test. We used exact tests and report uncorrected P values, which were considered significant if less than 0.05. Our local ethical committee at the Medical University of Graz approved our study (24-095 ex 11/12).
RESULTS Thirty-two patients were eligible, 30 patients (15 treated with standard soybean-based emulsion, 15 with SMOF lipids) met our inclusion criteria. Two patients who were eligible were excluded because of preexisting liver disease (i.e., marked bilirubin and liver enzyme elevation at start of chemotherapy and/or PN). Patients in both groups were comparable regarding gender and mean age. There were no statistical differences in the amount of PN (49.02 vs. 52.93 ml/kg body weight/day, n.s.) or the dose of total fat given (0.92 vs. 0.98 g/kg body weight/day, mean difference 0.06, n.s.; Table 2). All included patients had their PN started after the beginning of their chemotherapy (Table 2). The oncologic diagnoses were heterogenous: 21 patients had hemato-oncologic diagnoses (9 AML, 2 T-ALL, 10 other hematologic disease), 4 had brain tumors, and 5 had solid tumors. Table 2 gives an overview of the chemotherapeutic drugs used in the study groups. Twenty-three of the 30 patients (11 in the soybean-based group, 12 in the SMOF lipids group) received their PN during stem cell transplantation. The remaining 7 patients had conventional chemotherapy for AML (chemotherapeutic drugs used in these patients included cytarabine, daunorubicine, etoposide, mitoxanthron, idarubicine, and thioguanine). In the group who had received standard soybean-based fat emulsion (within-group comparison, Table 3), there was no significant increase in serum values of ALT, AP, and LDH between the chosen time points (baseline, PN14, post). There was a significant decrease of total bilirubin between the time points baseline and post (0.500 vs. 0.385; P = 0.030), whereas there was no statistical difference between baseline and PN14 or between PN14 and post. γ GT increased significantly from baseline to PN14 and from baseline to post (26.4 at baseline vs. 63.0 at PN14 and 63.0 at post; P = 0.0497 and 0.048, respectively; Table 3). In the SMOF lipid group there was no significant difference in serum values of total bilirubin, γ GT, AST, ALT, AP, and LDH (Table 4). When comparing both groups there was
no significant difference in total bilirubin, γ GT, AST, ALT, AP, and LDH (data not shown). When analyzing the changes of values (PN14-baseline, postbaseline) we found no significant difference in change of serum levels of total bilirubin, γ GT, AST, ALT, and AP (data not shown). There was a significant difference in the change in serum LDH levels between the 2 groups (Fig. 1): LDH levels in the SMOF lipid group changed to lower levels, whereas in the soybean-based group there was a change to higher levels (−32.75 vs. 29.57; P = 0.016). DISCUSSION Our study showed that when used for short periods of time for PN both lipid solutions used were safe in regard to the development of PNLD, because none of our patients developed cholestasis during the study period. A fish-oil containing fat emulsion (SMOF lipids) seems to have an advantage over standard soybean-based formulas: While none of the patients showed elevated bilirubin levels above levels considered abnormal, we found a significant increase in γ GT levels in the patient group treated with conventional soybean-based fat emulsion, which can be interpreted as an early sign of cholestasis(21–23), but not in patients treated with SMOF lipids. Recent clinical studies have demonstrated, that in children who require long-term PN changing to novel PN regimens with fat emulsions that are rich in omega-3 fatty acids (fish oil) and ∝-tocopherol, and have low phytosterol content can positively influence the natural course of cholestasis in PNLD: Gura et al. first demonstrated in two children that adding fish-oil to PN fat emulsions led to improvement of cholestasis in children with PNLD. (6) The same authors later showed that changing soybean-based formula to a regimen containing fish oil (SMOF lipids) led to resolution of already existing cholestasis caused by long-term PN in most patients.(7) These findings were confirmed by Muhammed et al. (8), who switched children with short bowel syndrome to a fish oil-based regimen once patients showed signs of cholestasis during PN containing a conventional soybean-based formula. Seven of 8 patients showed improvement or resolution of bilirubin levels. Compared to a historical cohort the intervention group had significantly lower bilirubin levels in the change of values analysis. A recent double blind randomized controlled study by Goulet et al. (5) prospectively enrolled patients requiring home parenteral nutrition to receive either fish oil containing formula (SMOF lipids) or standard soybean-based emulsion for more than 4 wk. Apart from demonstrating that fish oil-based formula was well tolerated; they found a significant decrease in bilirubin levels (within levels considered normal) in the group receiving fish oil. With respect to the short-term application of fish oil-based regimens compared to a standard soybean-based emulsion 2 randomized controlled studies in preterm neonates demonstrated a beneficial effect on liver function and development of cholestasis. Although Tomsits et al. (18) showed a significant increase
1073
SMOF LIPIDS IN PEDIATRIC ONCOLOGIC PATIENTS
TABLE 2 Descriptive data on patient characteristics, received parenteral nutrition, chemotherapy, and chemotherapeutic agents used
Female Mean age, years Mean weight, KG Parenteral nutrition dose parenteral nutritionf ml/kg BW/d Mean dose fatf g/kg BW/d Mean dose carbohydrates (glucose)f g/kg BW/d Chemotherapy Chemotherapy without PN, days Chemotherapeutics used for SCT induction§
Total 30#
SMOF 15 (50)
Soybean-based 15 (50)
P value
17 (56.7) 10.69 ± 7.11 35.97 ± 21.39
8 (53.3) 10.70 ± 7.2 33.41 ± 17.5
9 (60) 10.69 ± 7.27 38.54 ± 25.08
0.71 0.26 0.87
59.98 ± 24.49 0.95 ± 0.45 7.65 ± 3.67
49.02 ± 16.85 0.92 ± 0.33 7.35 ± 2.53
52.93 ± 30.81 0.98 ± 0.55 7.94 ± 4.62
0.67 0.76 0.67
6.03 ± 5.70
5.3 ± 4.90 Cyclophosphamide (8) ATG (5) Thiopeta (5) Treosulfan (4) Fludarabin (3) Etoposide (1) Melphalan (2) Busulfan (1) Carboplatin (0) Cytarabin (0)
6.6 ± 6.17 Cyclophosphamide (3) ATG (5) Thiopeta (5) Treosulfan (4) Fludarabin (3) Etoposide (3) Melphalan (5) Busulfan (2) Carboplatin (3) Cytarabin (1)
0.54
There were no statistical differences between the groups in the clinical parameters shown. PN = parenteral nutrition; BW = body weight; SCT = stem cell transplantation; ATG = anti-hymocyte globulin. Two-tailed student’s t-test was used for group comparisons. # Numbers are expressed in n (%) and mean ± SD. § Chemotherapeutic is followed by the number (in parenthesis) of patients who received the drug in this group. f Daily total parenteral nutrition (TPN) doses were calculated for 20-hour infusion periods in order to minimize TPN infusion rates.
TABLE 3 Within-group comparisons of bilirubin, liver enzymes, and LDH levels in patients receiving soybean-based fat emulsion Soybean-based fat emulsion Within-groups∗ P =
Bili-tot GGT AST ALT AP LDH
mg/dl U/l U/l U/l U/l U/l
Standard 0.10–1.20 −38 −30 −35 35–105 120–240
Median (IQR)
Median (IQR)
Median (IQR)
PN14 vs. baseline
Post vs. baseline
Post vs. PN14
Baseline1 0.500 (0.615) 26.43 (32.90) 47.83 (53.88) 45.86 (63.00) 123.4 (128.8) 244.6 (137.7)
PN142 0.410 (0.390) 63.00 (49.00) 31.00 (36.00) 32.00 (44.50) 154.5 (128.0) 239.0 (103.0)
Post3 0.385 (0.200) 63.00 (61.00) 46.00 (27.00) 49.00 (48.00) 141.0 (76.0) 272.0 (100.0)
0.121 0.0497 0.041 0.252 0.639 0.804
0.030 0.048 0.421 0.762 0.847 0.208
0.169 0.795 0.095 0.492 0.879 0.397
IQR = interquartile range Q3–Q1;Bili-tot = total bilirubin; GGT = gamma-glutamyl transferase; AST = asparatate transaminase; ALT = alanine transaminase; AP = alkaline phosphatase; LDH = lactate dehydrogenase. ∗ Exact P-values of Wilcoxon Signed Rank Test. 1 Baseline = mean value of Day 7-1 pre parenteral nutrition (15 children observed). 2 PN14 = Day 14 on parenteral nutrition (15 children observed; offset −4 for 2, −3, −2, and −1 for 1 patient, respectively). 3 Post = Day +7 post PN14 phase (15 children observed; offset −5 for one patient, −1 and +1 for 1 patient, respectively).
1074
K. M. HOFFMANN ET AL.
TABLE 4 Within group comparisons of bilirubin, liver enzymes and LDH levels in patients receiving SMOF lipids SMOF lipids Within groups∗ P =
Bili-tot GGT AST ALT AP LDH
mg/dl U/l U/l U/l U/l U/l
Standard 0.10–1.20 −38 −30 −35 35–105 120–240
Median (IQR)
Median (IQR)
Median (IQR)
PN14 vs. baseline
Post vs. PN14
Post vs. baseline
Baseline1 0.363 (0.312) 27.45 (30.71) 52.93 (71.71) 53.40 (57.98) 95.0 (104.6) 237.6 (129.0)
PN142 0.370 (0.340) 50.00 (97.00) 44.00 (23.00) 60.00 (65.00) 126.0 (70.0) 286.0 (149.0)
Post3 0.500 (0.380) 56.00 (66.00) 45.00 (41.00) 46.00 (66.50) 137.0 (39.0) 251.0 (103.0)
0.235 0.083 0.188 0.639 0.890 0.454
0.814 0.636 0.987 0.247 0.890 0.638
0.720 0.107 0.203 < 0.999 0.169 0.075
IQR = interquartile range Q3–Q1;Bili-tot = total bilirubin; GGT = gamma-glutamyl transferase; AST = asparatate transaminase; ALT = alanine transaminase; AP = alkaline phosphatase; LDH = lactate dehydrogenase. ∗ Exact P-values of Wilcoxon Signed Rank Test. 1 Baseline = mean value of Day 7-1 pre parenteral nutrition (15 children observed). 2 PN14 = Day 14 on parenteral nutrition (15 children observed; offset −4 for 2, −3, −2, and −1 for 1 patient, respectively). 3 Post = Day +7 post PN14 phase (15 children observed; offset −5 for one patient, −1 and +1 for 1 patient, respectively).
of γ GT in the group receiving standard soybean lipid emulsion, there was no such change in the group treated with the fish oil containing fat emulsion. Rayyen et al. (17), who used a similar study design, did not observe this change in γ GT levels, however, they found significant increases and decreases of bilirubin levels (within ranges considered normal at this age) in the groups receiving soybean-based and fish oil-based lipids, respectively. In our study we found a significant increase of γ GT levels only in the patient group receiving standard soybean-based lipids (PN14 and post). We observed a decrease of bilirubin (well within the normal ranges) in the soybean-based group, which was an unexpected finding. The study of Gura et al. provides data, which allow the interpretation that PNLD potentially can be reversed when continuing parenteral soybean-based lipid solutions. Furthermore, animal studies have shown reversal of cholestasis using soybean oil (6,24). However, our observation of increased γ GT levels in the group receiving soybean-based lipids, and the growing evidence of clinical trials showing reversal of cholastasis (bilirubin and γ GT) support the conclusion of the beneficial effects of fish oil or fish oil-containing lipid solutions on PNLD. Although studies investigating short-time effects of fish oil-based PN (14 days), including our study, show changes that could be interpreted as early signs of cholestasis (i.e., rise in γ GT) (22,23), it has to be stated that most of these observed changes were within or near ranges considered normal. Whether this rise in γ GT is of clinical relevance might be investigated in a long-time follow-up cohort study in patients who have been treated with standard soybean-based emulsions for short periods of time (e.g.
