www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0 1
Antimicrobial treatment for early, limited
Mycobacterium ulcerans infection: a randomised
Willemien A Nienhuis, Ymkje Stienstra, William A Thompson, Peter C Awuah, K Mohammed Abass, Wilson Tuah, Nana Yaa Awua-Boateng,
Edwin O Ampadu, Vera Siegmund, Jan P Schouten, Ohene Adjei, Gisela Bretzel, Tjip S van der Werf
Background Surgical debridement was the standard treatment for Mycobacterium ulcerans infection (Buruli ulcer
disease) until WHO issued provisional guidelines in 2004 recommending treatment with antimicrobial drugs
(streptomycin and rifampicin) in addition to surgery. These recommendations were based on observational studies
and a small pilot study with microbiological endpoints. We investigated the effi cacy of two regimens of antimicrobial
treatment in early-stage M ulcerans infection.
Methods In this parallel, open-label, randomised trial undertaken in two sites in Ghana, patients were eligible for
enrolment if they were aged 5 years or older and had early (duration <6 months), limited (cross-sectional diameter
<10 cm), M ulcerans infection confi rmed by dry-reagent-based PCR. Eligible patients were randomly assigned to
receive intramuscular streptomycin (15 mg/kg once daily) and oral rifampicin (10 mg/kg once daily) for 8 weeks
(8-week streptomycin group; n=76) or streptomycin and rifampicin for 4 weeks followed by rifampicin and
clarithromycin (7·5 mg/kg once daily), both orally, for 4 weeks (4-week streptomycin plus 4-week clarithromycin
group; n=75). Randomisation was done by computer-generated minimisation for study site and type of lesion
(ulceration or no ulceration). The randomly assigned allocation was sent from a central site by cell-phone text message
to the study coordinator. The primary endpoint was lesion healing at 1 year after the start of treatment without lesion
recurrence or extensive surgical debridement. Analysis was by intention-to-treat. This trial is registered with
ClinicalTrials.gov, number NCT00321178.
Findings Four patients were lost to follow-up (8-week streptomycin, one; 4-week streptomycin plus 4-week
clarithromycin, three). Since these four participants had healed lesions at their last assessment, they were included in
the analysis for the primary endpoint. 73 (96%) participants in the 8-week streptomycin group and 68 (91%) in the
4-week streptomycin plus 4-week clarithromycin group had healed lesions at 1 year (odds ratio 2·49, 95% CI 0·66 to
infi nity; p=0·16, one-sided Fisher’s exact test). No participants had lesion recurrence at 1 year. Three participants had
vestibulotoxic events (8-week streptomycin, one; 4-week streptomycin plus 4-week clarithromycin, two). One
participant developed an injection abscess and two participants developed an abscess close to the initial lesion, which
was incised and drained (all three participants were in the 4-week streptomycin plus 4-week clarithromycin group).
Interpretation Antimycobacterial treatment for M ulcerans infection is eff ective in early, limited disease. 4 weeks of
streptomycin and rifampicin followed by 4 weeks of rifampicin and clarithromycin has similar effi cacy to 8 weeks
of streptomycin and rifampicin; however, the number of injections of streptomycin can be reduced by switching to
oral clarithromycin after 4 weeks.
Funding European Union (EU FP6 2003-INCO-Dev2-015476) and Buruli Ulcer Groningen Foundation.
Buruli ulcer is a necrotising infection of subcutaneous
tissue caused by Mycobacterium ulcerans.1 The name Buruli
ulcer comes from a region near the Nile River delta in
Uganda, named Buruli County, where the disease was
highly endemic in the 1960s.2 Today, the disease is emerging
in west African countries with thousands of cases every
year, mainly in children.3,4 A plasmid of M ulcerans encodes
the production of mycolactone,4,5 an immuno modula tory
macrolide toxin that causes tissue necrosis.6 M ulcerans is
acquired near slow-fl owing and stagnant water in tropical
and subtropical environments. The natural reservoir and
mode of transmission of the infection remain largely
obscure and might diff er between endemic foci around the
world.7,8 However, skin injury9 and insect bites10 have been
proposed as modes of transmission.
M ulcerans infection usually starts as a nodule, papule,
plaque, or oedema. When left alone, the lesion breaks
open and a typical painless ulcer with undermined
edges appears, which can progress to a large necrotic
lesion. WHO has defi ned lesions with a cross-sectional
diameter of less than 5 cm as category I, 5–15 cm as
category II, and more than 15 cm, lesions on important
sites (eye, breast, and genitalia), or multiple lesions as
category III. M ulcerans infection can be self-limiting,
but scar tissue and contractures in joints leave patients
February 4, 2010
Department of Internal
Medicine (W A Nienhuis MD,
Y Stienstra MD), Department of
(J P Schouten MSc), and
Infectious Diseases Service and
Tuberculosis Unit, Department
of Internal Medicine and
Department of Pulmonary
Diseases and Tuberculosis
(Prof T S van der Werf MD),
University Medical Centre
Groningen, University of
Agogo Presbyterian Hospital,
(W A Thompson MD, K M Abass);
Hospital, Nkawie, Ghana
(P C Awuah MD, W Tuah);
Kumasi Centre for
Collaborative Research in
Tropical Medicine, Kwame
Nkrumah University of Science
and Technology, Kumasi,
Ghana (N Y Awua-Boateng,
Prof O Adjei PhD); National
Buruli Ulcer Program, Accra,
Ghana (E O Ampadu MD); and
Department of Infectious
Diseases and Tropical Medicine,
of Munich, Munich, Germany
(V Siegmund PhD, G Bretzel MD)
Prof Tjip S van der Werf,
University Medical Centre
Groningen, PO Box 30 001,
9700 RB Groningen, Netherlands
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0
with functional limitations and can result in social
stigma.11,12 The diagnosis can be made clinically but
culture is the gold standard. However, this method is
diffi cult and has low sensitivity.1,3,4,13 Since the develop-
ment of PCR targeting insertion sequence 2404
(IS2404)—a repetitive oligonucleotide unit with more
than 200 copies in the genome of M ulcerans14—
diagnostic confi rmation has improved substantially.13,15,16
Buruli ulcer is one of 19 neglected tropical diseases
addressed by WHO in its Global plan to combat neglected
tropical diseases 2008–2015.17 In this plan, the organisation
describes Buruli ulcer as a disease for which there are no
cost-eff ective control methods. Since the disease’s fi rst
description in 1948,18 diff erent treatments have been
investigated. Extensive surgical debridement, with or
without subsequent skin grafting, is standard treatment.
However, surgery cannot com pletely remove all bacilli19
and recurrence is common, with reported rates varying
between 6% and 47%.20–22 Although larger excisions might
be more eff ective, they can increase chances of residual
functional limitations. In the fi rst of two randomised
controlled trials for M ulcerans infection, clofazimine did
not show a signifi cant benefi t compared with placebo.23
In individuals with small (<5 cm), non-ulcerated lesions,
recurrence-free healing without surgery was reported in
fi ve of eight participants who were treated with
clofazimine compared with fi ve of 17 who were treated
with placebo. In ten patients with larger and ulcerated
lesions, all except one (in the placebo group) needed
surgery. A second study compared the eff ect of dapsone
plus rifampicin with placebo. Of 41 randomised patients,
30 completed the 2-month trial. Rate of healing did not
diff er between groups. Uneven baseline characteristics
might partly explain why patients assigned to active
treatment had a larger reduction in lesion size than did
patients assigned to placebo.24
Many antimycobacterial agents show activity against
M ulcerans in vitro, and experiments in animals, such as
the mouse footpad model, show that streptomycin in
combination with rifampicin is highly bactericidal. In a
pilot study sponsored by WHO, 31 patients clinically
diagnosed with pre-ulcerative M ulcerans infection were
treated with streptomycin and rifampicin for 0, 2, 4, 8, or
12 weeks.25 All lesions were excised; M ulcerans infection
was confi rmed by PCR in 21 cases. In ten patients who
were treated for 2 weeks or less, viable bacilli could be
isolated from excised tissues, whereas M ulcerans could
not be cultured from tissue taken from 11 patients who
were treated for 4 weeks or longer. Lesions either reduced
or stabilised in size in all patients.25 On the basis of these
fi ndings, preliminary guidelines were issued by WHO
recommending streptomycin in combination with
rifampicin as standard treatment for M ulcerans infection,26
with or without additional surgical debridement or skin
grafting.27 When our study was designed, clarithromycin
was believed to have only bacteriostatic activity in vivo.3,28
We assessed the effi cacy of antibiotic therapy with oral
rifampicin and intramuscular streptomycin given for
8 weeks for treatment of early M ulcerans infection in
patients from Ghana. This regimen was compared with
rifampicin and streptomycin given for 4 weeks, followed
by an oral combination of clarithromycin and rifampicin
for 4 weeks. Our aim was to identify an eff ective
alternative treatment to extensive surgical debridement,
and to explore possibilities to keep the use of injectable
antimicrobial treatment to a minimum.
The study design was partly based on discussions within
a WHO expert group on Buruli ulcer that took place
between 2001, and 2003. Between April, 2006, and
January, 2008, patients were recruited at two sites (Nkawie-
Toase Government Hospital, Nkawie, and Agogo
Presbyterian Hospital, Agogo) in Ghana. Patients
clinically diagnosed with M ulcerans disease were recruited
by active case fi nding. Patients were eligible for enrolment
18 lesion too large
16 too young
2 hearing impairment
7 recurrent disease
20 doubtful diagnosis and PCR and AFB negative
3 withdrew/lost before randomisation
1 accidentally operated
26 with clinical disease but without PCR confirmation*
143 with infection confirmed by PCR
5 with infection confirmed by other method‡
3 without confirmation of infection
75 assigned to streptomycin and
rifampicin for 4 weeks followed by
rifampicin and clarithromycin for
76 assigned to streptomycin and
rifampicin for 8 weeks
1 died (in week 16)†§
2 migrated (in week 32 and week 36)§
1 withdrew (in week 6)§
76 assessed for primary endpoint 75 assessed for primary endpoint
154 enrolled and started on treatment
250 patients screened
Figure 1: Trial profi le
AFB=acid-fast bacilli.*Patients not enrolled but given 8 weeks of treatment with streptomycin and rifampicin.
†Participant died of cause unrelated to M ulcerans infection. ‡See text for details. §Healed at time of last
assessment, included in the fi nal analysis.
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0 3
if they were aged 5 years or older, had a reported disease
duration of less than 6 months, and had lesions with a
cross-sectional diameter (indurated area) of 10 cm or less.
M ulcerans infection was confi rmed by IS2404 dry-reagent-
based PCR.29 Exclusion criteria were pregnancy, drug in-
tolerance, and renal, hepatic, and acoustic impairment.
The protocol and consent forms were approved by the
Committee on Human Research, Publication, and Ethics
of the School of Medical Science, Kwame Nkrumah
University of Science and Technology, Kumasi, and
the Komfo Anokye Teaching Hospital, Kumasi
(CHRPE/07/01/05), and by the Ethical Review Committee
of Ghana Health Services (GHS-ERC-01/01/06). The
Medical Ethics Review Committee of the University
Medical Centre Groningen, Netherlands, reviewed the
protocol before ethics clearance in Ghana. Written and
verbal informed consent was obtained from all participants
aged 12 years or older, and from parents, carers, or legal
representatives of participants aged 18 years or younger.
After participants had given informed consent, we
obtained demographic and clinical information and took
blood samples. We undertook pregnancy tests in female
participants aged 10 years or older, and hearing tests in
all participants (AS208 portable equipment; Inter-
acoustics, Assens, Denmark) to obtain baseline audio-
metric data. HIV antibody testing was done with
cold-stored sera after completion of the study.
Lesions were photographed and traced onto acetate
sheets. Three 3 mm punch biopsy samples were taken
under local anaesthesia; two swabs of ulcerated lesions
were also taken. All samples were transported to the
Kumasi Centre for Collaborative Research in Tropical
Medicine laboratory in Kumasi, Ghana, for IS2404 dry-
reagent-based PCR and Ziehl-Neelsen staining to detect
acid-fast bacilli; mycobacterial culture was done on
Löwenstein-Jensen slopes at 32ºC.13 One punch biopsy
was reserved for histopathological examination.
Participants started streptomycin (15 mg/kg once daily
intra muscularly) and rifampicin (10 mg/kg once daily
orally) after the diagnostic procedures. After assess-
ments and start of treatment at the hospital, most parti-
cipants were treated as outpatients. Once a week,
parti ci pants were given study drugs to take to the nearest
health facility to receive directly observed treatment
(DOT) for the subsequent days, with daily wound care.
Only parti cipants that had extensive oedema or lesions
at diffi cult sites (joints, eye, or genitalia), or lesions with
suspected secondary infection were admitted to hospital;
participants who could not receive DOT or wound care
at home were also admitted to hospital. DOT was
recorded on forms by the health-care worker or helper
who was observing the treatment. Participants were
followed up at weekly inter vals during the fi rst 8 weeks.
At these visits, clinical assess ments and digital
photographs were taken, DOT forms were checked, and
participants were invited to report any adverse events.
Once every 2 weeks, the size of the lesion was traced
onto an acetate sheet and blood cell counts were taken;
we also undertook liver and kidney function tests and
hearing tests in all participants, and pregnancy tests in
female participants aged 10 years or older.
Randomisation and masking
Before the end of week 4, participants with M ulcerans
infection confi rmed by PCR were randomly assigned to
receive streptomycin intramuscularly and rifampicin
orally for 4 more weeks (8-week streptomycin group) or
rifampicin and clarithromycin (7·5 mg/kg once daily),
both orally, for another 4 weeks (4-week streptomycin
plus 4-week clarithromycin group). Randomisation was
done with minimisation for study site and type of lesion
(ulceration or no ulceration). The study coordinator
(WAN) forwarded the information of every enrolled
participant by cell-phone text messaging to a statistician
(JPS) at the Department of Epidemiology, University
Medical Centre Groningen, Netherlands. There, a
computer-generated randomisation program was used,
and the randomly assigned allocation was then sent by
text message to the study coordinator. Individuals who
were clinically diagnosed with M ulcerans disease but
4-week streptomycin plus
Body-mass index (kg/m²)
Duration of disease (weeks)
Lesion surface area (cm²)
Category of lesion
II or III
Type of lesion
Lesion distribution (side of body)
Data are n (%) or median (IQR). Patients in the 8-week streptomycin group were
assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks.
Patients in the 4-week streptomycin plus 4-week clarithromycin group were
assigned to receive streptomycin and rifampicin for 4 weeks followed by
rifampicin and clarithromycin, both orally, for 4 weeks.
Table 1: Patient baseline characteristics
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0
who did not have confi rmation by PCR continued
treatment with streptomycin plus rifampicin and
were not randomised; these individuals were followed up
and analysed separately. This was an open-label trial.
Follow-up and study outcomes
After 8 weeks of antimicrobial treatment, missed doses
were not supplemented. Participants were followed up
at week 10 and week 12 after start of treatment, and
then monthly to week 36, and bimonthly to week 52.
Study visits included clinical assessment with reporting
of adverse eff ects, measurement of lesion size (if not
healed) by tracing onto an acetate sheet, and
photography of the lesion. Participants’ travel costs
were reimbursed and small monthly incentives (sugar,
condensed milk, and cocoa powder) were off ered for
time spent in the study.
Treatment failure was recorded if a participant’s lesion
had not healed by week 52, lesion recurrence occurred
within 1 year, or lesion size increased to 150% or more at
any timepoint compared with baseline with surgical
debridement undertaken as deemed necessary by the
attending doctor in the hospital. Neither the investigators
who took measurements of the lesions, nor the attending
doctor in the hospital making the fi nal decision for
extensive surgical debridement were masked to treatment
assignment. Removal of necrosis and slough is part of
normal wound care and skin grafting speeds up healing
Stage Size of lesion
Participants with treatment failure recorded before week 52
1 Agogo 8-week
2 Agogo 8-week
Male6 II Ulcer 120×98 Extensive
Male10 III Ulcer and
3 Agogo 4-week
Female12II 73×60 Lesion progression,
8 Surgical resected tissuePCR negative, ZN negative,
Female 11I Nodule30×24 Ulceration of lesion;
4 additional weeks of
New ulceration after
12 Punch biopsy and swab PCR negative, ZN negative,
Female5I Plaque48×47 34 Swab and surgical
PCR negative, ZN negative,
Participants who were not healed at time of primary endpoint (week 52)
Male 22 II (with oedema
leading to III)
Ulcer 112×80 ···· ····
Female20 Ulcer and
··Swab (32), swab (52)32 weeks: PCR positive,
ZN negative, culture postive;
52 weeks: PCR positive,
ZN positive, culture negative
29 weeks: PCR negative,
ZN negative, culture negative;
72 weeks: PCR negative,
ZN negative, culture positive;
80 weeks: PCR negative,
ZN negative, culture negative
Male 12IIUlcer95×95·· Swab (29), swab (72),
113×86HIV positive ···· ··
Treatment failure was recorded if a participant’s lesion had not healed by week 52, lesion recurrence occurred within 1 year, or lesion size increased to 150% or more at any timepoint compared with baseline with
surgical debridement undertaken as deemed necessary by the attending doctor in the hospital. *Weeks after start of treatment. †Results for insertion sequence 2404 dry-reagent-based PCR, Ziehl-Neelsen (ZN)
staining to identify acid-fast bacilli, and M ulcerans culture. Patients in the 8-week streptomycin group were assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week
streptomycin plus 4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks.
Table 2: Characteristics of ten participants with treatment failure
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0 5
but does not aff ect bacterial load. These interventions were
therefore not regarded as evidence of treatment failure.
The primary clinical endpoint was lesion healing
(complete re-epithelialisation) at 1 year after the start of
treatment without recurrence or extensive surgical
debridement. Secondary outcomes were time to wound
healing and time to complete wound coverage by a crust.
Daily sterile dressings were only applied at the health
facility if lesions were open and discharging.
Before fi nal healing occurs, lesions might turn dry with
a crust. At this stage, participants could cover the lesions
for protection at home, without visiting the health facility
to receive wound care and sterile dressings. Since
participants reported this stage of wound healing as
benefi cial, we also measured time to complete wound
coverage by a crust without complete re-epithelialisation
as a secondary endpoint. The safety outcome measure
was occurrence of adverse events.
When the study was designed, there was no information
available about healing rates for the proposed regimens;
therefore, we assumed a healing rate of 80% in the
8-week streptomycin group. We calculated that a sample
size of 148 randomised and fully assessable participants
(74 in each group) would be needed to detect a diff erence
in healing rate of 20% or more (<60% in the 4-week
streptomycin plus 4-week clarithromycin group) with a
one-sided alpha of 0·05 and a power of 80%.
We calculated an odds ratio for the primary clinical
endpoint by use of Fisher’s exact test. Because secondary
outcome data were interval-censored, we analysed the
cumulative incidence of healing by use of actuarial life
table analysis and weighted log-rank tests for interval-
censored data, in particular the group proportional
hazards model30 and a generalised Wilcoxon-Mann-
Whitney test,31 which emphasises early events. We
calculated the exact permutation p value for the scores of
the group proportional hazards model and Wilcoxon-
Mann-Whitney tests and the non-parametric maximum
likelihood estimate of the survival distribution function.32
Other secondary outcome measures were assessed by
actuarial life table analysis. All analyses were by intention
to treat. Statistical analysis was done with SPSS version
16.0, R version 2.9.2, and Stata version 10.1.
An independent data safety monitoring board
reviewed the data for safety purposes after inclusion of
57 and 115 participants. Interim reports were discussed
at the annual WHO meeting on Buruli ulcer in Geneva
in 2007 and 2008, and presented at the 2008 combined
ICAAC/IDSA Annual Meeting in Washington, DC,
USA.33 After the trial had been completed, two
independent wound experts from University Medical
Centre Groningen, who were masked to treatment
assignment, assessed the primary study end point
(healing at 1 year) using the digital photographs of the
lesions taken during the trial. This trial is registered
with ClinicalTrials.gov, number NCT00321178.
Role of the funding source
The sponsors of the study had no role in study design,
data collection, data analysis, data interpretation, writing
8-week streptomycin group 4-week streptomycin plus 4-week clarithromycin group
Patients in the 8-week streptomycin group were assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus
4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks.
Table 3: Actuarial life table for cumulative proportion of healing for both treatment groups, by time interval
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0
of the report, or in decisions about submission of results
for publication. The corresponding author had full access
to all the data in the study and had fi nal responsibility for
the decision to submit for publication.
Figure 1 shows the trial profi le. 180 eligible patients
started treatment. 26 patients with suspected but
unconfi rmed M ulcerans infection received streptomycin
and rifampicin for 8 weeks. Of 151 participants who were
enrolled and randomised, eight had a clinical diagnosis
without confi rmation of M ulcerans infection by PCR.
Five of these eight participants had infection later
confi rmed by one or more diagnostic tests (Ziehl-Neelsen
staining, two; culture, one; histopathology, two). Three
randomised participants did not have diagnosis
confi rmed by any test. Table 1 shows baseline
characteristics of study participants. Lesions were more
frequently seen on the right side of the body (64%) than
on the left side (36%; p<0·0001). Three (2%) participants
were HIV positive; these individuals had initial lesions
and clinical presentations that were indiscernible from
those of HIV-negative participants.
One participant in the 8-week streptomycin group
withdrew from the study at week 6. In the 4-week
streptomycin plus 4-week clarithromycin group, two
participants moved out of the study area and were lost to
follow-up (week 32 and 36) and one participant, who later
tested positive for HIV infection, died in week 16 of
urosepsis. Since these four participants had healed
lesions at their last assessment, they were included in the
analysis for the primary clinical endpoint and in the
analyses for time to healing.
Compliance to study treatment was assessed by use of
DOT forms, signed by health personnel at the health
facilities. Adherence to treatment protocol was 98% in
the 8-week streptomycin group and 99% in the 4-week
streptomycin plus 4-week clarithromycin group.
Treatment failure was recorded in ten participants,
three in the 8-week streptomycin group and seven in
the 4-week streptomycin plus 4-week clarithromycin
group. Table 2 shows the characteristics of these
individuals. Five participants were not healed at
week 52, all of whom had a substantial decrease in
lesion size. One participant had several lesions, and
four had large lesions at the start of treatment (one of
whom had HIV infection). Of the fi ve participants with
treatment failure before week 52, two had large lesions,
one had a pre-ulcerative lesion that ulcerated later, one
had a progressive lesion, and one had a lesion that
Cumulative proportion of participants with healed lesions
48 12 162024 28
3236 4044 48 52
Number at risk
8-week streptomycin group
4-week streptomycin plus
4-week clarithromycin group
8-week streptomycin group
4-week streptomycin plus
4-week clarithromycin group
Group proportional hazards model p=0·26 (99% CI 0·22–0·29)
Generalised Wilcoxon-Mann-Whitney test p=0·60 (99% CI 0·56–0·64)
Figure 2: Non-parametric maximum likelihood estimates for time to healing
The non-parametric maximum likelihood estimates for each treatment group are plotted with shaded rectangles denoting the indeterminate rises in the proportion
healed during each time interval. Linear interpolation lines of healing within these indeterminate regions are also shown. Patients in the 8-week streptomycin group
were assigned to intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus 4-week clarithromycin group were assigned to
receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks.
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0 7
almost healed, but opened up again. No participants
with healed lesions had a recurrence at week 52.
73 (96%) participants in the 8-week streptomycin group
and 68 (91%) participants in the 4-week streptomycin
plus 4-week clarithromycin group had healed lesions at
week 52 (odds ratio for failure in healing for 4-week
streptomycin plus 4-week clarithromycin vs 8-week
streptomycin 2·49, 95% CI 0·66 to infi nity, p=0·16). We
obtained consistent fi ndings when the four participants
who were not followed up to week 52 were excluded from
the analysis and when two wound experts masked to
treatment assignment assessed the primary endpoint by
use of photographs available at the diff erent timepoints
(data not shown).
Table 3 shows the actuarial life table for cumulative
proportion of healing. The estimated cumulative
proportion of patients healed at week 52 was 0·99 (95%
CI 0·94–1·00) in the 8-week streptomycin group and
0·96 (95% CI 0·88–0·99) in the 4-week streptomycin
plus 4-week clarithromycin group; a diff erence of 0·034
(95% CI –0·024 to 0·091) between groups. Figure 2
shows the non-parametric maximum likelihood estimates
for healing in the intention-to-treat population. Neither
the group proportional hazards model (p=0·26; 99% CI
0·22–0·29) nor the generalised Wilcoxon-Mann-Whitney
test (p=0·60; 99% CI 0·56–0·64) showed a signifi cant
diff erence in time to healing between groups. The group
proportional hazards model suggested a shorter time to
healing in the 8-week streptomycin group whereas the
Wilcoxon-Mann-Whitney test suggested that time to
healing was shorter in the 4-week streptomycin plus
4-week clarithromycin group. Adjustment for study site
and type of lesion (ulceration or no ulceration) did not
aff ect the results (data not shown).
Five participants received skin grafts, four in the 8-week
streptomycin group (at week 16, 24, 24, and 28), and one
in the 4-week streptomycin plus 4-week clarithromycin
group (at week 20). Time to healing of category I lesions
(median 18 weeks, 95% CI 14–22) was signifi cantly
shorter than that for category II and III lesions (30 weeks,
95% CI 26–34, p=0·002; data pooled for the two treatment
groups; fi ve participants with skin grafts not included).
Time to complete wound coverage by a crust was
also signifi cantly shorter for category I lesions than
category II and III lesions (14 weeks, 95% CI 11–18, vs
22 weeks, 95% CI 22–26; p=0·002).
Three participants had vestibulotoxic events, one in
the 8-week streptomycin group (aged 49 years, starting
after 7 weeks of treatment) and two in the 4-week
streptomycin plus 4-week clarithromycin group (aged
24 years and 38 years, starting after 4 weeks and 3 weeks
of treatment, respectively).
photographs showed that three participants had mild to
moderate functional limitations at the end of the study:
one had a contracture with substantial decrease in range
of movement of the thumb and index fi nger (4-week
streptomycin plus 4-week clarithromycin group); two
had ulcers on the back of the hand and wrist that resulted
in claw-hands (one in each group). No liver or kidney
function test abnormalities or audiological deterioration
occurred that necessitated termination of streptomycin
treatment. One participant developed an injection
abscess (4-week streptomycin plus 4-week clarithro -
mycin group) and two participants (both in the 4-week
strepto mycin plus 4-week clarithromycin group)
developed an abscess close to the initial lesion which
was incised and drained. One participant in the 8-week
streptomycin group and two participants in the 4-week
streptomycin plus 4-week clarithromycin group reported
Some participants had additional diagnostic tests not
specifi ed in the protocol. Table 4 shows the characteristics
of the fi ve participants in whom M ulcerans was isolated
by culture after treatment; all were in the 4-week
streptomycin plus 4-week clarithromycin group. Three of
these fi ve participants had treatment failure: in two,
surgical debridement was done; in the third, multiple
nodules ulcerated successively over 52 weeks before fi nal
healing. Two participants had lesion healing without
further intervention within the study period.
Analysis of digital
Our study has shown that early, limited M ulcerans
infection can be safely and eff ectively managed by
debridement. The drug regimen proposed by WHO,
alone, without surgical
Stage Indication for cultureTimepoint
Additional information Diagnostic specimen
Development of a second lesion
Confi rmation of nodule
Extensive surgical debridement†
Lesion healed without further intervention
Lesion healed without further intervention
Ulcer healed; multiple nodules ulcerated
Inadequate wound care; surgical
Surgically resected tissue
Patient 5 NkawieMale12 IINo complete healing at week 52 72Swab
This analysis was not specifi ed by the protocol. All fi ve participants were in the 4-week streptomycin plus 4-week clarithromycin group (streptomycin and rifampicin for 4 weeks followed by rifampicin and
clarithromycin for 4 weeks). *Time of tissue specimen collection (weeks after start of treatment). †Participant with treatment failure.
Table 4: M ulcerans isolated by culture of tissue specimens after 8-week treatment period
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0
consisting of 8 weeks of streptomycin and rifampicin,
seemed eff ective and was not associated with
deterioration requiring subsequent surgical debridement.
Treatment with oral clarithromycin plus rifampicin
during the second 4-week period resulted in similar
outcomes to continuation of treatment with streptomycin
and rifampicin. Our fi ndings are important for patients
with M ulcerans infection who live in remote, resource-
poor areas in west Africa, where people often need to
walk for several hours to reach health-care facilities,
skilled personnel are scarce, and patients tend to refrain
from treatment because of fear of surgery. Our results
also support the use of antimicrobial treatment in
individuals who are unable to receive streptomycin—
eg, pregnant women or those who cannot tolerate
aminoglycosides. With few reported side-eff ects, the
treatment regimens used in this trial seemed well
tolerated, although vestibulotoxicity remains a concern.
The rate of lesion recurrence in our study at 52 weeks
was lower than that reported in retrospective studies
assessing the eff ect of surgery, in which rates of between
6% and 47% were reported.20–22
Time to healing was a median of 18 weeks for category I
lesions and 30 weeks for larger lesions. The length of this
healing period might have obscured the potential of
antimicrobial treatment in earlier studies that either
looked at healing after 2 months,24 or assigned participants
to surgery when early healing was not seen during follow-
up.23 HIV was not an important confounder; most case-
control studies from west Africa report a low incidence of
HIV in patients with M ulcerans infection.34,35
One strength of this study is the large proportion of
participants (147 of 151) who were followed up to week 52.
Second, most participants (148 of 151) had laboratory-
confi rmed M ulcerans infection, which contrasts with
previous trials that were partly undertaken before PCR-
based diagnostic confi rmation tests were avail able.23,24
Finally, the sample size in our study was substantially
larger than that in earlier studies.
A potential weakness of our study is the open-label
design. However, masking would have substantially
increased costs, and a trial in which children can be
assigned to placebo injections is not justifi ed for safety
reasons. Moreover, although only one injection abscess
was recorded, intramuscular injections in rural Africa
are not the preferred option. Another limitation of our
study is that no formal external monitoring was done.
However, limited auditing was organised. Additionally,
consistent results were obtained when two wound experts
who were masked to treatment assignment reviewed all
digital photographs available at the diff erent timepoints.
We therefore believe that the study was robust.
One concern is that healing took a fairly long time.
Additionally, we could not address the issue of
prevention of disabilities in a formal way, although our
analysis of digital photographs combined with clinical
assessment showed that only three participants had
mild to moderate functional limitations at week 52 (all
three involving hand function). Contractures and
functional limitations are common in ulcers that are
close to joints.11 Future studies should assess prevention
of disabilities, include all categories of lesions, and
investigate oral drug regimens.
Thus, antimicrobial treatment is highly eff ective for
treatment of early, limited M ulcerans infection, and the
number of intramuscular injections of streptomycin can
be reduced without compromising effi cacy.
TSvdW and YS designed and supervised the study. WAN coordinated the
study. WAN, WAT, PCA, and EOA were responsible for patient
screening and enrolment. KMA, WT, and WAN provided patient care
and requested informed consent from participants, participants’ parents,
or legal representatives, and collected the clinical and laboratory data.
GB, VS, NYA-B, and OA were responsible for the laboratory
confi rmation. JPS, WAN, and YS did the statistical analyses. TSvdW,
WAN, and YS contributed to the interpretation of the results and the
writing and critical review of the report. All authors have seen and
approved the fi nal version of the report.
Confl icts of interest
We declare that we have no confl icts of interest.
We thank all study participants and health-care workers involved in the
trial. We thank Richard O Phillips for invaluable help during the start of
the trial, Cari J Stek for occasional monitoring activities, the “de Sprong”
outpatient pharmacy team, University Medical Centre Groningen for
drug supply, Marieke M de Waard and Janneke Huizinga for
undertaking masked assessment of digital images of the participants’
lesions, William R Faber, Paul D R Johnson, and Alan J Knell for serving
on the data safety monitoring board, and Kingsley Asiedu, Global Buruli
Ulcer Initiative, WHO, Geneva, Switzerland. This study was supported
by the European Union (EU FP6 2003-INCO-Dev2-015476) and Buruli
Ulcer Groningen Foundation. YS received a ZonMW AGIKO grant
(Netherlands Foundation for Research). WAN is currently at the
Department of Internal Medicine, Medical Center Leeuwarden,
Leeuwarden, Netherlands. YS is currently at Department of Internal
Medicine, Division of infectious Diseases, Tropical Medicine and AIDS,
Academic Medical Centre, University of Amsterdam, Amsterdam,
1 Van der Werf TS, Van der Graaf WTA, Tappero JW, Asiedu K.
Mycobacterium ulcerans infection. Lancet 1999; 354: 1013–18.
2 Clancey JK, Dodge OG, Lunn HF, Oduori ML. Mycobacterial skin
ulcers in Uganda. Lancet 1961; 278: 951–54.
3 van der Werf TS, Stienstra Y, Johnson RC, et al. Mycobacterium
ulcerans disease. Bull World Health Organ 2005; 83: 785–91.
4 Wansbrough-Jones M, Phillips R. Buruli ulcer: emerging from
obscurity. Lancet 2006; 367: 1849–58.
5 George KM, Chatterjee D, Gunawardana G, et al. Mycolactone: a
polyketide toxin from Mycobacterium ulcerans required for virulence.
Science 1999; 283: 854–57.
6 van der Werf TS, Stinear T, Stienstra Y, van der Graaf WTA,
Small PL. Mycolactones and Mycobacterium ulcerans disease.
Lancet 2003; 362: 1062–64.
7 Johnson PD, Azuolas J, Lavender CJ, et al. Mycobacterium ulcerans
in mosquitoes captured during outbreak of Buruli ulcer,
southeastern Australia. Emerg Infect Dis 2007; 13: 1653–60.
8 Benbow ME, Williamson H, Kimbirauskas R, et al. Aquatic
invertebrates as unlikely vectors of Buruli ulcer disease.
Emerg Infect Dis 2008; 14: 1247–54.
9 Meyers WM, Shelly WM, Connor DH, Meyers EK. Human
Mycobacterium ulcerans infections developing at sites of trauma to
skin. Am J Trop Med Hyg 1974; 23: 919–23.
10 Portaels F, Elsen P, Guimaraes-Peres A, Fonteyne PA, Meyers WM.
Insects in the transmission of Mycobacterium ulcerans infection.
Lancet 1999; 353: 986.
Articles Download full-text
www.thelancet.com Published online February 4, 2010 DOI:10.1016/S0140-6736(09)61962-0 9
11 Stienstra Y, van Roest MH, van Wezel MJ, et al. Factors associated
with functional limitations and subsequent employment or schooling
in Buruli ulcer patients. Trop Med Int Health 2005; 10: 1251–57.
12 Stienstra Y, van der Graaf WTA, Asamoa K, van der Werf TS. Beliefs
and attitudes towards Buruli ulcer Ghana. Am J Trop Med Hyg 2002;
13 Herbinger KH, Adjei O, Awua-Boateng NY, et al. Comparative
study of the sensitivity of diff erent diagnostic methods for the
laboratory diagnosis of Buruli ulcer disease. Clin Infect Dis 2009;
14 Portaels F, Aguiar J, Fissette K, et al. Direct detection and
identifi cation of Mycobacterium ulcerans in clinical specimens by
PCR and oligonucleotide-specifi c capture plate hybridization.
J Clin Microbiol 1997; 35: 1097–100.
15 Stienstra Y, van der Werf TS, Guarner J, et al. Analysis of an IS2404-
based nested PCR for diagnosis of Buruli ulcer disease in regions of
Ghana where the disease is endemic. J Clin Microbiol 2003; 41: 794–97.
16 Phillips R, Horsfi eld C, Kuijper S, et al. Sensitivity of PCR targeting
the IS2404 insertion sequence of Mycobacterium ulcerans in an assay
using punch biopsy specimens for diagnosis of Buruli ulcer.
J Clin Microbiol 2005; 43: 3650–56.
17 WHO. Global plan to combat neglected tropical diseases 2008–2015.
pdf (accessed Nov 13, 2009).
18 MacCallum P, Tolhurst JC, Buckle G, Sissons HA. A new
mycobacterial infection in man. J Pathol Bacteriol 1948; 60: 93–122.
19 Rondini S, Horsfi eld C, Mensah-Quainoo E, Junghanss T,
Lucas S, Pluschke G. Contiguous spread of Mycobacterium
ulcerans in Buruli ulcer lesions analysed by histopathology and
real-time PCR quantifi cation of mycobacterial DNA. J Pathol
2006; 208: 119–28.
20 Amofah G, Asamoah S, Afram-Gyening C. Eff ectiveness of excision
of pre-ulcerative Buruli lesions in fi eld situations in a rural district
in Ghana. Trop Doct 1998; 28: 81–83.
21 Teelken MA, Stienstra Y, Ellen DE, et al. Buruli ulcer: diff erences in
treatment outcome between two centres in Ghana. Acta Trop 2003;
22 Debacker M, Aguiar J, Steunou C, Zinsou C, Meyers WM, Portaels F.
Buruli ulcer recurrence, Benin. Emerg Infect Dis 2005; 11: 584–89.
23 Revill WDL, Morrow RH, Pike MC, Ateng J. A controlled trial of the
treatment of Mycobacterium ulcerans infection with clofazimine.
Lancet 1973; 302: 873–77.
24 Espey DK, Djomand G, Diomande I, et al. A pilot study of
treatment of Buruli ulcer with rifampin and dapsone. Int J Infect Dis
2002; 6: 60–65.
25 Etuaful S, Carbonnelle B, Grosset J, et al. Effi cacy of the
combination rifampin-streptomycin in preventing growth of
Mycobacterium ulcerans in early lesions of Buruli ulcer in humans.
Antimicrob Agents Chemother 2005; 49: 3182–86.
26 WHO. Provisional guidance on the role of specifi c antibiotics in the
management of Mycobacterium ulcerans disease (Buruli ulcer).
pdf (accessed Dec 14, 2009).
27 Chauty A, Ardant MF, Adeye A, et al. Promising clinical effi cacy of
streptomycin-rifampin combination for treatment of Buruli ulcer
(Mycobacterium ulcerans disease). Antimicrob Agents Chemother 2007;
28 Sizaire V, Nackers F, Comte E, Portaels F. Mycobacterium ulcerans
infection: control, diagnosis, and treatment. Lancet Infect Dis 2006;
29 Siegmund V, Adjei O, Racz P, et al. Dry-reagent-based PCR as a
novel tool for laboratory confi rmation of clinically diagnosed
Mycobacterium ulcerans-associated disease in areas in the tropics
where M ulcerans is endemic. J Clin Microbiol 2005; 43: 271–76.
30 Finkelstein DM. A proportional hazards model for interval-censored
failure time data. Biometrics 1986; 42: 845–54.
31 Fay MP. Rank invariant tests for interval censored data under the
grouped continuous model. Biometrics 1996; 52: 811–22.
32 Fay MP. Interval: weighted logrank tests and NPMLE for interval
censored data. R package version 0.7–5.5. 2009. http://
Oct 21, 2009).
33 Nienhuis WA, Stienstra Y, Thompson WA, et al. BURULICO drug
trial for M ulcerans infection in Ghana: Interim results. 48th Annual
ICAAC/IDSA 46th Annual Meeting; Washington, DC, USA;
Oct 25–28, 2008. Poster L-626A.
34 Raghunathan PL, Whitney EA, Asamoa K, et al. Risk factors for
Buruli ulcer disease (Mycobacterium ulcerans infection): results
from a case-control study in Ghana. Clin Infect Dis 2005;
35 Debacker M, Portaels F, Aguiar J, et al. Risk factors for Buruli ulcer,
Benin. Emerg Infect Dis 2006; 12: 1325–31.