Endoscopic Therapy of Barrett's: What More Do We Need to Know?

COMMENTARY 779

Barrett’s oesophagus Barrett’s segment and normalisation of ...... acid exposure as the only independent

predictive factors for sustained long Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from term re-epithelialisation.78 However, Endoscopic therapy of Barrett’s: what other factors may be involved such as biliary reflux, or microenvironmental or more do we need to know? genetic abnormalities.7910 In summary, the study of Hage et al J-L Van Laethem, J Devie`re confirms results emerging from endoscopic treatments to ablate Barrett’s ...... oesophagus: APC alone or ALA-PDT in combination with APC achieves com- Results emerging from endoscopic treatments to ablate Barrett’s plete clearance of Barrett’s epithelium in oesophagus indicate that APC alone or ALA-PDT in combination approximately two thirds of patients. with APC achieves complete clearance of Barrett’s epithelium in The direct impact on the cancer risk in approximately two thirds of patients Barrett’s oesophagus is unknown. Given the risk of developing adenocarcinoma arising under squamous re-epithelialisa- ver the past few years, different in this scenario, only long term studies tion,11 12 the potential complications of endoscopic ablative techniques (more than five years) will tell us if this such endoscopic therapies, and the high Ohave been used in combination type of treatment is effective in prevent- cost of such management, we should with antireflux therapy with the aim of ing the need for surveillance and ulti- follow the authors’ recommendations of reversing Barrett’s oesophagus and mately in reducing the risk of cancer not using these techniques routinely. On replacing it with squamous epithelium. development. the other hand, these modalities, espe- The end goal of these procedures is to The results provided by the present cially PDT and its newer development in impact directly on the risk of tumour randomised study are relatively similar photosensitisation, offer fascinating per- development. to those reported in previous ones which spectives for the curative management To date, most of the published results assessed a single modality, particularly of early neoplastic changes and severe on ablative therapy in Barrett’s oeso- with regard to treatment procedures, dysplasia arising in Barrett’s oesopha- phagus have dealt with a single treat- median length of Barrett’s oesophagus gus. ment modality, and long term results as (3 cm), proton pump inhibitor therapy, Finally, these experimental groups of well as the real impact on the course of and duration of follow up (12 patients, having already received treat- the condition remain unanswered. months).1–3 ment to eradicate Barrett’s oesophagus, The most widely used and studied Clearly, there was no significant should be monitored in order to obtain procedures are photochemical (that is, difference between the two procedures valuable clinical information on the rate photodynamic therapy (PDT) using at six weeks, with a complete histologi- of tumour development over the next

Photofin or more recently 5-aminolevu- cal response rate of 33% and 36% for the 10 years. All of these patients should be http://gut.bmj.com/ linic acid (5-ALA)12) and thermal (that ALA-PDT fractionated dose and APC, followed up with further endoscopic is, argon plasma coagulation (APC)).3 respectively. As additional treatment evaluations at five and 10 years. 4 The study by Hage and colleagues, in with APC was permitted in both groups Gut 2004;53:779–780. this issue of Gut, is the first to compare for residual metaplastic mucosa, inter- doi: 10.1136/gut.2003.03639 two modalities [see page 785]. This preting the 12 month results is difficult. was a well designed, prospective, ran- However, it is interesting to note that ...... domised, three arm study involving 40 residual Barrett’s oesophagus was Authors’ affiliations on September 23, 2021 by guest. Protected copyright. patients with non or low dysplastic observed in only 10% of the ALA-PDT J-L Van Laethem, J Devie`re, Department of Barrett’s oesophagus, randomised to group and in 33% of the APC group, Gastroenterology, Erasme University Hospital, treatment with ALA-PDT as a single although the difference was not signifi- Brussels, Belgium dose of 100 J/cm2 (n = 13), ALA-PDT as cant as the number of patients enrolled a fractionated dose of 20 and 100 J/cm2 was relatively small. Correspondence to: Dr J-L Van Laethem, Department of Gastroenterology, Erasme (n = 13), or APC 65 W for two sessions The results from the ALA-PDT group University Hospital, 808 Route de Lennik, (n = 14). All patients received omepra- compare favourably with previous Brussels, Belgium; [email protected] zole 40 mg daily. After the designated reports: PDT is effective in eradicating treatment, any residual Barrett’s oeso- high grade dysplasia or superficial REFERENCES phagus was treated with additional tumours in 90–100% of cases but squa- sessions of APC. mous re-epithelialisation was found in 1 Wang KK, Kim JY. Photodynamic therapy in Barrett’s esophagus. Gastrointest Endosc The end points of the study were only two thirds of patients and was Clin N Am 2003;13:483–9. endoscopic reduction of the Barrett’s incomplete in all.25The rate of residual 2 Gossner L, Stolte M, Sroka R, et al. Photodynamic oesophagus surface and the microscopic Barrett’s oesophagus after eradication ablation of high-grade dysplasia and early cancer in Barrett’s oesophagus by means of 5- presence or absence of intestinal meta- using APC ranged from 0% to 68% in aminolevulinic acid. Gastroenterology plasia at different time points (that is, various series367after a mean follow up 1998;114:448–55. six weeks and 6, 12, 18, and period of approximately 12 months. It is 3 Franchimont D, Van Laethem JL, Devie`re J. Argon plasma coagulation in Barrett’s oesophagus. 24 months). In the setting of endoscopic not easy to explain this high variability Gastrintest Endosc Clin N Am 2003;13:457–66. therapy of Barrett’s oesophagus, com- which could depend on the technical 4 Hage M, Siersema PD, Van Dekken H, et al. 5- plete histological ablation of specialised procedure, length of the abnormal Aminolevulinic acid photodynamic therapy versus argon plasma coagulation for ablation of Barrett’s metaplastic epithelium should be con- mucosa, and perhaps most importantly oesophagus: a randomised trial. Gut sidered the unique relevant end goal of acid suppressive maintenance therapy. 2004;53:785–90. these procedures as only complete his- In long term follow up studies (med- 5 Overholt BF, Panjehpour M, Haydech JM. Photodynamic therapy for Barrett’s esophagus. tological clearance will influence the ian follow up 36 months), we and others Follow-up in 100 patients. Gastrointest Endosc risk of cancer in this condition. Even have clearly identified the length of the 1999;49:1–7.

www.gutjnl.com 780 COMMENTARY

6 Schulz H, Miehlke S, Antos D, et al. Ablation of 8 Kahaleh M, Van Laethem JL, Nagy N, et al. Long- oesophagus after photodynamic therapy. Barrett’s epithelium by endoscopic argon plasma term follow-up and factors predictive of Gastroenterology 2000;119:624–30. coagulation in combination with high-dose recurrence in Barrett’s esophagus treated by 11 Van Laethem JL, Peny MO, Salmon I, et al.

omeprazole. Gastrointest Endosc argon plasma coagulation and acid suppression. Intramucosal adenocarcinoma arising under Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from 2000;51:659–63. Endoscopy 2002;34:950–5. squamous re-epithelialisation of Barrett’s 7 Basu KK, Pick B, Bale R, et al. Efficacy and 9 Ouatu-Lascar R, Fitzgerald RC, Triadafilopoulos G. oesophagus. Gut 2000;46:575–7. 1 year follow-up of argon plasma coagulation Differentiation and proliferation in Barrett’s 12 Shand A, Dallal H, Palmer K, et al. therapy for ablation of Barrett’s oesophagus: esophagus and the effects of acid suppression. Adenocarcinoma arising in columnar lined factors determining persistence and Gastroenterology 1999;117:327–35. Barrett’s oesophagus following treatment recurrence of Barrett’s epithelium. Gut 10 Krishmadath KK, Wang KK, Tomiguchi K, et al. with argon plasma coagulation. Gut 2002;51:776–80. Persistent genetic abnormalities in Barrett’s 2001;48:580.

Infliximab marketing safety reports or warnings ...... from the US Food and Drug Administration have yielded additional information regarding the potential for Balancing the risks and benefits of opportunistic infections, including tuberculosis,9 disseminated histoplas- infliximab in the treatment of mosis,10 coccidioidomycosis, listeriosis, and Pneumocystis carinii pneumonia,11 inflammatory bowel disease and possibly non-Hodgkin’s lymphoma.12 Many of these post-marketing W J Sandborn, E V Loftus events occurred in patients with rheu- matoid arthritis who have a median age ...... 20 years older than that of patients with Patients with moderate to severely active Crohn’s disease treated Crohn’s disease. In this issue of Gut, Ljung and with infliximab may have a small but real risk of developing colleagues8 report on the clinical benefit severe infections, opportunistic infections, and non-Hodgkin’s and toxicity associated with the use of lymphoma infliximab in a population based cohort of 217 patients with inflammatory bowel disease (Crohn’s disease n = 191, nfliximab, a monoclonal antibody to a reflection of the uncertainties regard- ulcerative colitis n = 22, indeterminate tumour necrosis factor (TNF) a,isan ing safety.

colitis n = 4) in Stockholm County, http://gut.bmj.com/ important advance in the treatment of Subsequently, two blinded, placebo I 1–5 Sweden [see page 849]. Patients Crohn’s disease. The efficacy of inflix- controlled, phase 4 maintenance trials received a mean of 2.6 infusions of imab for the treatment of ulcerative (designed as infliximab withdrawal infliximab (range 1–11). Fifty four per colitis is still unclear.67 Infliximab was trials) were conducted in 573 patients cent of patients were also receiving approved for the treatment of Crohn’s with moderate to severely active Crohn’s azathioprine or 6-mercaptopurine, 51% disease in 1998 based on a 12 week disease4 and in 306 patients with fistu- were receiving corticosteroids, and 25% phase 2 trial in 108 patients1 (followed lising Crohn’s disease.5 All patients in 3 were receiving both. The authors reported by a 36 week extension trial) and a these two maintenance studies initially on September 23, 2021 by guest. Protected copyright. 2 that the clinical benefit observed in small phase 3 trial in 94 patients, both received at least one induction dose of of which showed compelling efficacy. infliximab. Thus safety data in patients patients with Crohn’s disease was Because of the possibility of open label with Crohn’s disease treated only with comparable with that reported in the crossover at week 4, 102 of 108 patients placebo who were naı¨ve to infliximab controlled trials referenced above and in the phase 2 trial received infliximab was available in only 34 of 1081 patients in other uncontrolled reports of inflix- 13 by week 12. Thus only six patients in the enrolled in placebo controlled trials. imab in clinical practice. The authors phase 2 study and 31 patients in the Potential treatment emergent safety also reported that severe adverse events phase 3 study who received placebo issues in the other 1047 patients treated occurred in 41 (18.9%) patients were available for safety follow up with infliximab in these clinical trials (Crohn’s disease n = 35, ulcerative coli- without crossover to infliximab. included non-Hodgkin’s lymphoma in tis n = 6) including: non-Hodgkin’s Although not all patients enrolled in 2/1047 (0.2%),34 opportunistic infec- lymphomas in 3/217 (1.4%) patients these studies were receiving concomi- tions in 3/1047 (0.3%) (tuberculosis (Crohn’s disease n = 3) of whom two tant corticosteroids and/or azathioprine n = 1; cytomegalovirus n = 1; cutaneous died; severe infections in 18/217 (8.3%) or 6-mercaptopurine, regulatory Nocardia n = 1),45 serious infections patients (Crohn’s disease n = 11, ulcera- approval in the USA was granted only which occurred at rates of 4% and tive colitis n = 5); serum sickness-like in patients with moderate to severely 4.6% in the two large maintenance reaction in 5/217 (2.3%) (Crohn’s dis- active Crohn’s disease unresponsive to trials,45 serum sickness-like reactions ease n = 5); and drug induced lupus in conventional therapy (not defined) and in 19/1047 (1.8%),45drug induced lupus 1/217 (0.5%) (Crohn’s disease n = 1). for fistulising Crohn’s disease; and in in 2/1047 (0.2%),34and death in 4/1047 The severe infections included opportu- Europe in patients with moderate to (0.4%) patients, including two deaths nistic infection in 2/217 (0.9%) patients severely active Crohn’s disease or fistu- from lymphoma and a death from sepsis (listeriosis n = 1 in a patient with lising Crohn’s disease unresponsive to a that were potentially related to inflix- Crohn’s disease, Pneumocystis carinii full and adequate course of corticoster- imab3–5 (one patient developed lym- n = 1 in a patient with ulcerative colitis oids and immunosuppressive therapy. phoma during a maintenance trial and who died) and fatal sepsis in 2/217 These restrictions were, at least in part, then died during follow up48). Post- (0.9%) patients (Crohn’s disease n = 1,

www.gutjnl.com COMMENTARY 781

ulcerative colitis n = 1). In total, 6/217 Crohn’s disease occurred at a crude infliximab, or are the result of other (2.8%) patients died in 28 months for a annual rate of 0.4% in the clinical important confounders such as age,

crude annual mortality of 1.2%: 3/191 trials, 1.2% of patients in the Ljung et al severity of illness, concomitant therapy Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from (1.6%) patients with Crohn’s disease study, and 1.3% of patients in the with corticosteroids, concomitant ther- died (non-Hodgkin’s lymphoma n = 2, Colombel et al study. The mortality rate apy with azathioprine or 6-mercapto- sepsis n = 1); 3/22 (13.6%) patients with in these three data sets is comparable purine, and combination therapy with ulcerative colitis died (sepsis n = 1, with what has previously been described corticosteroids and azathioprine or Pneumocystis carinii n = 1, pulmonary in several studies of the natural history 6-mercaptopurine, is impossible to embolus n = 1). All patients who died of Crohn’s disease.15–17 Non-Hodgkin’s determine from the available data. were receiving corticosteroids and most lymphoma occurred at a rate of 0.2% in Preliminary data from one large registry were elderly; at least two of the three the clinical trials and in the Colombel et study in 5000 patients has reported that patients who developed lymphoma were al study, and at a rate of 1.4% in the the severity of illness and corticosteroids receiving or had previously received Ljung et al study. Based on these results account for virtually all of the infections azathioprine. from clinical trials, a referral centre, and and any excess mortality in patients Another study recently reported the a population based cohort, we can with Crohn’s disease treated with inflix- safety experience in 500 consecutive conclude that patients with moderate imab.38 patients with Crohn’s disease treated to severely active Crohn’s disease trea- To definitively address these issues, with infliximab at the Mayo Clinic.14 ted with infliximab may have a small population based studies of patients Patients received a median of three but real risk of developing severe infec- treated with infliximab adjusted for infusions and had a median follow up tions, opportunistic infections, and non- important confounders are needed. of 17 months. Forty three patients Hodgkin’s lymphoma. However, it must Patient registries containing comparator (8.6%) experienced a serious adverse be pointed out that all three data sets patients not treated with infliximab will event of which 30 (6%) were considered lack adequate controls, and one cannot also be useful in this regard. Placebo to be possibly related to infliximab. be certain to what degree the potential controlled trials with infliximab (with- Serum sickness-like disease occurred in bias of infliximab being given to the out crossover of placebo patients to 19/500 patients and was attributed to most refractory patients, and concomi- infliximab) in patients with mild to infliximab in 14 (2.8%). Three patients tant immunosuppressive therapy, may moderately active Crohn’s disease who (0.6%) developed drug induced lupus. contribute to any possible risk. are not receiving concomitant therapy One patient (0.2%) developed a new Thus the important unanswered ques- with corticosteroids and/or immunosup- demyelination disorder. Forty eight tion is to what degree infliximab ther- pressives would also give a more clear patients had an infectious event of apy caused or contributed to these signal regarding safety outcomes, but which 41 (8.2%) were attributed to serious adverse events and to the such studies may no longer be feasible. infliximab. Twenty patients (0.4%) had observed crude annual mortality rates? In general, clinicians should restrict the a serious infection: two fatal sepsis, Population based studies in patients use of infliximab to patients with eight pneumonias of which two were with Crohn’s disease have shown only moderate to severely active Crohn’s fatal, six viral infections, two abdominal a slightly increased mortality15 18–22 (with disease who have failed conventional

abscesses requiring surgery, one arm three exceptions where no increased therapy such as corticosteroids and http://gut.bmj.com/ cellulitis, and one histoplasmosis mortality was reported)23–25 and no immunosuppressive therapy with (opportunistic infection). Nine patients increased risk of non-Hodgkin’s lym- azathioprine, 6-mercaptopurine, or had a malignant disorder, three of phoma26–29 (with one exception).30 methotrexate. Patients with ulcerative which were possibly related to inflix- However, these population based stu- colitis and indeterminate colitis should imab, including one lymphoma (0.2%). dies have not provided mortality or not be treated with infliximab until A total of 10 deaths were observed over lymphoma rates adjusted for patient definitive evidence of efficacy from

a median of 17 months, yielding a crude age, disease severity, or concomitant placebo controlled trials is available. on September 23, 2021 by guest. Protected copyright. annual mortality of 1.3%. For five of therapy with corticosteroids and/or Gut 2004;53:780–782. these patients (1%), the events leading azathioprine or 6-mercaptopurine. doi: 10.1136/gut.2003.020552 to death were possibly related to inflix- Clinical trials and observational studies imab. Most of the patients who died have reported that corticosteroids31 32 ...... were elderly. and azathioprine33 may result in abdom- Authors’ affiliations These three data sets (controlled inal abscess, sepsis, and death; and that W J Sandborn, E V Loftus, Inflammatory clinical trials, Ljung et al study,8 azathioprine and 6-mercaptopurine may Bowel Disease Clinic, Division of Colombel et al study14) show remarkable be associated with non-Hodgkin’s lym- Gastroenterology and Hepatology, Mayo convergence for the frequency of the phoma.34–37 Thus the data reported in the Clinic and Mayo Foundation, Rochester, most important adverse events. Serious clinical trials and by Ljung et al and Minnesota, USA or severe infections occurred at a rate of Colombel et al must be interpreted with 4.0–4.6% in clinical trials, 8.3% in the caution. At the present time all that we Correspondence to: Dr W J Sandborn, Mayo Ljung et al study, and 8.2% in the can really say is that patients with Clinic, 200 First Street SW, Rochester, MN 55905, USA; [email protected] Colombel et al study. Opportunistic moderately to severely active Crohn’s infection occurred at a rate of 0.3% in disease who are failing therapy with Conflict of interest. Dr Sandborn has received the clinical trials, 0.9% in the Ljung et al corticosteroids and/or immunosuppres- research support from, served as a consultant study, and 0.2% in the Colombel et al sive therapy and are subsequently for, and participated in continuing medical education events sponsored indirectly by study. Serum sickness-like reactions selected for therapy with infliximab Centocor Inc. occurred at a rate of 1.9% in the clinical have a small but apparently real risk of trials, 2.3% in the Ljung et al study, and serious infection, opportunistic infec- 2.8% in the Colombel et al study. Drug tion, and possibly non-Hodgkin’s lym- REFERENCES induced lupus occurred at a rate of 0.2% phoma. It is unclear if the crude annual 1 Targan SR, Hanauer SB, van Deventer SJ, et al. A in the clinical trials, 0.5% in the Ljung et al mortality rate is increased or not. short-term study of chimeric monoclonal antibody cA2 to tumor necrosis factor alpha for Crohn’s study, and 0.6% in the Colombel et al Whether these adverse outcomes are disease. Crohn’s Disease cA2 Study Group. study. Finally, death in patients with directly caused by or exacerbated by N Engl J Med 1997;337:1029–35.

www.gutjnl.com 782 COMMENTARY

2 Present DH, Rutgeerts P, Targan S, et al. disease: the Mayo Clinic experience in 500 27 Ekbom A, Helmick C, Zack M, et al. Extracolonic Infliximab for the treatment of fistulas in patients patients. Gastroenterology 2004;126:19–31. malignancies in inflammatory bowel disease. with Crohn’s disease. N Engl J Med 15 Jess T, Winther KV, Munkholm P, et al. Mortality Cancer 1991;67:2015–19. 1999;340:1398–405. and causes of death in Crohn’s disease: follow-up 28 Persson PG, Karlen P, Bernell O, et al. Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from 3 Rutgeerts P,D’Haens G, Targan S, et al. Efficacy of a population-based cohort in Copenhagen Crohn’s disease and cancer: a and safety of retreatment with anti-tumor necrosis County, Denmark. Gastroenterology population-based cohort study. Gastroenterology factor antibody (infliximab) to maintain remission 2002;122:1808–14. 1994;107:1675–9. in Crohn’s disease. Gastroenterology 16 Loftus EV jr, Silverstein MD, Sandborn WJ, et al. 29 Loftus EV Jr, Tremaine WJ, Habermann TM, et al. 1999;117:761–9. Crohn’s disease in Olmsted County, Minnesota, Risk of lymphoma in inflammatory bowel disease. 4 Hanauer SB, Feagan BG, Lichtenstein GR, et al. 1940–1993: incidence, prevalence, and survival. Am J Gastroenterol 2000;95:2308–12. Maintenance infliximab for Crohn’s disease: the Gastroenterology 1998;114:1161–8. 30 Bernstein CN, Blanchard JF, Kliewer E, et al. ACCENT I randomised trial. Lancet 17 Farmer RG, Whelan G, Fazio VW. Long-term Cancer risk in patients with inflammatory bowel 2002;359:1541–9. follow-up of patients with Crohn’s disease. disease: a population-based study. Cancer 5 Sands BE, Anderson FH, Bernstein CW, et al. Relationship between the clinical pattern and 2001;91:854–62. Infliximab maintenance therapy for fistulizing prognosis. Gastroenterology 1985;88:1818–25. 31 Malchow H, Ewe K, Brandes JW, et al. European Crohn’s disease. N Engl J Med 18 Palli D, Trallori G, Saieva C, et al. General and Cooperative Crohn’s Disease Study (ECCDS): 2004;350:876–85. cancer specific mortality of a population based results of drug treatment. Gastroenterology 6 Sands BE, Tremaine WJ, Sandborn WJ, et al. cohort of patients with inflammatory bowel 1984;86:249–66. Infliximab in the treatment of severe, steroid- disease: the Florence Study. Gut 1998;42:175–9. 32 Aberra FN, Lewis JD, Hass D, et al. refractory ulcerative colitis: a pilot study. Inflamm 19 Ekbom A, Helmick CG, Zack M, et al. Survival Corticosteroids and immunomodulators: Bowel Dis 2001;7:83–8. and causes of death in patients with inflammatory postoperative infectious complication risk in 7 Probert CS, Hearing SD, Schreiber S, et al. bowel disease: a population-based study. inflammatory bowel disease patients. Infliximab in moderately severe glucocorticoid Gastroenterology 1992;103:954–60. Gastroenterology 2003;125:320–7. resistant ulcerative colitis: a randomised 20 Persson PG, Bernell O, Leijonmarck CE, et al. 33 Connell WR, Kamm MA, Ritchie JK, et al. Bone controlled trial. Gut 2003;52:998–1002. Survival and cause-specific mortality in marrow toxicity caused by azathioprine in 8 Ljung T, Karle´n P, Schmidt D, et al. Infliximab in inflammatory bowel disease: a population-based inflammatory bowel disease: 27 years of inflammatory bowel disease: clinical outcome in a cohort study. Gastroenterology experience. Gut 1993;34:1081–5. population based cohort from Stockholm County. 1996;110:1339–45. 34 Present DH, Korelitz BI, Wisch N, et al. Gut 2004;53. 21 Loftus EV jr, Schoenfeld P, Sandborn WJ. The Treatment of Crohn’s disease with 9 Keane J, Gershon S, Wise RP, et al. Tuberculosis epidemiology and natural history of Crohn’s 6-mercaptopurine. A long-term, randomized, associated with infliximab, a tumor necrosis factor disease in population-based patient cohorts from double-blind study. N Engl J Med alpha-neutralizing agent. N Engl J Med North America: a systematic review. Aliment 1980;302:981–7. 2001;345:1098–104. Pharmacol Ther 2002;16:51–60. 35 Bouhnik Y, Lemann M, Mary JY, et al. Long-term 10 Lee JH, Slifman NR, Gershon SK, et al. Life- 22 Card T, Hubbard R, Logan RFA. Mortality in follow-up of patients with Crohn’s disease treated threatening histoplasmosis complicating inflammatory bowel disease: a population based with azathioprine or 6-mercaptopurine. Lancet immunotherapy with tumor necrosis factor alpha cohort study. Gastroenterology 1996;347:215–19. antagonists infliximab and etanercept. Arthritis 2003;125:1583–90. 36 Dayharsh GA, Loftus EV jr, Sandborn WJ, Rheum 2002;46:2565–70. 23 Farrokhyar F, Swarbrick ET, Grace RH, et al. Low et al. Epstein-Barr virus-positive lymphoma 11 Remicade (infliximab) for IV injection. Package mortality in ulcerative colitis and Crohn’s disease in patients with inflammatory bowel Insert. Centocor, Malvern, Pennsylvania, USA, in three regional centers in England. disease treated with azathioprine or 2003. Am J Gastroenterol 2001;96:501–7. 6-mercaptopurine. Gastroenterology 12 Brown SL, Greene MH, Gershon SK, et al. Tumor 24 Cottone M, Magliocco A, Rosselli M, et al. 2002;122:72–7. necrosis factor antagonist therapy and lymphoma Mortality in patients with Crohn’s disease. 37 Farrell RJ, Ang Y, Kileen P, et al. Increased development: twenty-six cases reported to the Scand J Gastroenterol 1996;31:372–5. incidence of non-Hodgkin’s lymphoma in Food and Drug Administration. Arthritis Rheum 25 Probert CS, Jayanthi V, Wicks AC, et al. Mortality inflammatory bowel disease patients on 2002;46:3151–8. from Crohn’s disease in Leicestershire, 1972– immunosuppressive therapy but overall risk is 13 Ricart E, Panaccione R, Loftus EV, et al. Infliximab 1989: an epidemiological community based low. Gut 2000;47:514–19. 38 Lichtenstein GR, Cohen RD, Feagan BG, for Crohn’s disease in clinical practice at the study. Gut 1992;33:1226–8. http://gut.bmj.com/ Mayo Clinic: the first 100 patients. 26 Lewis JD, Bilker WB, Brensinger C, et al. et al. Safety of infliximab in Crohn’s Am J Gastroenterol 2001;96:722–9. Inflammatory bowel disease is not associated with disesse: data from the 5000-patient TREAT 14 Colombel JF, Loftus EV jr, Tremaine WJ, et al. The an increased risk of lymphoma. Gastroenterology registry. Gastroenterology 2004;(abstract in safety profile of infliximab in patients with Crohn’s 2001;121:1080–7. press). on September 23, 2021 by guest. Protected copyright.

Spontaneous bacterial peritonitis know how to diagnose, treat, and even ...... prevent this potentially fatal infection.34 Even the ‘‘spontaneous’’ nature of this infection has been largely resolved Early events in spontaneous bacterial in recent years. We now know that the gut is the source of most of the bacteria peritonitis that eventually cause SBP.5 As cirrhosis develops in animals, gram negative B A Runyon bacteria increase in numbers in the gut.5 We know that the gut of animals ...... and patients with advanced cirrhosis is Insight into the very early events in the pathogenesis of more permeable to bacteria than the normal gut and more permeable than spontaneous bacterial peritonitis the gut in less advanced cirrhosis.67 Once bacteria reach a critical concentra- he database regarding spontaneous deaths, which were probably viewed as tion in the gut lumen, they ‘‘spill over’’, bacterial peritonitis (SBP) has mysterious at the time. and escape the gut, ‘‘translocating’’ to Tincreased dramatically in the past Now we know that SBP is quite mesenteric lymph nodes. Then they can 33yearssincethisphrasewasfirstcoined common, with a prevalence of .20% enter lymph, blood, and eventually byHaroldConn.1 In the remote past this on admission to the hospital, prior to ascitic fluid.6 If the ability of the ascitic easily treatable cause of deterioration of the era of prevention.2 We know who is fluid to assist macrophages and neutro- patients with advanced cirrhosis was at high risk—patients with cirrhosis and phils in killing the errant bacteria is underdiagnosed and undertreated. This (a) prior SBP, (b) low protein ascites, or deficient, uncontrolled growth occurs.8 undoubtedly led to many unnecessary (c) gastrointestinal bleeding.3 We also This is SBP. In general, the animal or

www.gutjnl.com COMMENTARY 783 patient dies if they develop this infection translocated pieces of bacteria (for Now that this new subset of patients and it is not promptly diagnosed and example, DNA).23 24 Bacterial DNA can with molecular evidence of transloca-

treated. bind to the toll-like receptor 9 of cells of tion has been identified, it is perhaps Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from Thus SBP is the result of failure of the the innate immune system and activate time to perform a prospective study gut to contain bacteria and failure of the them.25 Host immune defences are able following those patients who are posi- immune system to kill the virulent to kill the bacteria in these episodes of tive for bacterial DNA and determining bacteria once they have escaped the colonisation that do not progress to SBP. if their risks of SBP, hepatorenal syn- gut. Patients and animals have duplica- SBP occurs (a) when the organism is drome, and death are higher than those tive mechanisms of protection from more virulent than the bacteria that of DNA negative controls. If excessive bacteria. This makes great sense tele- were killed by host immune defences at morbidity and/or mortality are docu- ologically. Opsonins assist motile and the stage of colonisation, (b) when mented, the next step would be to fixed ‘‘professional’’ killers of bacteria, immune defences weaken, or (c) a conduct a randomised controlled trial the neutrophils and Kupffer cells, combination of these events. It is the of selective intestinal decontamination respectively. Innate defenders against peritoneal macrophage that is the first versus placebo in the DNA positive bacterial invasion include macrophages, line of defence against bacterial coloni- group and determine if hepatorenal dendritic cells, and natural killer cells. sation of ascitic fluid.26 SBP occurs when syndrome and death can be prevented. These cells synthesise proinflammatory macrophages fail to kill the bacteria and Gut 2004;63:782–784. cytokines and effector molecules which the second line of defence is called in, doi: 10.1136/gut.2003.035311 assist in killing bacteria. the neutrophils. Unfortunately, patients with Based on the information presented Correspondence to: Professor B A Runyon, Chief, Liver Service, Loma Linda University advanced cirrhosis have been reported above, one would predict that the Medical Center, 11234 Anderson Street, Room to have defects and dysfunction in many presence of whole bacteria or DNA in 1556, Loma Linda, California, 92354, USA; of these systems of protection.8–11 It is serum and ascitic fluid would have [email protected] then no surprise that these patients are consequences (for example, stimulation vulnerable to infection by their own gut of immune defences, effector molecules, REFERENCES flora. To make matters worse, some of and cytokines). This could in turn 1 Conn HO, Fessel JM. Spontaneous bacterial the effector molecules and cytokines impact on haemodynamics, renal func- peritonitis in cirrhosis: variations on a theme. that help kill the bacteria have un- tion, and survival. These effector mole- Medicine 1971;50:161–97. desired side effects. Nitric oxide (NO) cules and cytokines are two edged 2 Pinzello G, Simonetti RG, Craxi A, et al. Spontaneous bacterial peritonitis: a prospective is one of these effector molecules. swords. They can protect from bacterial investigation in predominantly nonalcoholic Tumour necrosis factor (TNF) is one of infection but they can also initiate a cirrhotic patients. Hepatology 1993;3:545–9. the relevant cytokines. NO is probably sequence of events than can lead to the 3 Rimola A, Garcia-Tsao G, Navasa M, et al. Diagnosis, treatment and prophylaxis of the long sought after agent responsible patient’s death. spontaneous bacterial peritonitis: a consensus for vasodilation that characterises The elegant study in this issue document. J Hepatol 2000;32:142–53. advanced liver disease.12 Bacterial infec- furthers this line of investigation and 4 Runyon BA. Management of adult patients with ascites caused by cirrhosis. Hepatology tion leads to further elevations in these provides the scientific rationale for new 1998;27:264–72. 13 14 molecules. NO and TNF are impor- clinical studies, including a randomised 5 Guarner C, Runyon BA, Young S, et al. Intestinal http://gut.bmj.com/ tant mediators of the further vasodila- controlled trial.20 The authors harvested bacterial overgrowth and bacterial translocation tion and renal failure that too often in an experimental model of cirrhosis in rats. peritoneal macrophages from patients J Hepatol 1997;26:1372–8. 14–16 accompany SBP. with cirrhosis and ascites. Approxi- 6 Runyon BA, Squier SU, Borzio M. Translocation The good news here is that selective mately one third of their patients had of gut bacteria in rats with cirrhosis to mesenteric intestinal decontamination with poorly detectable bacterial DNA. They divided lymph nodes partially explains the pathogenesis of spontaneous bacterial peritonitis. J Hepatol absorbed antibiotics reduces gut bacter- the patients into two groups: those with 1994;21:792–6. 7 Cirera I, Bauer TM, Navasa M, et al. Bacterial ial counts, reduces translocation rates, and without bacterial DNA in serum on September 23, 2021 by guest. Protected copyright. can prevent SBP in high risk subgroups, and ascitic fluid. They measured macro- translocation of enteric organisms in patients with cirrhosis. J Hepatol 2001;34:32–7. and can improve the hyperdynamic phage production of NO metabolites and 8 Runyon BA. Patients with deficient ascitic fluid circulatory state of these patients.31718 cytokines, including TNF, and compared opsonic activity are predisposed to spontaneous Selective intestinal decontamination can the results between the two groups. The bacterial peritonitis. Hepatology 1988;8:632–5. 9 Runyon BA, Morrissey R, Hoefs JC, et al. Opsonic even improve survival of rats with authors convincingly demonstrate that activity of ascitic fluid: a potentially important cirrhosis and ascites.19 peritoneal macrophages from patients protective mechanism against spontaneous This brings us to the most recent with cirrhosis and bacterial DNA in bacterial peritonitis. Hepatology 1985;5:634–7. 10 Fiuza C, Salcedo M, Clemente G, et al. In vivo contribution to this line of investigation serum and ascitic fluid are markedly neutrophil dysfunction in cirrhotic patients with published by France´s and colleagues20 in activated, as evidenced by increased NO advanced liver disease. J Infect Dis this issue of Gut [see page 860]. These synthesising ability and enhanced cyto- 2000;182:526–33. 20 11 Rimola A, Soto R, Bory F, et al. investigators have previously shown kine production. This study provides Reticuloendothelial system phagocytic activity in that some patients with cirrhosis have further insight into very early events in cirrhosis and its relation to bacterial infections bacterial DNA in their serum and ascitic the pathogenesis of SBP. Pieces of and prognosis. Hepatology 1984;4:53–8. fluid, and that the DNA is always bacteria commonly escape the gut and 12 Wiest R, Groszmann RJ. The paradox of nitric oxide in cirrhosis and portal hypertension: too present simultaneously in both body end up in blood and ascitic fluid. In the much, not enough. Hepatology fluids.21 This provides molecular evi- process, a complex sequence of events 2002;35:478–91. dence of bacterial translocation. They occurs. The immune system is stimu- 13 Jimenez W, Ros J, Morales-Ruiz M, et al. Nitric oxide production and inducible nitric oxide have also shown that patients who lated to contain the bacterial colonisa- synthase expression in peritoneal macrophages subsequently develop SBP have a higher tion and protect the host from fatal of cirrhotic patients. Hepatology 1999;30:670–6. baseline ascitic fluid TNF level than infection. However, as a consequence, 14 Such J, Hillebrand DJ, Guarner C, et al. Nitric 22 oxide in ascitic fluid is an independent predictor patients who do not develop SBP. It the effector molecules and cytokines are of renal impairment in patients with cirrhosis and is probable that in patients and in rats increased, setting the stage for worsen- spontaneous bacterial peritonitis. Eur J with cirrhosis that SBP is preceded by ing of the haemodynamic status, devel- Gastroenterol Hepatol, 2004 (in press). 15 Follo A, Llovet JM, Navasa M, et al. Renal episodes of colonisation of blood and opment of functional renal failure, and impairment after spontaneous bacterial peritonitis ascitic fluid with viable bacteria or the possibility of death. in cirrhosis: incidence, clinical course, predictive

www.gutjnl.com 784 COMMENTARY

factors and prognosis. Hepatology 19 Guarner C, Runyon BA, Heck M, et al. Effect of oxide in sterile ascitic fluid and serum from 1994;20:1495–501. long-term trimethoprim/sulfamethoxazole patients who subsequently develop ascitic fluid 16 Ruiz-del-Arbol L, Urman J, Fernandez J, et al. prophylaxis on ascites formation, bacterial infection. Dig Dis Sci 2001;46:2360–6.

Systemic, renal, and hepatic hemodynamic translocation, spontaneous bacterial peritonitis 23 Runyon BA. Monomicrobial nonneutrocytic Gut: first published as 10.1136/gut.2003.035311 on 11 May 2004. Downloaded from derangement in cirrhotic patients with and survival in cirrhotic rats. Dig Dis Sci bacterascites: a variant of spontaneous bacterial spontaneous bacterial peritonitis. Hepatology 1999;44:1957–62. peritonitis. Hepatology 1990;12:710–15. 2003;38:1210–18. 20 France´sR, Mun˜oz C, Zapater P, et al. Bacterial 24 Runyon BA, Sugano S, Kanel G, et al. A rodent 17 Runyon BA, Borzio M, Young S, et al. Effect of DNA activates cell mediated immune response model of cirrhosis, ascites, and bacterial selective intestinal decontamination with and nitric oxide overproduction in peritoneal peritonitis. Gastroenterology 1991;100:489–93. norfloxacin on spontaneous bacterial macrophages from patients with cirrhosis and 25 Wagner H. Interactions between bacterial CpG- peritonitis, translocation, and survival in an ascites. Gut 2004;53:860–4. DNA and TLR9 bridge innate and adaptive animal model of cirrhosis. Hepatology 21 Such J, France´s R, Mun˜oz C, et al. Detection and immunity. Curr Opin Microbiol 2002;5:62–9. 1995;21:1719–24. identification of bacterial DNA in patients with 26 Dunn DL, Barke RA, Knight NB, et al. Role of 18 Rasaratnum B, Kaye D, Jennings G, et al. The cirrhosis and culture-negative, nonneutrocytic resident macrophages, peripheral neutrophils effect of selective intestinal decontamination on ascites. Hepatology 2002;36:135–41. and translymphatic absorption in bacterial the hyperdynamic circulatory state in cirrhosis. 22 Such J, Hillebrand DJ, Guarner C, et al. Tumor clearance from the peritoneal cavity. Infect Immun Ann Intern Med 2003;139:186–93. necrosis factor-alpha, interleukin-6 and nitric 1985;49:257–64. http://gut.bmj.com/ on September 23, 2021 by guest. Protected copyright.