R

H

TR

a

ARRAA

KHSRA

1

(iEHbpuHet‘r

H

0h

Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 395– 399

Contents lists available at SciVerse ScienceDirect

Transactions of the Royal Society ofTropical Medicine and Hygiene

j ourna l ho me pag e: ht t p: / /www.e lsev ier .com/ locate / t rs tmh

eview

aemolytic uraemic syndrome during shigellosis

homas Butler ∗,1

oss University School of Medicine, Dominica, West Indies

r t i c l e i n f o

rticle history:eceived 19 September 2011eceived in revised form 12 April 2012ccepted 12 April 2012vailable online 10 May 2012

eywords:

a b s t r a c t

Haemolytic uraemic syndrome (HUS), which is comprised of the triad of haemolyticanemia, thrombocytopenia and renal insufficiency, occurs in about 13% of dysentericpatients with shigellosis due to Shigella dysenteriae type 1 (SD1) infections, who are mostlychildren less than five years old in Africa and Asia. With a case-fatality rate of about 36%,it is the leading cause of death in SD1 outbreaks. Research suggests that Shiga toxin andlipopolysaccharide from the causative bacteria play roles in pathogenesis. The risk of HUSis increased when inappropriate antimicrobial drugs, against which infecting bacteria are

aemolytic uraemic syndrome

higellosisisk factorsntimicrobial treatment

resistant, are used or when any antimicrobial drug is given more than four days after thestart of diarrhoea. To prevent HUS, it is advised to initiate an appropriate drug early andto consider withholding antimicrobial therapy in patients presenting more than four daysafter the onset of diarrhoea.

© 2012 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.

. Introduction

Diarrhoea-associated haemolytic uraemic syndromeHUS), caused by Shiga-toxin-producing Escherichia colis currently an important disease in North America andurope, but the first recognition of diarrhoea-associatedUS was in shigellosis in Bangladesh in 19751 followedy a warning in 1987 that antibiotic treatment couldrecipitate HUS.2 The purpose of this review is to bringp to date worldwide reports of shigellosis-associatedUS and our understanding of pathogenesis in addition tofforts at aiming to prevent it. The method was a search ofhe literature using Pubmed with the keywords ‘HUS’ andshigella’, and cross-referencing papers to select clinicallyelevant investigations.

∗ Tel.: +767 255 6529.E-mail address: tbutler@rossmed.edu.dm

1 Present address: Ross University School of Medicine, 630 USighway 1, Suite 500, North Brunswick NJ 08902, USA.

035-9203/$ – see front matter © 2012 Royal Society of Tropical Medicine and Http://dx.doi.org/10.1016/j.trstmh.2012.04.001

2. Occurrence and geographic distribution

In the last 25 years reports of 488 cases of HUS duringshigellosis outbreaks have accumulated from 11 countries(Table 1). These reports used the standard triad of labora-tory and clinical signs to diagnose the complication of HUS:haemolytic anemia as shown by low blood hemoglobinconcentrations and presence of schistocytes in bloodsmears, thrombocytopenia, and elevated serum creatinineconcentrations or anuria, oliguria or haematuria. Asidefrom eight adults with HUS,5,18 all cases were children, withthe four sites contributing the largest numbers of casesshowing a median age of 29 months and 65% males.2,3,7,10

Other clinical features that accompanied or precededHUS in these patients were leucocytosis with meanvalues of blood leucocytes >30 000 per mm2,5,8,12,13 andhyponatraemia with serum sodium concentrations <125meq/L.7,8,15 Stool cultures yielded Shigella dysenteriaetype 1 (SD1) in most patients or was assumed to be the

cause of dysentery because this serotype was present inthe respective communities. Exceptions were six patientsin whom S. flexneri was isolated.7,8,10,19 Resistance ofSD1 isolates to ampicillin, chloramphenicol, tetracycline,

ygiene. Published by Elsevier Ltd. All rights reserved.

396 T. Butler / Transactions of the Royal Society of Tropical Medicine and Hygiene 106 (2012) 395– 399

Table 1Reports of haemolytic uraemic syndrome (HUS) during shigellosis

Country No. cases Age in months, meanor median

Interval in days between onset ofdiarrhoea and detection of HUS

Case fatality rate, % References

South Africa 151 55 7 17 3–6Zimbabwe 110 18 11 41 7,8India 74 28 8 59 9,10Nepal 55 25 17 23 11,12Saudia Arabia 33 36 8 26 13,14Bangladesh 30 40 6 37 2Kenya 21 19 4 52 15France 5 50 6–15 0 16Burma 4 42 NA 50 17USA 3 84 10 0 18

2

Thailand 2 22 1

NA: Not available.

sulfonamides, and trimethoprim-sulfamethoxozole (T-S)was present at most sites, whereas isolates were sensitiveto nalidixic acid, ciprofloxacin, and ceftriaxone.

Haemolytic uraemic syndrome was manifested latein the course of shigellosis, detected often more thana week after the onset of dysentery sometimes whenthe diarrhoea was waning. Clinical reports indicate thatintervals between the onset of bloody diarrhoea and signsof HUS can be as short as four days15 but longer intervalsof up to 17 days have been reported (Table 1). In seven ofthe reports, the incidence of HUS was provided in relationto all dysentery cases observed in a particular outbreakor clinic setting, ranging from 6–45% with an overallincidence of 95 HUS cases per 721 dysentery cases, or13%.4,5,8,13,14,16,18 Case-fatality rates given for 384 cases ofHUS were overall 36%, which was much higher than ratesfor uncomplicated dysentery cases. Thus, the importanceof HUS during shigellosis is that this complication hasemerged as the leading cause of death in epidemics ofdysentery caused by SD1.

3. Pathogenesis of haemolytic uraemic syndrome

Haemolytic uraemic syndrome is described as a disor-der of blood clotting known as disseminated intravascularcoagulation (DIC) with clots containing fibrin and plateletsforming preferentially in the renal glomeruli.20,21 Kidneysshow ischaemic cortical necrosis in some cases explainingrenal insufficiency and urinary abnormalities. Thrombo-cytopenia follows from consumption of platelets in clots,and microangiopathic haemolytic anemia results frommechanical injury to red blood cells traversing partiallythrombosed blood vessels.

The occurrence of HUS during shigellosis has beenlinked almost exclusively to infection with the Shigellaserotype SD1, which is unique among species of Shigella aswell as other serotypes of S. dysenteriae for producing largeamounts of Shiga toxin and causing epidemics of severedisease in developing countries. In patients with dysen-tery due to SD1, Shiga toxin has been measured in stoolspecimens,22 but the extent of its absorption into blood

to cause HUS is not known. Closely related Shiga toxinsproduced by E. coli O157:H7, with designations Stx1 andStx2, are believed important in causing HUS during bloodydiarrhoea caused by that infection.23 Although E. coli

0 19

O157:H7 and SD1 are similar for causing bloody diarrhoeafollowed by HUS, SD1 is different by being enteroinva-sive, producing fever regularly as well as bacteraemiaoccasionally, and its HUS occurs almost exclusively inyoung children, whereas E. coli O157:H7 causes afebrilediarrhoea, never bacteraemia, and HUS more broadlyacross the age spectrum in adults and children.23,24 Stx1 isidentical to Shiga toxin of SD1, which is an enzymaticallyactive protein with molecular weight of 67 000 daltonsand A-B subunit structure, whereas Stx2 is 56% homol-ogous with Stx1 and is the primary virulence factor forE. coli O157:H7.23 The A subunits of both toxins act byinhibiting protein synthesis in ribosomes. The B subunitsbind Shiga toxin to the glycolipid globotriasylceramide(Gb3) receptor, which is present on glomerular endothelialcells, mesangial cells, podocytes, and renal tubular cells.21

Another candidate molecule for causing HUS islipopolysaccharide (LPS or endotoxin), which Gram-negative Shigellae contain as a potential mediator ofvascular injury, DIC, and sepsis. With the limulus assayLPS was detected in the blood of patients with Shigella-associated HUS but rarely in controls with uncomplicateddysentery.25 LPS of Shigella without participation of Stxcan experimentally produce the renal lesion of HUS inrabbits by two intravenous injections of sublethal doses ofendotoxin that result in a generalized Shwartzman reac-tion (GSR).26 The first, or preparatory, dose sensitizes theanimal and the second dose the next day provokes renalcortical thrombosis and necrosis that can be preventedby rendering animals neutropenic by busulfan treatment.Free haemoglobin in the blood during haemolysis of HUSwould have the capacity to enhance toxicity of LPS,27 thusincreasing severity of blood coagulation in the kidney. Sitesof bleeding in the colon serve as potential portals of entryinto the blood stream of luminal bacterial products includ-ing LPS linked to haemoglobin released by red blood cellsthat had been transported into the lumen and broken downby harsh chemical conditions there. Additionally, hypo-tonicity of serum due to hyponatraemia could contributeto haemolysis by decreasing deformability of swollenred blood cells in capillary beds. Young children may be

at greater risk than adults because, when subjected tohypotonicity of plasma, their considerably smaller kidneysalong with red cell volumes nearly equal to adults, woulddevelop a disproportionate reduction in ratio of vascular

Tropica

lgt

4d

cccoiobwd4mkdpHeefimamobTbtscwaaitaoai

TA

T. Butler / Transactions of the Royal Society of

umen diameter to red blood cell diameter, resulting inreater mechanical stress to swollen erythrocytes passinghrough vessels narrowed by expansion of renal cells.

. Role of antimicrobial therapy and advancedisease

A striking feature of the shigellosis-associated HUSases is that the complication regularly followed antimi-robial treatment of dysentery in most reports that gave ahronology of therapy and evolution of HUS signs. Antibi-tic therapy as a precondition for HUS was first recognizedn Bangladesh in 1987.2 The first clear example of an antibi-tic effect for HUS in a patient with SD1 was a case reporty Ullis and Rosenblatt in 197328 of a 5-year-old boy whoas admitted six days after the onset of dysentery acquireduring vacation in Mexico. His white blood cell count was2 800 per mm3 and serum sodium concentration 111eq/L. After four days of treatment with ampicillin and

anamycin, to which his SD1 isolate was susceptible, heeveloped oliguria followed by signs of HUS, but theseioneer authors designated their case as neither beingUS nor showing an antibiotic effect. In 1975, Rahamant al1 reported 16 children with SD1 in Bangladesh whoxhibited leucocytosis and haemolytic anaemia but onlyve of them showed renal impairment. These patients alsoanifested their leucocytosis and haemolytic anaemia

fter receiving antibiotics, to which the strains of SD1 wereostly susceptible. The authors did suggest that some

f their patients had HUS but didn’t make a connectionetween treatment and subsequent development of HUS.1

wo different forms of such an antibiotic effect have sinceeen proposed: appropriate therapy consisting of a drugo which an isolate of SD1 is susceptible and which hashown benefits in clinical trials; and inappropriate therapyonsisting of a drug to which a strain of SD1 is resistant orhich has not shown benefits in clinical trials. Examples of

ppropriate drugs for SD1 in the last 25 years were nalidixiccid, ciprofloxacin, and ceftriaxone, whereas examples ofnappropriate drugs were ampicillin and T-S. Appropriateherapy might cause release of toxins from dying SDI or

ctivate release of bacterial products from other speciesf intestinal flora; inappropriate therapy might give SD1n advantage by eliminating competing flora and alsoncrease release of bacterial products from other species

able 2ntimicrobial drugs started before the onset of haemolytic uraemic syndrome (H

Drug No. cases

Ampicillin 14

11

1

Ceftriaxone 4

3Ciprofloxacin 2

1

1

Nalidixic acid 7

Trimethoprim-sulfamethoxazole Most common drug in 60 cases

1

1

l Medicine and Hygiene 106 (2012) 395– 399 397

of dying intestinal flora. In a two-step process analogousto the GSR in rabbits, the preparatory step might occurduring bloody diarrhoea when Stx and LPS leak into thecirculation through the damaged colonic mucosa, followedby a second burst of LPS release into blood a few days latercaused by antibiotic-induced lysis of gut bacteria.

Authors of nine reports of shigellosis-associated HUSgave specific drugs employed and intervals betweentherapy and appearance of HUS (Table 2). Inappropriatetreatment due to resistant strains of SD1 was emphasizedas a likely predisposition to HUS, but many patients werealso acknowledged to have received appropriate drugs. Inoutbreaks in Saudi Arabia, inappropriate use of ampicillinfor dysentery patients was associated with incidencesof HUS of 36%13 and 71%14 whereas appropriate use ofnalidixic acid was associated with incidences of 7%13 and12%.14 Owing to antimicrobial therapy being the standardof care for dysentery in most countries, very few untreatedpatients are available for comparison, but Bin Saeed et al.13

reported that no cases of HUS developed in 12 patientswho received no antibiotic. Timing of initiation of therapyis critical. When the appropriate drug ciprofloxacin wasadministered within the first three days after onset ofdysentery due to SD1 in a prospective trial, only one childout of 128 developed HUS.6

Before HUS can make its appearance, a patient with SD1infection needs to progress, without antibiotic treatmentor with inappropriate therapy, from mild colitis to severecolitis that Gilman et al.29 by proctoscopy describedas heavy exudates, friability, and pseudomembranes.Examining the entire colon by colonoscopy in adults withshigellosis, Speelman et al.30 observed colitis that wasmost frequent and intense in the rectosigmoid colon butpatients with histories of symptoms for four days or morewere likely to show extension of inflammation proximallyto the splenic flexure, transverse colon and even pancolitis,with patients infected with SD1 exhibiting a higher inci-dence of proximal involvement than patients infected withother species or serotypes. When death occurs for reasonsother than HUS, patients sometimes show colonic perfo-ration, toxic megacolon, and extensive necrosis of colonic

mucosa with formation of a pseudomembrane.31 Damagedcolonic mucosa would be permeable to bacteria as well asbacterial breakdown products after antibiotic treatment.An unexplored factor in pathogenesis of HUS is that

US)

Interval between start of drug anddetection of HUS in days, median or range

References

5–7 136 25 195–10 16Within 5 35–10 165 65 97 711 76 188 19

f Tropica

398 T. Butler / Transactions of the Royal Society o

anatomic contiguity of the inflamed proximal colon withkidneys in the mid-abdomen and upper abdomen maypermit delivery of high concentrations of bacterial prod-ucts through lymphatic channels between the colon andkidneys,32 which would be enhanced in the colonic laminapropria by lymphangiogenesis during inflammation.33

Formation of pseudomembranes after antibiotics couldsuggest overgrowth of Clostridium difficile, but testing ofstool for toxins of C. difficile has not been carried out in HUSpatients.

5. Prevention of haemolytic uraemic syndrome

Therapeutic interventions to prevent HUS after theonset of diarrhoea caused by E. coli O157:H7 that might berelevant for SD1 infections have been attempted, but nonehas been proven effective.34 Administering corticosteroids,anticoagulants, and antiplatelet drugs to patients withdiarrhoea did not prevent HUS. Monoclonal antibodies spe-cific for Shiga toxin that neutralize toxicity have been madeand show promise in animal models for preventing HUS.23

Prevention of HUS in SD1 infections can be best achievedby starting an appropriate antibiotic early, that is, withinthe first four days of symptoms. Choosing an appropri-ate antibiotic requires culture and susceptibility testingof stool specimens. In field situations where laboratoryservices are scarce, however, not all patients need to be cul-tured because SD1 tends to circulate in outbreaks caused bya single strain that can be identified for predicting appro-priate treatment of other patients. Early appropriate use ofthese drugs is beneficial to shorten the course of symptomsas well as to clear viable SD1 from stool cultures and is prob-ably life-saving.35 In prospective clinical trials of patientsinfected with SD1 using early appropriate antimicrobials,including nalidixic acid,36,37 norfloxacin,38 ciprofloxacin,39

ceftriaxone40 and pivmecillinam,37 HUS did not occur. It islikely that some cases of HUS could be prevented by with-holding antibiotics in patients with severe symptoms whopresent themselves for treatment more than four days afterthe onset of illness. A four-day period of diarrhoea beforeHUS develops is the same duration described in an out-break of E. coli O157:H7 in Utah, in which HUS appeared 5–9days after onset of diarrhoea.41 At these late times, immuneresponses should be mounting to produce antibodies thatwill result in self-limited infection. Other measures sug-gested by Tarr et al.20 to preserve renal blood flow duringdiarrhoea due to E. coli O157:H7 are expansion of intravas-cular volume with isotonic intravenous fluids and avoid-ance of nonsteroidal anti-inflammatory drugs, which canreduce renal blood flow. Use of isotonic saline would alsoreverse low osmolarity in patients with hyponatraemia.Oral rehydration would also be beneficial for good renalblood flow. Anti-motility drugs, which increase the risk ofHUS, should be avoided. A warning against antibiotic usefor HUS associated with E.coli O157:H7 has been in effect

for the past decade, but withholding antimicrobial drugsfor that infection has been easy for physicians to imple-ment because treatment of the E. coli infections at any timeafter the onset of symptoms seemed to provide no clinicalbenefit.20,41,42

l Medicine and Hygiene 106 (2012) 395– 399

6. Conclusions

In SD1 infections, for which early appropriate antimi-crobial therapy is beneficial, this proposal to avoidantimicrobials after four days of symptoms to preventHUS must be considered to carry risks as well as potentialbenefits. For individual patients, physicians would needto follow clinical and laboratory signs daily in decidingwhether to withhold or institute therapy. A warning aboutantimicrobials in SD1 infection has practical implicationsand can be considered supported by clinical featuresshared with E. coli O157:H7 of bloody diarrhoea, severecolitis with pseudomembranes, Shiga toxin, leucocytosis,hyponatraemia, HUS risk and antibiotic effect. Whetherwarnings about antimicrobial treatment will result inany benefits for patients with either shigellosis or E. coliinfections will depend on future reports about clinicaloutcomes correlated with therapeutic choices.

Author’s contribution: TB fulfilled all the duties ofauthorship and is guarantor of the paper.

Funding: None.

Competing interests: None declared.

Ethical approval: Not required.

References

1. Rahaman MM, Alam AKMJ, Islam MR, Greenough WB, LindenbaumJ. Shiga bacillus dysentery associated with marked leukocytosis anderythrocyte fragmentation. Johns Hopkins Med J 1975;136:65–70.

2. Butler T, Islam MR, Azad MAK, Jones PK. Risk factors for devel-opment of hemolytic uremic syndrome during shigellosis. J Pediatr1987;110:894–7.

3. Bhimma R, Rollins NC, Coovadia HM, Adhikari M. Post-dysenterichemolytic uremic syndrome in children during an epidemic of Shigelladysentery in Kwazulu/Natal. Pediatr Nephrol 1997;11:560–4.

4. Rollins NC, Wittenberg DF, Coovadia HM, Pillay DJ, Karas AJ, SturmAW. Epidemic Shigella dysenteriae Type 1 in Natal. J Trop Pediatr1995;41:281–4.

5. Bloom PD, MacPhail AP, Klugman K, Louw M, Raubenheimer C,Fischer C. Haemolytic-uraemic syndrome in adults with resistantShigella dysenteriae type 1. Lancet 1994;344:206.

6. The Zimbabwe, Bangladesh, South Africa (Zimbasa) Dysentery StudyGroup. Multicenter, randomized, double blind clinical trial of shortcourse versus standard course oral ciprofloxacin for Shigella dysente-riae type 1 dysentery in children. Pediatr Infect Dis J 2002;21:1136–41.

7. Nathoo KJ, Sanders JA, Siziya S, Mucheche C. Haemolytic uraemic syn-drome following Shigella dysenteriae type 1 outbreak in Zimbabwe: aclinical experience. Cent Afr J Med 1995;41:267–74.

8. Oneko M, Nyathi MN, Doehring E. Post-dysenteric hemolytic uremicsyndrome in Bulawayo, Zimbabwe. Pediatr Nephrol 2001;16:1142–5.

9. Taneja N, Lyngdoh VW, Sharma M. Haemolytic uraemic syndromedue to ciprofloxacin-resistant Shigella dysenteriae serotype 1. J MedMicrobiol 2005;54:997–8.

10. Srivastava RN, Moudgil A, Bagga A, Vasudev AS. Hemolytic uremicsyndrome in children in northern India. Pediatr Nephrol 1991;5:284–8.

11. Baranwal AK, Ravi RNM, Singh R. Diarrhea associated with hemolyticuremic syndrome: a 3-year PICU experience from Nepal. Ind J Pediatr2009;76:1180–2.

12. Jha DK, Singh R, Raja S, Kumari N, Das BK. Clinico-laboratory profile ofhaemolytic uremic syndrome. Kathmandu Univ Med J 2007;5:468–74.

13. Bin Saeed AAA, El Bushra HE, Al-Hamdan NA. Does treatment of bloody

diarrhea due to Shigella dysenteriae Type 1 with ampicillin precipitatehemolytic uremic syndrome? Emerg Infect Dis 1995;1:134–7.

14. Al-Qarawi S, Fontaine RE, Al-Qahtani M-S. An outbreak of hemolyticuremic syndrome associated with antibiotic treatment of hospitalinpatients for dysentery. Emerg Infect Dis 1995;1:138–40.

Tropica

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

T. Butler / Transactions of the Royal Society of

5. Olotu AI, Mithwani S, Newton CRJC. Haemolytic uraemic syndromein children admitted to a rural district hospital in Kenya. Trop Doctor2008;38:165–7.

6. Houdouin V, Doit C, Mariani P, et al. A pediatric cluster of Shigelladysenteriae serotype 1 diarrhea with hemolytic uremic syndrome in2 families from France. Clin Infect Dis 2004;38:e96–9.

7. Khin-Maung U, Khin M, Aye T, et al. Clinical features, includinghaemolytic-uraemic syndrome, in Shigella dysenteriae type 1 infectionin children in Rangoon. J Diarrhoeal Dis Res 1987;5:175–7.

8. Parsonnet J, Greene KD, Gerber AR, Tauxe RV, Aguilar OJV, Blake PA.Shigella dysenteriae type 1 infections in US travelers to Mexico. Lancet1989;2:543–5.

9. Rongnoparat C, Panpanit R. Hemolytic uremic syndrome associatedwith shigellosis: report of two cases. Southeast Asian J Trop Med HygPub Hlth 1987;18:226–8.

0. Tarr PI, Gordon CA, Chandler WL. Shiga-toxin-producingEscherichia coli and haemolytic uraemic syndrome. Lancet2005;365:1073–86.

1. Chaisri U, Nagata M, Kurzono H, et al. Localizations of Shiga toxins ofenterohaemorrhagic Escherichia coli in kidneys of paediatric and geri-atric patients with fatal haemolytic uraemic syndrome. Microb Pathog2001;31:59–67.

2. Bennish ML, Khan WA, Begum M, et al. Low risk of hemolyticuremic syndrome after early effective antimicrobial therapy forShigella dysenteriae type 1 infection in Bangladesh. Clin Infect Dis2006;42:356–62.

3. Tzipori S, Sheoran A, Akiyoshi D, Donohue-Rolfe A, Trachtman H. Anti-body therapy in the management of Shiga toxin-induced hemolyticuremic syndrome. Clin Microbiol Rev 2004;17:926–41.

4. Gould LH, Demma L, Jones TF, et al. Hemolytic uremic syndrome anddeath in persons with Escherichia coli O157:H7 infection, foodbornediseases active surveillance network sites, 2000–2006. Clin Infect Dis2009;49:1480–5.

5. Koster F, Levin J, Walker L, et al. Hemolytic-uremic syndrome aftershigellosis. Relation to endotoxemia and circulating immune com-plexes. N Engl J Med 1978;298:927–33.

6. Butler T, Rahman H, Al-Mahmoud KA, et al. An animal modelof hemolytic-uremic syndrome in shigellosis: lipopolysaccharides

of Shigella dysenteriae 1 and S. flexneri produce leucocyte-mediated renal cortical necrosis in rabbits. Br J Exp Path 1985;66:7–15.

7. Su D, Roth RI, Yoshida M, Levin J. Hemoglobin increases mortality frombacterial endotoxin. Infect Immun 1997;65:1258–66.

4

l Medicine and Hygiene 106 (2012) 395– 399 399

8. Ullis KC, Rosenblatt RM. Shiga bacillus dysentery complicated bybacteremia and disseminated intravascular coagulation. J Pediatr1973;83:90–3.

9. Gilman RH, Koster F, Islam S, McLaughlin J, Rahaman MM. Random-ized trial of high- and low-dose ampicillin therapy for treatment ofsevere dysentery due to Shigella dysenteriae type 1. Antimicrob AgentsChemother 1980;17:402–5.

0. Speelman P, Kabir I, Islam M. Distribution and spread of colonic lesionsin shigellosis: a colonoscopic study. J Infect Dis 1984;150:899–903.

1. Butler T, Dunn D, Dahms B, Islam M. Causes of death and thehistopathologic findings in fatal shigellosis. Pediatr Infect Dis J1989;8:767–72.

2. Goldenberg S. Lymphatic connections between the ascending anddescending colon and the corresponding kidneys. Acta Anat (Basel)1959;39:264–70.

3. Kenney BC, Jain D. Identification of lymphatics within the coloniclamina propria in inflammation and neoplasia using the monoclonalantibody D2-40. Yale J Biol Med 2008;81:103–13.

4. Ochoa TJ, Cleary TG. Epidemiology and spectrum of disease ofEscherichia coli O157. Curr Opin Infect Dis 2003;16:259–63.

5. Traa BS, Walker CLF, Munos M, Black RE. Antibiotics for the treatmentof dysentery in children. Int J Epidemiol 2010;39(suppl 1):i70–4.

6. Salam MA, Bennish ML. Therapy for shigellosis. I. Randomized,double-blind trial of nalidixic acid in childhood shigellosis. J Pediatr1988;113:901–7.

7. Alam AN, Islam MR, Hossain MS, Mahalanabis D, Hye HKMA. Com-parison of pivmecillinam and nalidixic acid in the treatment of acuteshigellosis in children. Scand J Gastroenterol 1994;29:313–7.

8. Bhattacharya SK, Bhattacharya MK, Dutta P, et al. Randomized clinicaltrial of norfloxacin for shigellosis. Am J Trop Med Hyg 1991;45:683–7.

9. Bennish ML, Salam MA, Khan WA, Khan AM. Treatment of shigellosis:III. Comparison of one- or two-dose ciprofloxacin with standard 5-daytherapy. Ann Intern Med 1992;117:727–34.

0. Kabir I, Butler T, Khanam A. Comparative efficacies of single intra-venous doses of ceftriaxone and ampicillin for shigellosis in a placebo-controlled trial. Antimicrob Agents Chemother 1986;29:645–8.

1. Pavia AT, Nichols CR, Green DP, et al. Hemolytic-uremic syndromeduring an outbreak of Escherichia coli O157;H7 infections in insti-

tutions for mentally retarded persons: clinical and epidemiologicobservations. J Pediatr 1990;116:544–51.

2. Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI. The riskof the hemolytic-uremic syndrome after antibiotic treatment ofEscherichia coli O157:H7 infections. N Engl J Med 2000;342:1930–6.