Clinical review

Science, medicine, and the future Pathogenesis of sepsis: new concepts and implications for future treatment
Pierre-Yves Bochud, Thierry Calandra
Severe sepsis and septic shock are important causes of death in intensive care units. Although our understanding of the pathogenesis of inflammation and sepsis has improved, until recently this has not translated into clinical benefit. Several new treatment approaches have given encouraging results. Evidence suggests that the way forward is to develop pathogen specific regimens rather than assume that one treatment fits all.

Institute for Systems Biology, 1441 North 34th Street, Seattle, WA 98103-8904, USA Pierre-Yves Bochud research fellow Division of Infectious Diseases, Centre Hospitalier Universitaire Vaudois, CH-1011 Lausanne, Switzerland Thierry Calandra associate professor Correspondence to: T Calandra Thierry.Calandra@ chuv.hospvd.ch
BMJ 2003;326:2626

Summary points
Bacterial cell walls, endotoxins, and exotoxins are powerful activators of innate and acquired immune responses Molecules expressed by pathogens interact with Toll-like receptors on immune cells, activating the immune response Cytokines are important in the pathogenesis of sepsis Susceptibility to sepsis may be due to inherited or acquired mutations of innate immune genes Severe sepsis and septic shock are clinical manifestations of a dysregulated immune response to invasive pathogens Adjunctive therapy with low dose steroids, activated protein C or early supportive care can reduce mortality from severe sepsis and septic shock Pathogen recognition receptors (such as Toll-like receptors) and mediators of sepsis (such as macrophage migration inhibitory factor) might be novel targets for treatment

Sources and methods

We selected articles for this review by searching Medline using the keywords sepsis, therapy, and Toll-like receptors. We concentrated on publications on the pathogenesis of sepsis and treatment of septic shock. As the number of references that could be cited was limited, we have often referenced review articles rather than original publications.

Epidemiology and importance of severe sepsis and septic shock

Severe sepsis and septic shock are life threatening complications of infections and the most common cause of death in intensive care units. However, a lack of widely accepted definitions of these complications has made it difficult to obtain accurate estimates of their frequency. A study published by the Centers for Disease Control in the United States indicated that the incidence of septicaemia had increased from 73.6 per 100 000 patients in 1979 to 175.9 per 100 000 patients in 1987.1 Recent US and European surveys have estimated that severe sepsis accounts for 2-11% of all admissions to hospital or intensive care units.2 Although Gram negative infections were predominant in the 1960s and early 1970s, Gram positive infections have increased in the past two decades and now account for about half of cases of severe sepsis.3 Fungal infections are also increasing in many countries. Despite better supportive care, the hospital mortality from severe sepsis and septic shock (30% and over 60%, respectively) has not changed much over recent decades.

Further details about pattern recognition receptors and genetic susceptibility are available on bmj.com

Innate immune responses to microbial products

The innate immune system is the first line of defence against infection and is activated when a pathogen

crosses the hosts natural defence barriers.4 It consists of soluble elements (the alternative and mannanbinding lectin pathways of the complement system, acute phase proteins, and cytokines) and cellular elements (monocytes, macrophages, neutrophils, dendritic cells, and natural killer cells). Innate immune responses must be tightly regulated as unbalanced inflammatory and immune reactions can result in either uncontrolled microbial growth or devastating inflammatory responses with tissue injury, vascular collapse, and multiorgan failure. Detection of invading microorganisms is mediated by pattern recognition receptors expressed on the surface of innate immune cells (figure). Pattern recognition receptors recognise structures common to many
BMJ VOLUME 326 1 FEBRUARY 2003 bmj.com

262

Clinical review
microbial pathogens. These structures are called pathogen associated molecular patterns and include endotoxins (lipopolysaccharide), peptidoglycan, lipoteichoic acid, lipopeptides, flagellin, mannan, and viral RNA. The structures are essential for survival of the microorganisms and therefore do not undergo major mutations. When a pathogen associated molecular pattern binds to a pattern recognition receptor, it activates several intracellular signalling pathways resulting in the activation of transcription factors (NF- B, AP-1, Fos, Jun). The transcription factors control the expression of immune response genes and the release of numerous effector molecules, such as cytokines. Cytokines have an essential role in orchestrating the innate and acquired immune responses to an invading pathogen.5

Gram negative bacteria

Gram positive bacteria

Lipopolysaccharide TLR4 MD-2 CD14

CpG DNA TLR9

Flagellin TLR5

Peptidoglycan TLR6-TLR2

Lipoteichoic acid TLRX-TLR2

CD14

Bacterial sepsis
Gram negative bacilli (mainly Escherichia coli, Klebsiella species, and Pseudomonas aeruginosa) and Gram positive cocci (mainly staphylococci and streptococci) are the commonest microbes isolated from patients with severe sepsis and septic shock.3 Fungi, mostly Candida, account for only about 5% of all cases of severe sepsis. Gram negative sepsis Most cases of Gram negative sepsis are caused by Enterobacteriaceae such as E coli and Klebsiella species. Pseudomonas aeruginosa is the third commonest cause. Gram negative infections usually occur in the lung, abdomen, bloodstream, or urinary tract. Lipopolysaccharide is an important component of the outer membrane of Gram negative bacteria and has a pivotal role in inducing Gram negative sepsis.6 Lipopolysaccharide binding protein in host cells binds to lipopolysaccharide in the bacteria and transfers it to CD14.7 CD14 is a protein anchored in the outer leaflet of the plasma membrane, although it also exists as a soluble plasma protein that attaches lipopolysaccharide to CD14-negative cells, such as endothelial cells. CD14 is located in the extracellular space and therefore cannot induce cellular activation without a transmembrane signal transducing coreceptor. A series of remarkable investigations have recently led to the identification of Toll-like receptor 4 (TLR4) as the coreceptor for lipopolysaccharide. Toll receptors were first discovered in Drosophila, where they were found to have a role in the defence of flies against fungi and Gram positive bacteria.8 Toll-like receptors were then identified in other species. Human Toll-like receptors, like their homologues in insects and other mammalian species, are type I transmembrane proteins with an extracellular leucine rich repeat domain and an intracellular domain homologous to the interleukin 1 receptor. Genetic studies in mice showed that mutations in the Tlr4 gene were linked to resistance to lipopolysaccharide, providing evidence that TLR4 was an essential component of the lipopolysaccharide receptor complex.9 MD-2, a secreted protein associated with the extracellular domain of TLR4, has also recently been shown to have an important role in responsiveness to lipopolysaccharide.10 Gram positive sepsis Staphylococci (mainly Staph aureus and coagulasenegative staphylococci) and streptococci (Strep pyogenes, viridans streptococci, Strep pneumoniae) are the
BMJ VOLUME 326 1 FEBRUARY 2003 bmj.com

Specific signal

Common signal

Specific signal

Response 1

Response 2

Interaction between bacterial products and pattern recognition receptors expressed on immune cells. Components of bacterial cell walls (such as lipopolysaccharide, peptidoglycan, lipoteichoic acid, flagellin, and unmethylated CpG DNA sequences) interact with specific Toll-like receptors (TLR) expressed on immune cells. The receptors then activate intracellular signalling pathways and transcription factors resulting in expression of the gene for immune response

commonest causes of Gram positive sepsis. They are usually responsible for infections of skin and soft tissue, infections associated with intravascular devices, primary bloodstream infections, or respiratory infections. Gram positive organisms can cause sepsis by at least two mechanisms: by producing exotoxins that act as superantigens (see definition below) and by components of their cell walls stimulating immune cells.11 Superantigens are molecules that bind to MHC class II molecules of antigen presenting cells and to V chains of T cell receptors. In doing so, they activate large numbers of T cells to produce massive amounts of proinflammatory cytokines. Staphylococcal enterotoxins, toxic shock syndrome toxin-1, and streptococcal pyrogenic exotoxins are examples of bacterial superantigens. Gram positive bacteria without exotoxins can also induce shock, probably by stimulating innate immune responses through similar mechanisms to those in Gram negative sepsis. Indeed, Toll-like receptor 2 (TLR2) has been shown to mediate cellular responses to heat killed Gram positive bacteria and their cell wall structures (peptidoglycan, lipoproteins, lipoteichoic acid, and phenol soluble modulin).12

Pathways to sepsis
The clinical manifestations of sepsis produced by different Gram positive and Gram negative bacteria vary. For example, the clinical pictures of streptococcal toxic
263

Clinical review
shock syndrome and meningococcaemia are very different. In addition, E coli urosepsis follows a more benign course than nosocomial pneumonia due to P aeruginosa. Moreover, Gram positive and Gram negative sepsis result in different expression and release of proinflammatory mediators, such as the cytokine tumour necrosis factor- .13 These observations suggest that there are specific host immune responses to each pathogen mediated by various sets of pathogen associated molecular patterns and pattern recognition receptors. Toll-like receptors Of the 10 human Toll-like receptors identified so far, seven interact with microbial motifs (table 1).4 For example, TLR2 binds components of the cell wall of Gram positive bacteria as well as ligands derived from other pathogens, TLR5 is the receptor for bacterial flagellin,14 and TLR9 is required for cellular activation by unmethylated CpG motifs of bacterial DNA.15 Cooperation between Toll-like receptors is necessary to respond to certain pathogens, such as Gram positive bacteria and yeast (zymosan).16 Several signal transducing pathways are activated after microbial ligands bind to Toll-like receptors (figure).17 The fact that different microbial products bind to different Toll-like receptors, the existence of receptor specific signalling pathways, and the idea of differential expression of Toll-like receptors by tissues and organs strongly suggest that the innate immune system is tailored in a pathogen and tissue specific manner. Expression of immune genes and host responses to infections will vary depending on the structural and biochemical composition of the invading pathogen. If confirmed, these hypotheses point to the need to develop pathogen specific approaches to treatment. Other soluble and membrane associated proteins have recently been shown to be involved in recognising bacteria or microbial products. These include peptidoglycan recognition proteins and triggering receptor expressed on myeloid cells (TREM-1). Additional information on these proteins is available on bmj.com
Table 1 Ligands for human Toll-like receptors and their sources
Receptors TLR1-TLR2 TLR2-TLRX Ligands Soluble factors released by live bacteria Lipoproteins Lipoteichoic acid Lipoarabinomannan Phosphatidylinositol dimannoside Glycosylphosphatidylinositol anchors Endotoxin (LPS) TLR2-TLR6 Peptidoglycan Soluble phenol modulin Zymosan Macrophage activating lipopeptide-2 (MALP-2) TLR3 TLR4 TLR5 TLR7 TLR9 Double-stranded RNA Endotoxin (lipopolysaccharide) Taxol Flagellin Imidazoquinoline antiviral compounds (imiquimod and R-848) Unmethylated GpG DNA Source of ligand Neisseria meningitidis Several bacterial species Staphylococcus aureus Mycobacteria Staph aureus Trypanosoma cruzi Leptospira interrogans, Porphyromonas gingivalis Gram positive bacteria Staph aureus Yeasts Mycoplasma fermentans Virus Gram negative bacteria Plants Flagellated bacteria Chemical compounds Bacteria

Adjunctive therapies for sepsis

Numerous adjunctive treatments (that is, other than antibiotics and supportive care) for severe sepsis and septic shock have been tested in clinical trials (table 2). These include neutralisation of microbial toxins such as lipopolysaccharide, non-specific anti-inflammatory and immunosuppressive drugs, neutralisation of pro-inflammatory cytokines, and correction of abnormalities in coagulation. The results have been mixed,18 although several recent clinical trials have given encouraging results. Coagulation abnormalities, especially disseminated intravascular coagulation, are common in patients with sepsis and microvascular thrombosis. The ensuing tissue damage may have an important role in the pathophysiology of organ dysfunction. Treatment with activated protein C, a protein that has anti-thrombotic, pro-fibrinolytic, and anti-inflammatory effects, reduces mortality from severe sepsis at the price of a slight increase in bleeding events.19 Glucocorticoids exert broad metabolic and immunomodulating effects and have been used to treat several inflammatory diseases. Although high doses of steroids have no clinical benefit,18 a recent multicentre trial found that a seven day course of low doses of hydrocortisone and fludrocortisone reduced mortality in patients with septic shock and relative adrenal insufficiency.20 Finally, two studies of supportive care, one focusing on early therapy with fluids, vasopressors, and transfusions and the other on meticulous control of glycaemia with insulin, have shown reduced mortality in patients with severe sepsis and septic shock.21 22

Future treatment strategies

Microbial drugs and pattern recognition molecules Designing new drugs to neutralise microbial products or block their interaction with specific receptor on immune cells is an attractive concept. Potential targets include lipolysaccharide binding protein, CD14, TLR4, and MD-2 for Gram negative sepsis, and CD14, TLR2, and TLR6 for Gram positive sepsis. Monoclonal antibodies against CD14 are being evaluated in phase II studies. Several intracellular signalling molecules, such as MyD88 and the mitogen-activated protein kinase are other possible therapeutic targets. However, inactivating molecules that are pivotal to innate immunity can be harmful, as shown by the increased sensitivity to bacterial sepsis in mice with mutations of the Tlr4 gene.23 Careful selection of patients with severe infections associated with a high probability of death will therefore be essential. Macrophage migration inhibitory factor Macrophage migration inhibitory factor is a cytokine that has recently been shown to be important in innate immunity and sepsis.24 It is constitutively expressed in large amounts by immune, endocrine, and epithelial cells and is rapidly released after exposure to microbial products and pro-inflammatory cytokines. Macrophage migration inhibitory factor regulates innate immune responses to endotoxin and Gram negative bacteria by modulating the expression of TLR4, enabling macrophages and other cells at the front line of defences to respond quickly.25 High levels of macrophage migration inhibitory factor have been detected
BMJ VOLUME 326 1 FEBRUARY 2003 bmj.com

264

Clinical review

Table 2 Selected antibacterial, anti-inflammatory, and immunomodulating adjunctive therapies investigated in patients with severe sepsis and septic shock
Type of therapy Neutralisation of microbial toxins Non-specific anti-inflammatory and immunomodulating drugs Inhibition of specific mediators Target (s) Endotoxin Multiple inflammatory and immune mediators Pro-inflammatory cytokines: Tumour necrosis factor Interleukin-1 Phospholipid components: Phospholipase A2 Cyclo-oxygenase Thromboxane Platelet activating factor Oxygen free radicals Nitric oxide Bradykinin Correction of coagulopathy Other Coagulation cascade Phospholipase A2 inhibitor Ibuprofen Dazoxiben, ketoconazole Platelet activating factor antagonists, platelet activating factor acetylhydrolase N-acetylcysteine, selenium N-methyl-L-arginine Bradykinin antagonist Antithrombin III, tissue factor pathway inhibitor, activated protein C Prostaglandin E1, granulocyte colony stimulation factor Anti-tumour necrosis factor antibodies, soluble tumour necrosis factor receptors Interleukin-1 receptor antagonist Agents Anti-endotoxin antibodies, anti-lipid A antibodies, lipopolysaccharide analogues, lipopolysaccharide removal High dose corticosteroids, low dose corticosteroids, pentoxifylline, immunoglobulins, interferon gamma

in patients with inflammatory and infectious diseases, including severe sepsis and septic shock.26 Immunoneutralisation of macrophage migration inhibitory factor or deletion of the Mif gene protects mice against lethal endotoxaemia, Gram positive toxic shock syndromes, and experimental bacterial peritonitis. Conversely, mice injected with macrophage migration inhibitory factor together with live bacteria or microbial toxins have increased death rates.2628 This factor thus has the potential to endanger life when expressed in excess during sepsis. Development of drugs to block the production of macrophage migration inhibitory factor or inhibit its function may help treat severe sepsis and other inflammatory diseases. High mobility group protein 1 High mobility group protein 1, a protein previously known as DNA binding protein regulating gene transcription and stabilising nucleosome formation, has recently been described as a late mediator of inflammation and sepsis.29 Patients with septic or haemorrhagic shock have raised serum concentrations of high mobility group protein 1, and concentrations are associated with patients outcome. Use of polyclonal antibodies to block high mobility group protein 1 in mice protects them against lipopolysaccharide induced acute lung injury and lethal endotoxaemia.30 Genetic studies of susceptibility to sepsis Several gene polymorphisms have been associated with increased susceptibility to sepsis (see bmj.com for more information). Testing for polymorphisms of important genes may help to identify people who are at increased risk of sepsis when exposed to virulent bacteria and who may benefit from targeted immunomodulatory therapies.
Funding: TC and P-YB are supported by grants from the Swiss National Science Foundation (31-066972.01 and 81LA-65462). TC is the recipient of a career award from the Leenaards Foundation. Competing interests: TC has been reimbursed for travel expenses and received fees for speaking at conferences organised by Eli Lilly, the manufacturer of Xigris, recombinant activated protein C.

3 4 5

6 7 8 9

10

11 12

13

14

15 16

17 18

19

20

21

22

23

Centers for Disease Control. Increase in national hospital discharge survey rates for septicemiaUnited States, 1979-1987. JAMA 1990;263:937-8. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med 2001;29:1303-10. Bochud PY, Glauser MP, Calandra T. Antibiotics in sepsis. Intensive Care Med 2001;27(suppl 1):S33-48. Janeway CA Jr, Medzhitov R. Innate immune recognition. Annu Rev Immunol 2002;20:197-216. Calandra T, Bochud PY, Heumann D. Cytokines in septic shock. In: Remington JS, Swartz MN, eds. Current clinical topics in infectious diseases. Oxford: Blackwell Publishing, 2002:1-23. Alexander C, Rietschel ET. Bacterial lipopolysaccharides and innate immunity. J Endotoxin Res 2001;7:167-202. Ulevitch RJ, Tobias PS. Recognition of gram-negative bacteria and endotoxin by the innate immune system. Curr Opin Immunol 1999;11:19-22. Hoffmann JA, Reichhart JM. Drosophila innate immunity: an evolutionary perspective. Nat Immunol 2002;3:121-6. Poltorak A, He X, Smirnova I, Liu MY, Huffel CV, Du X, et al. Defective LPS signalling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998;282:2085-8. Nagai Y, Akashi S, Nagafuku M, Ogata M, Iwakura Y, Akira S, et al. Essential role of MD-2 in LPS responsiveness and TLR4 distribution. Nat Immunol 2002;3:667-72. Calandra T. Pathogenesis of septic shock: implications for prevention and treatment. J Chemother 2001;13:173-80. Takeuchi O, Hoshino K, Kawai T, Sanjo H, Takada H, Ogawa T, et al. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity 1999;11:443-51. Cohen J, Abraham E. Microbiologic findings and correlations with serum tumor necrosis factor-alpha in patients with severe sepsis and septic shock. J Infect Dis 1999;180:116-21. Hayashi F, Smith KD, Ozinsky A, Hawn TR, Yi EC, Goodlett DR, et al. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 2001;410:1099-103. Hemmi H, Takeuchi O, Kawai T, Kaisho T, Sato S, Sanjo H, et al. A Tolllike receptor recognizes bacterial DNA. Nature 2000;408:740-5. Ozinsky A, Underhill DM, Fontenot JD, Hajjar AM, Smith KD, Wilson CB, et al. The repertoire for pattern recognition of pathogens by the innate immune system is defined by cooperation between toll-like receptors. Proc Natl Acad Sci USA 2000;97:13766-71. Kaisho T, Akira S. Toll-like receptors as adjuvant receptors. Biochim Biophys Acta 2002;1589:1-13. Vincent JL, Sun Q, Dubois MJ. Clinical trials of immunomodulatory therapies in severe sepsis and septic shock. Clin Infect Dis 2002;34: 1084-93. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699-709. Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B, Korach JM, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 2002;288:862-71. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368-77. Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in the critically ill patients. N Engl J Med 2001;345:1359-67. Cross AS, Sadoff JC, Kelly N, Bernton E, Gemski P. Pretreatment with recombinant murine tumor necrosis factor alpha/cachectin and murine

BMJ VOLUME 326

1 FEBRUARY 2003

bmj.com

265

Clinical review
interleukin 1 alpha protects mice from lethal bacterial infection. J Exp Med 1989;169:2021-7. 24 Froidevaux C, Roger T, Martin C, Glauser MP, Calandra T. Macrophage migration inhibitory factor and innate immune responses to bacterial infections. Crit Care Med 2001;29:S13-5. 25 Roger T, David J, Glauser MP, Calandra T. MIF regulates innate immune responses through modulation of Toll-like receptor 4. Nature 2001;414:920-4. 26 Calandra T, Echtenacher B, Roy DL, Pugin J, Metz CN, Hultner L, et al. Protection from septic shock by neutralization of macrophage migration inhibitory factor. Nat Med 2000;6:164-70. 27 Bernhagen J, Calandra T, Mitchell RA, Martin SB, Tracey KJ, Voelter W, et al. MIF is a pituitary-derived cytokine that potentiates lethal endotoxaemia. Nature 1993;365:756-9. 28 Bozza M, Satoskar AR, Lin G, Lu B, Humbles AA, Gerard C, et al. Targeted disruption of migration inhibitory factor gene reveals its critical role in sepsis. J Exp Med 1999;189:341-6. 29 Wang H, Yang H, Czura CJ, Sama AE, Tracey KJ. HMGB1 as a late mediator of lethal systemic inflammation. Am J Respir Crit Care Med 2001;164:1768-73. 30 Wang H, Bloom O, Zhang M, Vishnubhakat JM, Ombrellino M, Che J, et al. HMG-1 as a late mediator of endotoxin lethality in mice. Science 1999;285:248-51.

Lesson of the week Low dose methotrexate and bone marrow suppression
Michael Sosin, Sunil Handa
Department of Haematology, Sandwell General Hospital, West Bromwich, West Midlands, B71 4HJ Michael Sosin senior house officer Sunil Handa consultant haematologist Correspondence to: Sunil Handa sunil.handa@swellhot. wmids.nhs.uk
BMJ 2003;326:2667

Methotrexate is an antimetabolite that, apart from its use in malignant disorders, is taken orally in low doses for the control of conditions such as rheumatoid arthritis and psoriasis. When used in chemotherapy it causes profound suppression of bone marrow. However, even at a low dose it may be associated with bone marrow suppressionparticularly in the presence of renal insufficiency or when other drugs are taken concomitantly (box 1). Its unusual weekly dosing regimen can result in dose error by patients or clinicians. We present three cases of bone marrow suppression in patients taking low dose methotrexate who presented at a district general hospital during a period of four years (table).

Case reports
Case 1 A 78 year old woman with rheumatoid arthritis had been taking a weekly dose of 17.5 mg of methotrexate for two months. Before this the dose had been gradually built up over several years. She was admitted with breathlessness and found to be pancytopenic (haemoglobin concentration 100 g/l, white cell count 3.0 109/l, neutrophils 2.5 109/l, platelets 13 109/l). A month earlier her full blood count had been normal. Methotrexate treatment was discontinued. She was treated with intravenous folinic acid and antibiotics and was given transfusions of blood products. Her blood count showed recovery (haemoglobin concentration 137 g/l, white cell count 11.0 109/l, neutrophils 8.1 109/l, platelets 178 109/l). After a prolonged admission of about four weeks she was discharged home well. Case 2 A 67 year old man with rheumatoid arthritis had been taking methotrexate for six months. He was admitted

with diarrhoea and vomiting. Both he and his family were unclear about the dose of methotrexate he had taken recently, but it seemed likely that he had been taking methotrexate daily instead of weekly. He was pancytopenic (haemoglobin concentration 102 g/l, white cell count 1.2 109/l, neutrophils 0.3 109/l, platelets 66 109/l). A month previously his blood count showed a neutrophilia (haemoglobin concentration 111 g/l, white cell count 15.5 109/l, neutrophils 13.7 109/l, platelets 393 109/l). His weekly dose of methotrexate had been increased from 5 mg to 7.5 mg at that time, and this may have resulted in the presumed dose error. Methotrexate was discontinued, and he was treated with blood transfusion and with intravenous antibiotics and folinic acid. His blood count recovered (haemoglobin concentration 117 g/l, white cell count 15.6 109/l, neutrophils 8.5 109/l, platelets 153 109/l), and he was discharged home well. Case 3 A 74 year old woman, who had been taking methotrexate (5 mg weekly) for 10 years, was admitted with abdominal pain and pancytopenia (haemoglobin concentration 78 g/l, white cell count 0.5 109/l, neutrophils 0.4 109/l, platelets 14 109/l). She had received a course of trimethoprim for a presumed urinary tract infection one week before admission. Three weeks earlier her blood count had been normal. Methotrexate was discontinued. She was treated with intravenous antibiotics and folinic acid and transfusion of blood products. She did not improve and was started on treatment with granulocyte colony stimulating factor. Her white cell count rose rapidly, and her last blood count showed a neutrophilia (haemoglobin concentration 132 g/l, white cell count 52.0 109/l, neutrophils 48.4 109/l, platelets 190 109/l). After

Details of patients with bone marrow suppression after taking methotrexate

Weekly dose of methotrexate (duration of treatment) 17.5 mg (2 months) 5 mg (6 months) 5 mg (10 years) 10 mg (7 months) Symptoms at presentation to hospital Shortness of breath Diarrhoea and vomiting Abdominal pain Fever and rigors

Patient 1 2 3 4*

Age, sex 78, female 67, male 74, female 40, male

Indication Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis Psoriasis

Precipitating factor Unknown Dose error Concomitant trimethoprim Unprescribed self treatment

Outcome Recovered Recovered Died Recovered

*Not described in the text.

266

BMJ VOLUME 326

1 FEBRUARY 2003

bmj.com