'tRANSACTIONS
OF THE
ROYAL
SOCIETY
OF -r:ROPICAL
Treatrnent of bacillary dysentery
versu~; 5-days nalidixic acid
MEDICINE
AND HYGIENE
in Vietnamese
(2000) 94, 323-326
children: two doses of ofloxacin
Ha Vinh1, John Wain2,3, Mai Thu -Chinh1, Cao Thi Ta~Phan
Thi Thu Tr~Diem
Ngal,
Peter Echeverria4,
To Song Diep1, Nicholas J. White2.3 and Christopher
M. Parry2'3*
lCentrefor Tropical
Diseases,190BenHam Tu, DistrictS, Ho ChiMinhCity, Vietnam; 2WellcomeTrust Clinical Research Unit, Centrefor Tropical
Diseases, 190 Ben Ham Tu, District 5, Ho Chi Minh City, Vietnam; 3Centre for Tropical Medicine, Nuffield Department of
Clinicall~edicine,
John Radcliffe Hospital, University of Oxford, Oxford, UK; 4US AFRIMS
component, Bangkok, Thailand
Abstract
Nalidixic acid (NA: 55 mg/kg daily for 5 days) is the recommended treatment for uncomplicated bacillary
dysentery in areas where multidrug-resistant Shigella are prevalent. An open randomized comparison of this
NA regimen with 2 doses of ofloxacin (total 15 mg/kg) was conducted in 1995/96 in 135 Viemamese
children with fever and bloody diarrhoea. Sixty-six children with a bacterial pathogen isolated were eligible
foJ~analysis. Of the 63 Shigella isolates, 39 (62%) were resistant to multiple antibiotics. Resolution times for
fe',er and diarrhoea were similar in the 2 groups, but excretion time of stool pathogen was significan
in the NA recipients [median (range) days 1 (1-9) vs 1 (1-2), p = 0.001] .There were 9 (25%) treatment
failures in the NA regimen and 3 (10%) in the ofloxacin group; p = 0.1. Two patients had NA-resistant
S~ligellaflexneri. One of these isolates was selected during NA treatment. From a clinical and public health
stlmdpoint a 2-dose regimen of ofloxacin is preferable to nalidixic acid in the treatment of bacillary
dysentery.
Kt~ywords: shigellosis,chemotherapy,nalidixic acid, ofloxacin, children, Vier Nam
Introduction
Diarrhoeal disease is a major cause of childhood death
in tropical countries. Widespread use of oral rehydration
solutions has reduced considerably the monality of
watery diarrhoea, yet bacillary dysentery continues to
exen a considerable toll, panicularly in the malnourished
and the very young. Provided that the infecting organism
is sensitive, bacillary dysentery usually responds rapidly
to appropriate antimicrobial treatment (LoLEKHA et at.,
1991). However, the emergence in some areas of Shigella
strains tJ1atare resistant to multiple antibiotics has made
empirical antibiotic treatment more difficult (BENNISH
& SALAM, 1992; JAMAL et al., 1998; l..EGROS et at.,
1998). Where such multiple drug-resistant (MDR)
strains are prevalent the 4-quinolone nalidixic acid is
considered the oral drug of choice for uncomplicated
Shigella infections (WHO, 1996). However, in recent
years reduced quinolone susceptibility has been reponed
from some areas of the tropics (ARMAN et al., 1994).
Nalidixic acid is given 4 times daily over 5 days, and poor
compliance may be common. The fluoroquinolones are
intrinsically more active than nalidixic acid, and several
studies in enteric infections indicate that they may be
effective: in shoner treatment courses (BENNISH &
SALAM, 1992) .Despite the lack of evidence for toxicity
in children these compounds are still not recommended
by some authorities because of their toxicity to the
articulalr canilage of developing weight-bearing bones
in immature experimental animals (beagle dogs). However, nalidixic acid, which is also toxic in this experimental model, was introduced over 30 years ag0--before
concerns over its use in paediatric practice were raised. In
order to improve compliance and to reduce drug exposure, we have compared the currently recommended
standard 5-day nalidixic acid regimen (WHO, 1996)
with a single-day treatment with ofloxacin in children
presenting to hospital with fever and bloody diarrhoea.
Methods
The study was carried out between 1995 and 1996ina
speciali:;t paediatric ward in the Centre for Tropical
Diseases, an infectious diseases referral hospital in Ho
Chi Minh City, Viet Nam. The study was approved by
* Author for correspondence:Dr Christopher M. Parry, Wellcome Trust Clinical ResearchUnit, Centre for Tropical Diseases,190Ben Ham Tu, District 5, Ho Chi Minh City, Vietnam;
phone +848835 3954, fax +848835 3904,
[email protected]
the Scientific and Ethical Committee of the Centre for
Tropical Diseases.
Patients
Children aged >3 months and <15 years, who were
admitted with fever and bloody diarrhoea (> 3 loose
stools with obvious blood) for <5 days were entered into
the study provided that their parents or guardian gave
fully informed consent. Children who were judged by the
admitting doctor to require parenteral treatment, who
gave a history of quinolone treatment within the previous
48 h, or who were known to be allergic to quinolone
drugs, were excluded from the study.
Clinical examination
On admission, a detailed history of the present illness,
and the treatment taken, was documented on a standard
form. Physical examination was performed on admission
and daily thereafter until discharge. Particular note was
taken of any bone or joint symptoms. Blood was taken for
complete blood count. The axillary temperature, pulse
rate, respiratory rate, blood pressure, and frequency and
character of stools were recorded 6-hourly.
Microbiological methods
A microscopy examination of fresh stool was performed to look for haematophagous trophozoites of
Entamoeba histolytica. Stool samples were cultured directly, and after overnight enrichment in selenite F broth
(Oxoid, Basingstoke, UK) on MacConkey and XLD
agar (Oxoid) at 37°C. Stool was also cultured directly on
blood-free Campylobacter selective agar (Oxoid) at 37°C
in a microaerophilic atmosphere. Colonies suggestive of
Salmonella or Shigella were subcultured on to nutrient
agar and were identified using a 'short set' of sugar
fermentation reactions [Kliger iron agar, urea agar,
citrate agar, SIM motility-indole
media (Oxoid)) and
then confirmed using specific antisera. The API 20E test
strip of biochemical reactions (Biomerieux,
Paris,
France) was used to confirm the identity of Shigella
isolates. In order to identify enteroinvasive Escherichia
coli (EIEC) 6 isolated colonies of E. coli were saved from
each patient's admission stool and stored at room tern:
perature on an agar slope before transport to the US
AFRIMS laboratory in Bangkok where they were identified'by DNAhybridization
and confirmed by polymerase
chain reaction (SETHABUTR et al., 1994). A fresh single
stool sample was cultured each day for 5 days, or for
longer if diarrhoea had not resolved.
HA VINH ET A,
Antimicrobial susceptibility was assessed initially by
disc-diffusion using a modified Bauer- Kirby method
(NCCLS,
1997a), and later by minimum inhibitory
concentration agar-dilution method (NCCLS, 1997b).
completely. The parents were asked to bring the children
back to the hospital if fever and/or diarrhoea recurred.
Treatment failures were treated with ofloxacin 5 mg/kg
every 12 h for 5 days.
Drug treatment
Following clinical assessment and enrolment into the
studya sealed envelope was opened which contained the
treatment allocation. Children were randomized in
blocks of 10 to receive oral treatment with either nalidixic
ac:id (Gateway Pharmaceuticals, Australia) 55 mg/kg
daily divided into 4 equal doses for a total of 5 days or
ofloxacin (Roussel-UCLAF,
Paris, France) 7.5 mg/kg
given immediately and then again 12 h later (total dose
I~) mg/kg).
Oral rehydration solution was given if the children
were dehydrated, and this was supplemented with
parenteral fluids if necessary .Paracetamol ( 15 mg/kg)
every 6 h was given if necessary for high fever, and
convulsions were treated with intravenous diazepam
(0.2 mg/kg).
Statistical analysis
Assuming a 90% cure rate in 1 arm of the study, 30
stool culture-positive children were required in each
group to exclude a 35% difference with an 80% power
and 5% confidence level. The analysis concentrated on
the children with a positive stool culture. The MannWhitney U test was used for comparison of continuous
variables, X2 test with Yates' correction was used for
categorical variables or Fisher's exact test if the number
in any cell was less than 5. The log rank test was used to
compare the bacteriological clearance time (EpiInfo
package version 6.0 CDC, Georgia; SPSS for Windows
v 7.5, SPSS Inc., Chicago).
Assessmentparameters
The times from the start of treatment until resolution
of fever and dysentery were recorded as follows. Fever
clearance time was the time until axillary temperature fell
to <37.5°C and remained at or below this value for
>48 h. Clearance of bloody diarrhoea was the time until
the last stool containing visible blood was passed.
Clearance of diarrhoea was the time until the first formed
stool. A clinical failure was defmed prospectively as
persistence of fever and/or diarrhoea (bloody or watery)
for more than 5 days after the beginning of treatment. A
microbiological failure was defined as positive stool
cultures for the original infecting pathogen on day 5 of
treatment or later. The child was considered cured if all
signs and symptoms and pathogenic microorganisms
wc:re cleared within 5 days of treatment.
Patients remained in hospital for at least 5 days and
were discharged when all clinical symptoms had resolved
Results
A total of 135 children were enrolled into the study.
Sixty-one caseswere excluded because the pre-treatment
stool culture was negative. In 74 cases the stool culture
was positive. Eight of these patients were excluded
because the parents discharged them from hospital
prematurely (these children were taken home early when
the clinical state had improved, but before complete
resolution of signs and symptoms). Therefore, 66 cases
with confirmed bacillary dysentery were eligible for
analysis. There were 32 boys and 34 girls. Of the 66
eligible cases, 41 (62%) were aged <3 years (overall
range 8 months-13 years). Overall, 36 of these children
were treated with nalidixic acid and 30 with ofloxacin.
The clinical features and severity of illness were similar in
the 2 groups (Table I).
Microbiology
None of the stool samples had evidence of amoebic
dysentery on microscopy. In total, 68 pathogens were
isolated from the stool cultures of 66 children (in 2 cases
T ~lble I. Signs and symptoms of 66 children in Viet Nam with confirmed bacillary dysentery and allocated to
tr,~atment with nalidixic acid or ofloxacin
-Signs and symptoms
Nalidixic acid (n = 36)
Ofloxacin (n = 30)
Pvalue"
Age (months)
Median
Interquartile range
Thlration of illness before admission (h)
Median
Interquartile range
48
24-72
Watery diarrhoea precedes bloody
,:iiarrhoea, cases (%)
V(]lmiting, cases (%)
History of convulsion, cases (%)
15 (42)
15 (42)
2 (6)
W,~ight (kg)
.Median
:[nterquartile range
White blood count (/mm3)
Median
InterquartiIe range
Platelets (/mm3)
j\iedian
J:nterquartile range
Microorganism isolatedb
.S-higellajlexneri
.S-higellasonnei
EIEC
24
14-51
21.5
17-36
10-0
9-0-12-9
11200
8025-15350
170000
163000-186000
22
12
2
32
1-48
0.11
14 (47)
11 (37)
2 (7)
0.44
0.53
0.89
!0-0
8-9-!4-!
9600
8775-15100
179
000
0'3'
170000-184
20
9
3
-.J
325
"REATMENT OF BACILLARY DYSENTER
2 pathogens were identified in the same stool sample).
The microbiological results are shown in Table 1.
Campylobacterspecieswere not isolated. Overall, 63 of
the 66 culture-identified infections were causedby either
Shigellaflexneri(n = 42) or S. sonnei(n = 21). Neither S.
dysenteriaenor S. boydii was isolated. In 1 case the S.
jlexneri cultured on admissionwasnalidixic acid sensitive
[nalidixic acid minimum inhibitory concentration
(MIC) 4 mg/L; ofloxacin MIC 0.06 mg/L] but, by the
third day of treatment with nalidixic acid, the stool
organism had become resistant (nalidixic acid MIC
32 mg/L; ofloxacin MIC 0.25 mg/L). Antimicrobial
susceptibility testing showed that all the Shigellastrains
wereresistantto at least 1 commonly used antibiotic, and
39 (62%) of the 63 isolateswere simultaneouslyresistant
to multiple antibiotics (chloramphenicol, ampicillin,
trimethoprim and tetracycline). Overall, 39 (62%) of
the 63 Shigellaisolateswere resistantto chloramphenicol
(all MICs ~64 mg/L), 45 (71%) to ampicillin (all MICs
~128 mg/L), 56 (89%) to tetracycline (all MICs
~64 mg/L), 58 (92%) to trimethoprim (all MICs
~256 mg/L) and 2 (3%) to nalidixic acid (MICs 32
and 64 mg/L). The overall MIC9o for nalidixic acid was
8 mg/L. All isolatesof Shigellaand EIEC were sensitive
to ofloxacin (ShigellaMIC9o 0-125 mg/L).
Treatment response
All children responded well to treatment; there were
no deaths or severe complications (in particular there
were no convulsions, shock, renal failure or coma). The
resolution of fever and dysentery was rapid and not
signicantly different between the 2 treatment groups
(Table 2). Clinical responses were similar in the culture-positive and the culture-negative patients (data not
shown). The clearance of Shigella was always rapid
following ofloxacin (range 1-2 days), but some children
continued to excrete Shigella for several days on nalidixic
acid treatment (range 1-9 days) (P = 0.001, Figure). In
this series there were only 5 strains ofEIEC isolated from
the stools, 3 in the ofloxacin group and 2 in the nalidixic
acid group. The symptoms resolved rapidly in each case
except for 1 child with S. flexneri and EIEC treated with
ofloxacin who had a prolonged fever and diarrhoea
clearance time. The cure rate was 75% (27/36) in the
nalidixic acid group and 90% (27/30) in the ofloxacin
group, RR 0-83 (95% confidence interval, 0-67-1-04);
p = 0.1- There were no significant adverseeffects with
either drug treannent, and in particular no evidence of
bone or joint toxicityDiscussion
The 5-day nalidixic acid regimen is the current WHO
recommended treatment for multiple drug-resistant
(MDR) shigellosis (WHO, 1996). This study demonstrates that a single day (2 doses) of treatment with
ofloxacin is at least as effective as the oral 5-day nalidixic
acid regimen for the treatment of childhood bacillary
dysentery in this part of Viet Nam. In this series in
children, most of the confirmed bacillary dysentery was
caused by ShigeUa species, with a small number ofEIEC
and no Campylobacter infections, and the majority of
ShigeUa isolated were resistant to multiple antibiotics.
Empirical treatment of dysentery in this area with drugs
other than quinolones would therefore not be appropriate. The 2-dose ofloxacin regimen proved highly
Table 2. Response to treatment with nalidixic acid or ofloxacin of 66 children with confinned dysentery in
Viet Nam in 1995/96
Nalidixic acid
(n = 36)
Ofloxacin
(n = 30)
30
12-54
24
8-36
Bloody diarrhoea clearance time (h)
Median
lnterquartile range
36
24- 72
29
24-54
Diarrhoea clearance time (h)
Median
lnterquartile range
49
36-96
50
25-81
1
1-2
1-9
I-I
1-2
Parameter
Fever clearance time (h)
Median
lnterquartile range
Bacteria clearance time (days)
Median
lnterquartile range
Range
Pvaluea
0.5
0-001
27 (90,
Cure, cases (%)
27 (75)
).1
3 (10:
Failure, cases (%)
9 (25)
Clinical failure
2
L.
Relapse
1
o
Microbiological failure
3
o
Clinico-microbiological
failure
3
-.All Pvalues arefrom Mann- Whitney Utest exceptfor the Fisher's exacttest usedin outcome comparison, and log rank test for bactel
clearancetime survival analysis.
HA VINH ET AL
effiectiveand well tolerated. S. dysenteriaetype 1 is the
most dangerousof the enteric pathogenscausingdysentery and MDRstrains are prevalent in many areas.This
organism was not isolated in this seriesand indeed has
been isolated rarely in this infectious disease referral
hospital in recent years. Thus it is not possible to
extrapolate from these results to the treatment of Shiga
dysentery.
The small number of children with confirmed bacillary
dysentery and the fact that administration of antibiotics
was not double-blinded are potential limitations of this
study. Despite this, nalidixic acid was associatedwith
delayed eradication of the faecal pathogen in some
children, resulting in a significantly slower pathogen
clearancetime. Administration of 2 oral dosesof fluoroquinolone is also considerablysimpler than the 20-dose
nalidixic acid regimen. Poor compliance with the
4-times-daily nalidixic acid regimen would increase
funher the likelihood of persistent excretion and thus
the risk of transmission. Shigella diaiThoea has a high
secondaryattack rate, especiallyin areaswhere personal
hygiene is difficult because of inadequate facilities. In
such circumstances early reduction of carriage and
excretion will be particularly important. Thus on both
therapeutic and public health grounds, the 1-day, 2dose, fluoroquinolone regimen is preferable.
"Nalidixicacid resistanceoccurred in 2 children with S.
flexneri infections in this study, and in 1 of these it was
selected during treatment. Although these strains remained susceptible to fluoroquinolones, albeit with a
higher MIC value than for fully susceptiblestrains, this
may be a prelude to the development of full resistanceas
in other Enterobacteriaceae. Fluoroquinolone use is
widespread in southern Viet Nam and nalidixic acid
resistancein SalmonellaTyphi in this area is already a
major problem (WAIN et al., 1997).In the treatment of
multidrug-resistant typhoid, we have shown in several
largeseriesthat short-coursefluoroquinolones ( <5 days)
are effective, provided the isolates remain fully susceptible to fluoroquinolones, and safe particularly in children (HIEN et al., 1995; VINH et al., 1996; W AIN et al.,
1997). Long-term follow-up studies in Viet Nam have
alsoshown no adverseeffectson bones,joints, or growth
in children who have received these treatments
(BETHEIL et al., 1996). As nalidixic acid has also been
associatedwith cartilage damage in experimental animals, the 2-dose ofloxacin regimen may be safer than a
20-dosenalidixic acid regimen. Short-course fluoroquinolone treatment has proved very effective in other
studies of the treatment of dysentery and also cholera.
BENNISHetal. (1992) reported that a singlegram doseof
ciprofloxacin is effective therapy for adult patients
infected with Shigella except for S. dysenteriaetype 1.
N orfloxacin in a singledoseof800 mg proved aseffective
astrimethoprim-sulphamethoxazole in acute shigellosis
in Egypt (GOTTUZOet al., 1989).Single-dose ciprofloxacin is alsoeffectivein the treatment of choleracausedby
Vibrio cholerae01 or 0139, and is better than doxycycline in the eradication of v: cholerae
from stool (KHAN et
al., 1996). Two dosesof ofloxacin are a simple, safeand
effective treatment for uncomplicated bacillary dysentery in children.
Acknowledgements
We thank the Director and staff of the Centre for Tropical
Disease for their help in this stUdy. We are very grateful to Dr
Jeremy Farrar for critical reading of this manuscript and
Professor A. Bryskier, Roussel-UCLAF,
for kindly providing
the ofloxacin used in this study. The study was funded by the
WeJIcomeTrust of Great Britain.
References
Arman, D., Willke, A. & Tural, D. (1994). In vitro activity of
eight antibiotics against Salmonella and Shigella species.
EuropeanJoumal of Epidemiology,10,345-347.
Bennish, M. L. & Salam, M. A. (1992). Rethinking the options
for the treatment of shigellosis. Journal of Antimicrobial
Chemorherapy,
30,243-247.
Bennish, M. L., Salam, M. A., Khan, W.A. & Khan, A. M.
(1992). Treatment of shigellosis:ill. Comparison of one- or
tWo-doseciprofloxacin with standard 5-day therapy. A randomized, blinded trial. Annals of Inrernal Medicine, 117,
727-734.
Bethell, D. B., Hi~n,T. T., Phi, L. T., Day, N. P. I.,Vinh, H.,
Duong, N. M., Len, N. M., Chuong, L. V. & White, N. I.
(1996). Effects on growth of single short courses of fluoroquinolones. Archivesof Diseasein Childhood,74,44-46.
Gonuzo, E., Obertelman, R. A., Maguina, C., Berry, S. I., Yi,
A., Guzman, M., Ruiz, R., Leon-Barua, R. & Sack, R. B.
(1989). Comparison of a single dosetreatment with norfloxacin and standard 5 day treatment with trimethoprimsulphamethoxazolefor acute shigellosisin adults. AntimicrobialAgentsand Chemotherapy,33, 1101-1104.
Hien, T. T.,Bethell,D.B.,Hoa,N.
T. T., Wain, I., Diep, T.S"
Phi, L. T., Cuong, B. M., Duong, N. M., Thanh, P. T.,
Walsh,A. L.,Day, N. P.I. &White,N.I. (1995). Short course
of ofloxacin for treatment of multidrug-resistant typhoid.
Clinical InfectiousDiseases,
20, 917-923.
Iamal, W. Y., Rotimi, V. 0., Chugh, T. D. & Pal, T. (1998).
Prevalenceand susceptibility of Shigellaspeciesto 11 antibiotics in a Kuwait teachinghospital. Journal ofChemorherapy,
10, 285-290.
Khan, W. A., Bennish, M. L., Seas,C., Khan, E. H., Ronan, A.,
Dhar, U., Busch, W. & Salam, M. A. (1996). Randomised
controlled comparison of single doseciprofloxacin and doxycycline for cholera caused by Vibrio cholerae01 or 0139.
Lancer,348,296-300.
Legros, D., Ochola, D., Lwanga, N. & Gumma,G. (1998).
Antibiotic sensitivity of endemic Shigellain Mbarara, Uganda. EasrAfrican MedicalJournal, 75, 160-161.
Lolekha, S., Vibulbandhitkit, S. & Poonyarit, P. (1991).
Responseto antimicrobial therapy for shigellosisin Thailand.
Reviewsof InfectiousDiseases,
13 (supplement 4), S342-S346.
NCCLS (1997a). Peifonnancesrandardsfor antimicrobial disk
susceptibilitytests,sixth edition: Approved Standard. NCCLS
Document M2-A6. Villanova, PA: National Committee for
Clinical Laboratory Standards.
NCCLS (1997b).Methodsfor dilution antimicrobialsusceptibility
restsfor bacteria rhar grow aerobically,seventh edition: Approved Standard.NCCLS document M7 -A4 .Villanova, PA:
National Committee for Clinical Laboratory Standards.
Sethabutr, 0., Echeverria, P., Hoge, C. W., Bodhidatta, L. &
Pitarangsi, C. (1994). Detection of Shigellaand enteroinvasive Escherichiacoli by PCR in the stool of patients with
dysenteryin Thailand. Journal of DiarrhoealDiseases
Research,
12,265-269.
Vinh, H., Wain, I., Hanh, V. T. N., Nga, C. N., Chinh, M. T.,
Bethell, D. B., Hoa, N. T. T., Diep, T. S., Dung, N. M. &
White, N. I. (1996). Two or three daysof ofloxacin treatment
for uncomplicated multidrug-resistant typhoid fever in children. AntimicrobialAgentsand Chemorherapy,
40, 958-961.
Wain,I.,Hoa,N.T.T.,Chinh,N.
T., Vinh,H.,Everett,M.I.,
Diep, T.S.,Day,N.P.I.,Solomon, T., White, N. I., Piddock,
L. I. V. & Parry, C. M. (1997). Quinolone-resistant Salmonella ryphi in Vietnam: molecular basis of resistance and
clinical responseto treatment. Clinical InfectiousDiseases,
25,
1404-1410.
WHO (1996). The Managemenrof Bloody Diarrhoea in Young
Children. Geneva: Programme for Control of Diarrhoeal
Diseases,World Health Organization.
Received 8 November
December 1999
1999;
;cepted for publicatioj