Post-Transplant Conditions

Oral eicosapentaenoic acid for complications of bone marrow transplantation

Article metrics


The ‘systemic inflammatory response syndrome’ (SIRS) may represent the underlying cause of complications after bone marrow transplantation (BMT). This study was conducted to determine whether blocking the etiologic factors of SIRS could improve the complications of BMT. Sixteen consecutive patients with unrelated donors were allocated alternately to two groups. Seven patients received 1.8 g/day of eicosapentaenoic acid (EPA) orally from 3 weeks before to about 180 days after transplantation, while nine patients did not. These two groups were compared with respect to complications, survival, and various cytokines and factors causing vascular endothelial damage. All seven patients receiving EPA survived and only two had grade III graft-versus-host disease (GVHD). Among the nine patients not receiving EPA, three had grade III or IV GVHD. In addition, thrombotic microangiopathy developed in four patients and cytomegalovirus disease occurred in four. Five patients died in this group. The levels of leukotriene B4, thromboxane A2, and prostaglandin I2 were significantly lower in patients receiving EPA than in those not receiving it (all P < 0.01). Cytokines such as tumor necrosis factor-α, interferon-γ, and interleukin-10 were also significantly decreased by EPA (P < 0.05), as were factors causing vascular endothelial damage such as thrombomodulin and plasminogen activator inhibitor-1 (P < 0.05). The survival rate was significantly higher in the group given EPA (P < 0.01). EPA significantly reduced the complications of BMT, indicating that these complications may be manifestations of the systemic inflammatory response syndrome. Bone Marrow Transplantation (2001) 28, 769–774.


Bone marrow transplantation (BMT) has become an increasingly common therapy not only for malignant hematopoietic diseases but also for autoimmune diseases. However, the complications of BMT still pose a serious problem. In our recent report,1 we raised the possibility that various complications of BMT may arise from a single pathological entity called the systemic inflammatory response syndrome. We also suggested that suppression of inflammatory cytokines, such as tumor necrosis factor (TNF)-α and interferon (IFN)-γ, as well as prevention of vascular endothelial damage were likely to be effective methods for preventing complications after BMT. Eicosapentaenoic acid (EPA) has been reported to be effective in the suppression of inflammatory cytokines and prevention of vascular endothelial damage without serious side-effects.2,3 The present study was performed in patients who received BMT from unrelated donors to assess the effectiveness of prophylactic oral EPA therapy in preventing complications.

Materials and nethods


After consent was obtained, patients were randomized to receive the drug or not receive it by the envelope method. For randomization, four envelopes were provided, consisting of two containing a sheet of paper indicating assignment to the drug and two indicating assignment to no drug. The investigator opened one envelope at a time and assigned the patient to the group that was indicated. This trial is a single-center, randomized, parallel-group, controlled study. The subjects were 16 consecutive patients who underwent unrelated allogeneic BMT at our institution from June 1998 to December 1999. The patients were randomized to a group that received EPA (n = 7) and a group that did not (n = 9) (EPA and non-EPA groups, respectively). EPA was administered orally at a dose of 600 mg three times daily (1.8 g/day) from day 21 before BMT to around day 180 after BMT. Table 1 shows the age, gender, underlying disease, and disease stage of all patients. All of the donors were unrelated, but the human leukocyte antigen (HLA) match was complete.

Table 1 Clinical characteristics of the EPA and non-EPA groups

Management of BMT

All patients received the same conditioning regimen, the same prophylaxis for graft-versus-host disease,4 and the same prophylaxis for cytomegalovirus infection.5 For antimicrobial therapy before and after BMT, polymyxin B sulfate (3 000 000 units), an oral solution of vancomycin hydrochloride (1500 mg), amphotericin B syrup (1200 mg), oral fluconazole (300 mg), and inhalation of amphotericin B (several times) were administered between around days −21 and +30. Administration of sulfamethoxazole-trimethoprim was started around day −10, while intravenous infusion of antibiotics and acyclovir was started around day −7. Based on clinical manifestations such as fever and laboratory data such as the C-reactive protein (CRP) and β-D glucan levels,6 the antibiotics and antifungal agents were adjusted as necessary. Granulocyte colony-stimulating factor (G-CSF) (lenograstim (5 μg/kg), filgrastim (5 μg/kg), or nartograstim (8 μg/kg)) was administerd to all patients from day 5 until the WBC exceeded 10 × 109/l.

Diagnosis of complications

The classification of acute GVHD was based on the diagnostic criteria of the 1994 Consensus conference on acute GVHD grading,7 and the highest grade up to day 100 after transplantation was determined. Cytomegalovirus (CMV) interstitial pneumonitis was diagnosed by the following five criteria: (1) symptoms of fever, nonproductive cough, and tachypnea, (2) hypoxia, (3) radiological findings suggestive of interstitial pneumonia, (4) detection of CMV-DNA in the bronchoalveolar lavage fluid, and (5) no evidence for another cause of pneumonia. CMV encephalitis was diagnosed by the following five criteria: (1) fever, (2) convulsions or disturbance of consciousness, (3) findings suggesting encephalitis on magnetic resonance imaging or single photon emission computed tomography, (4) detection of CMV-DNA in cerebrospinal fluid, and (5) no evidence for another cause of encephalitis. A diagnosis of thrombotic microangiopathy was made when the reticulocyte count increased and the platelet count decreased in addition to criteria corresponding to grade 2 or higher in the classification of Zeigler et al8 (fragmented erythrocytes comprising at least 1.3% of 2000 counted peripheral blood cells and elevation of LDH).

Blood sampling and assays

Blood samples for all assays were collected into tubes containing 3.8% citric acid at a ratio of 1:9. After separation by centrifugation at 1500 r.p.m. for 10 min at 4°C, plasma was stored at −80°C. All parameters were measured at the following times: (1) before conditioning (before EPA therapy), (2) on day 0 of BMT before bone marrow infusion, (3) during the aplastic phase on days 5–10 after BMT, (4) during rapid of the white blood cell count (WBC) recovery on days 11–20, (5) at the time of WBC stabilization after discontinuation of G-CSF on days 21–28, and (6) at the time of maximum symptoms.

Assay of cytokines

The levels of the following cytokines were investigated: tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), interleukin (IL)-1β, IL-2, IL-6, IL-8, IL-10, IL-12, and intercellular adhesion molecule-1 (ICAM-1). All cytokines were measured in duplicate each time using ELISA kits (Endogen, Woburn, MA, USA) and the mean values were determined.

Assay of other parameters

Plasminogen activator inhibitor type 1 (PAI-1) activity was determined using a commercially available chromogenic assay (Biopool, Ventura, CA, USA) (normal 4–43 ng/ml). Soluble thrombomodulin was measured by a commercially available sandwich enzyme immunoassay9 based on a monoclonal antibody (Fuji-rebio Co. Ltd, Tokyo, Japan) (normal 4.5 FU/ml). Metabolites of EPA and arachidonic acid such as leukotriene (LT) B4, thromboxane (TX) B2 (the stable metabolite of TXA2), and 6-keto-prostaglandin (PG) F1α (the stable metabolite of PGI2) were determined.

Statistical analysis

Statistical analysis was done using the Mann–Whitney U test or the log-rank test as appropriate, and P < 0.05 was considered to indicate a significant difference.


There were no significant differences in age, sex, underlying disease, disease stage, conditioning regimen, and GVHD prophylaxis between the EPA group and the non-EPA group (Table 1). All of the 16 recipients and their respective donors were positive for anti-CMV IgG antibodies (and were negative for IgM antibodies). Consequently, it seemed unlikely that there was a significant difference between the two groups in risk factors for CMV disease. Among the nine patients not receiving EPA therapy, three developed severe GVHD (grade III or IV), three had grade II GVHD, four had thrombotic microangiopathy, four had CMV disease (all four had interstitial pneumonitis and two had encephalitis), and two had no complications.

Among the seven patients receiving EPA therapy, two had grade III GVHD, one had grade II GVHD, and the others had no complications (Table 2). In the non-EPA group, five patients died whereas all of seven patients in the EPA group survived. When survival curves were calculated by the Kaplan–Meier method, the outcome was significantly better in the EPA group (P < 0.01; log-rank test, Figure 1). There were no significant differences between the two groups in terms of length of febrile period and maximum CRP level, and the regimen for prophylaxis of infection after BMT was the same in all patients (Table 2). The changes in various cytokines were studied from before conditioning (before starting EPA) until the time of maximum symptoms. The EPA group showed significantly lower levels of TNF-α, IFN-γ, and IL-10 compared with the non-EPA group (P < 0.05; t-test) (Figures 24). There were no significant differences between the two groups with regard to the other cytokines and adhesion molecules studied (data not shown). In the non-EPA group, the thrombomodulin level was significantly higher at the time of maximum symptoms (P < 0.05) and the PAI-1 level was significantly higher during the recovery phase (P < 0.05) as well as at the time of maximum symptoms (P < 0.05) compared with the levels in the EPA group (Figures 5 and 6). The changes in metabolites of arachidonic acid and EPA were also compared between the EPA and non-EPA groups. LTB4, TXB2, and 6-keto-PGF1α levels were significantly lower at the time of maximum symptoms in the EPA group compared with the non-EPA group (LTB4; P < 0.01, TXB2, P < 0.01, 6-keto-α PG, P < 0.01) (Figures 79). In the patients who developed grade III GVHD despite administration of EPA, the levels of TNF-α, IFN-γ, and IL-10 were increased, along with an elevation in LTB4, TXB2, 6-keto-PG F1α, PAI-1, and thrombomodulin. In the patients without complications from the non-EPA group, no elevation of these cytokines and arachidonic acid metabolites was noted, and the PAI-1 and thrombomodulin levels were also low.

Table 2 Outcome of the EPA and non-EPA groups
Figure 1

Survival curves for the EPA and non-EPA groups. Survival was significantly better in the EPA group (P < 0.01, log-rank test).

Figure 2

Profile of TNF-α in the two groups. The times of examination are explained in Materials and methods. Data are shown as the mean ± standard deviation (*, Mann–Whitney U test).

Figure 4

Profile of IL-10 in the two groups. Data are shown as the mean ± standard deviation. (*, Mann–Whitney U-test).

Figure 5

Thrombomodulin (TM) levels in the two groups. Data are shown as the mean ± standard deviation (*, Mann–Whitney U test).

Figure 6

PAI-1 levels in the two groups. Data are shown as the mean ± standard deviation (*, Mann–Whitney U-test).

Figure 7

Levels of LTB4 in the two groups. Data are shown as the mean ± standard deviation (*, Mann–Whitney U test).

Figure 9

Levels of 6-keto-α PG F1α (6 KP) in the EPA and non-EPA groups. Data are shown as the mean ± standard deviation. (*, Mann–Whitney U-test).


The present study was planned as a larger trial, but because a significant difference in survival appeared between the two groups at an early stage, the interim results are being reported here.

Among the various complications of BMT, it is known that GVHD,10,11 thrombotic microangiopathy,12 CMV disease,13 respiratory dysfunction resembling adult respiratory disease syndrome,14 and central nervous system dysfunction15 are accompanied by vascular endothelial dysfunction. We have previously shown that certain cytokines, chemokines, and adhesion molecules participate in the development of endothelial damage in each of these complications.10,11,12,13,14,15 In addition, we have reported that these various complications can all be considered as manifestations of the systemic inflammatory response syndrome (SIRS).1 In this syndrome, dysfunction of multiple organs is caused by vascular endothelial damage arising from increased production of inflammatory cytokines induced by certain stimuli or inflammation.16,17 Because our hypothesis suggests that vascular endothelial dysfunction in various target organs may cause the complications of BMT, these complications should be substantially reduced if SIRS can be prevented, ie if suppression of the post-BMT increase of inflammatory cytokines can prevent the development of vascular endothelial damage. EPA is reported to be effective in suppressing both increase of inflammatory cytokines and vascular endothelial damage, and also causes few side-effects.2,3 The present study was performed to assess whether EPA could prevent the complications of BMT.

EPA is known to suppress the production of inflammatory cytokines. Arachidonic acid in the membranes of white blood cells is metabolized to LTB4, which promotes the migration, adhesion, and aggregation of leukocytes. LTB4 also has various potent pro-inflammatory actions, including the release of lysosome, activation of natural killer cells, and promotion of the production of TNF-α, IL-2, and IFN-γ.18 Like arachidonic acid, EPA is metabolized by 5-lipoxygenase, but it forms LTB5 instead of LTB4. LTB5 also binds to the LTB4 receptor and it also activates leukocytes and promotes the development of inflammation, but it is far less damaging than LTB4.19 Because reagents for an LTB5 assay did not exist, only LTB4 was measured in the present study. The LTB4 level was significantly lower in the EPA group than in the non-EPA group (P < 0.01), indicating that EPA suppressed LTB4 production. It has been reported that EPA directly suppresses the production of inflammatory cytokines such as TNF and IL-1,2 while it also indirectly decreases the levels of inflammatory cytokines such as IL-6 and IL-2 through stimulation of concanavalin A and phytohemagglutinin,18,20 and scavenges free radicals.21 In the present study, the levels of cytokines such as TNF-α, IFN-γ, and IL-10 were significantly decreased in the EPA group, while the levels of IL-1β, IL-2, and IL-6 were not significantly different between the EPA and non-EPA groups. Thus, by exerting a direct anti-inflammatory effect and by acting through LTB4, EPA caused a significant decrease of the inflammatory cytokines that are thought to be involved in causing various complications of BMT. In patients who developed severe complications during treatment with EPA, the levels of these cytokines were increased despite EPA therapy, while the levels of these parameters were low in patients from the non-EPA group when no complications occurred. These findings support our hypothesis about the role of SIRS in the pathogenesis of complications after BMT.

The inhibition of platelet aggregation by EPA is related to its prevention of vascular endothelial dysfunction. Namely, EPA acts on platelets as well as the endothelial cells and smooth muscle cells in the vessel wall, by replacing arachidonic acid in the cell membrane.18 It has been shown that the incorporation of EPA into platelets and megakaryocytes alters the fatty acid composition of circulating platelets.22 In platelets, arachidonic acid is metabolized by cyclooxygenase to TXA2, which potently enhances platelet aggregation. In the vascular wall, in contrast, it is metabolized by cyclooxygenase to PGI2 that potently inhibits platelet aggregation. Vascular potency and blood flow are maintained by the balance between these two processes. When EPA is metabolized by cyclooxygenase in platelets and the vascular wall, it forms TXA3 and PGI3, respectively. Unlike TXA2, TXA3 does not enhance platelet aggregation,23 while PGI3 is as effective as PGI2 in inhibiting platelet aggregation. Thus, the net effect of EPA is to inhibit platelet aggregation. It is also important that EPA replaces arachidonic acid in the platelet membrane, and thus decreases the pool of this precursor of TXA2. Comparison of the EPA and non-EPA groups in terms of TXB2 and 6-keto PGF1α, which are stable metabolites of TXA2 and PGI2, respectively, showed that the levels of both metabolites were significantly lower in the EPA group (P < 0.01). In patients receiving EPA, adhesion of platelets to the vessel wall was reported to be markedly reduced.24 In addition, it has been reported that EPA can inhibit thrombosis,25 augment endothelial cell-dependent vasodilatation through nitric oxide and PG,26 increase nitric oxide production by endothelial cells,27 and increase the production and release of endothelium-derived relaxing factor that has a local antiplatelet effect in the vascular wall.28 Because of these properties, EPA may have prevented vascular endothelial dysfunction after BMT, thus significantly decreasing the levels of indicators of such dysfunctions like thrombomodulin and PAI-1 in the EPA group. The amelioration of the complications of BMT thus achieved may have contributed to the improvement of survival in this group.

In summary, the present study indicated that inflammatory cytokines increase after BMT, causing progressive vascular endothelial dysfunction which manifests as SIRS. Because SIRS seems to be involved in the pathogenesis of complications after BMT, EPA can reduce such complications by suppressing the two typical mechanisms which lead to SIRS, ie an increase of inflammatory cytokines and progressive vascular endothelial dysfunction. Accordingly, the prophylactic administration of EPA may be effective in reducing the complications of BMT. Outside Japan, EPA is available as fish oil, which is sold as a health food. An increased relapse rate after BMT due to a decreased graft-versus-leukemia effect may not be a problem because the mechanisms related to prevention of complications by EPA do not involve immunological cells directly. However, further studies are necessary to assess both recurrence rate and efficacy in a larger number of patients.

Figure 3

Profile of IFN-γ in the two groups. Data are shown as the mean ± standard deviation (*, Mann–Whitney U-test).

Figure 8

Levels of TXB2 in the two groups. Data are shown as the mean ± standard deviation (*, Mann–Whitney U-test).


  1. 1

    Takatsuka H, Takemoto Y, Yamada S et al. Complications after bone marrow transplantation are manifestation of systemic inflammatory syndrome Bone Marrow Transplant 2000 26: 419–426

  2. 2

    Endres S, Ghorbani R, Kelley VE et al. The effect of dietary supplementation with n-3 polyunsaturated fatty acid on the synthesis of interleukin-1 and tumor necrosis factor by mononuclear cells N Engl J Med 1989 320: 265–271

  3. 3

    Chetty N, Vickers JD, Kinlough-Rathbone RL et al. Eicosapentaenoic acid interferes with U46619-stimulated formation of inositol phosphates in washed rabbit platelets Thromb Haemost 1989 62: 1116–1120

  4. 4

    Chao NJ, Schmidt GM, Niland JC et al. Cyclosporin, methotrexate and prednisone compared with cyclosporin and prednisone for prophylaxis of acute graft-versus-host disease New Engl J Med 1993 329: 1225–1230

  5. 5

    Singhal S, Mehta J, Powles R et al. Three weeks of ganciclovir for cytomegaloviraemia after allogeneic bone marrow transplantation Bone Marrow Transplant 1995 15: 777–781

  6. 6

    Obayashi T, Yoshida M, Mori T et al. Plasma (1→3)-β-D-glucan measurement in diagnosis of invasive deep mycosis and fungal febrile episodes Lancet 1995 345: 17–20

  7. 7

    Przepiorka D, Weisdorf D, Martin P et al. 1994 Consensus conference on acute GVHD grading Bone Marrow Transplant 1995 15: 825–828

  8. 8

    Zeigler ZR, Shadduck RK, Nemunaitis J et al. Bone marrow transplant-associated thrombotic microangiopathy: a case series Bone Marrow Transplant 1995 15: 247–253

  9. 9

    Nunberger W, Michelmann I, Burdach S, Guel U . Endothelial dysfunction after bone marrow transplantation: increase of soluble thrombomodulin and PAI-1 in patients with multiple transplant-related complications Ann Hematol 1998 76: 61–65

  10. 10

    Takatsuka H, Takemoto Y, Okamoto T et al. Predicting the severity of graft-versus-host disease from interleukin-10 levels after bone marrow transplantation Bone Marrow Transplant 1999 24: 1005–1007

  11. 11

    Takatsuka H, Takemoto Y, Okada M et al. Changes of cytokines during the course of graft-versus-host disease following bone marrow transplantation: A case report Cytokine 2000 12: 1225–1227

  12. 12

    Takatsuka H, Takemoto Y, Okamoto T et al. Thrombotic microangiopathy following allogeneic bone marrow transplantation Bone Marrow Transplant 1999 24: 303–306

  13. 13

    Yamada S, Takatsuka H, Takemoto Y et al. Association of cytomegalovirus interstitial pneumonitis with HLA type following allogeneic bone marrow transplantation Bone Marrow Transplant 2000 25: 861–865

  14. 14

    Takatsuka H, Takemoto Y, Okamoto T et al. Adult respiratory distress syndrome-like disorders after allogeneic bone marrow transplantation Transplantation 1999 68: 1343–1347

  15. 15

    Takatsuka H, Takemoto Y, Okamoto T et al. The levels of soluble P-selectin, von Willebrand factor and thrombomodulin in patients with neurological complications after allogeneic bone marrow transplantation Bone Marrow Transplant 1998 21: 809–813

  16. 16

    Chen X, Christou NV . Relative contribution of endothelial cell and polymorphonuclear neutrophil activation in their interactions in systemic inflammatory response syndrome Arch Surg 1996 131: 1148–1154

  17. 17

    Members of the American College of Chest Physicians/Society of Critical Care Medicine ConsensusConference Committee: American college of chest physicians/society of critical care medicine consensus conference . Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis Crit Care Med 1922 20: 864–874

  18. 18

    Kinsella JE, Lokesh B, Broughton S, Whelan J . Dietary polyunsaturated fatty acids and eicosanoids: potential effects on the modulation of inflammatory and immune cells: an overview Nutrition 1990 6: 24–44

  19. 19

    Terano T, Salmon JA, Moncada S . Biosynthesis and biological activity of leukotrien B5 Prostaglandins 1984 27: 217–232

  20. 20

    Meydani SN, Endres S, Woods MM et al. Oral (n-3) fatty acid supplementation suppresses cytokine production and lymphocyte proliferation: comparison between young and older women J Nutr 1991 121: 547–555

  21. 21

    Fisher M, Upchurch KS, Levine PH et al. Effects of dietary fish oil supplementation on polymorphonuclear leukocyte inflammatory potential Inflammation 1986 10: 387–392

  22. 22

    Schick PK, Menon S, Wojenski C . Effects of marine oil-enriched diets on guinea pig megakaryocyte and platelet lipids: effects on thromboxane synthesis and platelet function Biochim Biophys Acta 1990 1022: 49–56

  23. 23

    Needleman P, Raz A, Minkes MS et al. Triene prostaglandins: prostacyclin and thromboxane biosynthesis and unique biological properties Proc Natl Acad Sci USA 1979 76: 944–948

  24. 24

    Tremoli E, Eligini S, Colli S et al. n-3 Fatty acid ethyl ester administration to healthy subjects and to hypertriglyceridemic patients reduces tissue factor activity in adherent monocytes Arterioscler Thromb 1994 14: 1600–1608

  25. 25

    Harker LA, Kelly AB, Hanson SR et al. Interruption of vascular thrombus formation and vascular lesion formation by dietary n-3 fatty acids in fish oil in nonhuman primates Circulation 1993 87: 1017–1029

  26. 26

    Shimokawa H, Vanhoutte PM . Dietary omega 3 fatty acids and endothelium-dependent relaxations in porcine coronary arteries Am J Physiol 1989 256: H968–973

  27. 27

    Okuda Y, Kawashima K, Sawada T et al. Eicosapentaenoic acid enhances nitric oxide production by cultured human endothelial cells Biochem Biophys Res Commun 1997 232: 487–491

  28. 28

    Lawson DL, Mehta JL, Saldeen K et al. Omega-3 polyunsaturated fatty acids augment endothelium-dependent vasorelaxation by enhanced released of EDRF and vasodilator prostaglandins Eicosanoids 1991 4: 217–223

Download references


We wish to thank Ms A Utsumi and Ms Y Shikita for their expert technical assistance.

Author information

Correspondence to H Takatsuka.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Takatsuka, H., Takemoto, Y., Iwata, N. et al. Oral eicosapentaenoic acid for complications of bone marrow transplantation. Bone Marrow Transplant 28, 769–774 (2001) doi:10.1038/sj.bmt.1703226

Download citation


  • bone marrow transplantation
  • systemic inflammatory response syndrome
  • eicosapentaenoic acid
  • cytokine
  • endothelial damage

Further reading