A pilot study evaluating the tolerability and effectiveness of miltefosine and sodium stibogluconate for the treatment of cutaneous leishmaniasis in Kandahar, Afghanistan

Abstract

Background

Oral miltefosine (MF) achieves per-protocol cure rates of 68–77% against Leishmania tropica in Iran and Pakistan, but it has never been compared with the standard of care in Afghanistan, where only intralesional (IL) or intramuscular (IM) antimonials are used. We, therefore, assessed the 6-month efficacy and tolerability of sodium stibogluconate (SSG) and MF, including MF pharmacokinetics, in L. tropica-predominant Kandahar city.

Methods

This single-centre, open-label, pilot trial randomised patients aged ≥5 years to either SSG or 28-day, partially supervised MF that was allometrically scaled from the adult dose of 2.5 mg/kg/day. For the SSG group, patients were treated with either IL-SSG ×1/week ×5 doses or IM-SSG 20 mg/kg/day x20d for IL-inaccessible lesions. All children < 5 years received MF in an observational arm. Epithelialisation of ulcerated lesions or return to normal skin for other lesions, with resolution of signs of acute inflammation, defined cure.

Results

Of 103 recruited patients (65 males), 15 were randomised to IL-SSG, 27 to IM-SSG, and 37 to MF; 24 under-5 s received MF. Respective median (interquartile range) ages were: 11.9 (7.2–13.8), 9.9 (8.2–15.1), 10 (7.8–12.5) and 4.3 (2.9–4.9) years. A total of 113/191 (59.2%) lesions were located on the face. 66 patients completed the 6-month follow-up. Combining the MF groups, the Kaplan–Meier cure rates (95% confidence intervals) were MF 25.3 (14.7‒40.9)%, IL-SSG 65.0 (36.0‒91.5)% (p=0.019 vs. MF), and IM-SSG 77.3 (53.7‒94.3)% (p=0.001 vs. MF). Median MF exposures_0-180d were similar in failed (629 µg*day/mL) and cured (555 µg*day/mL) patients. 52/61 (85.2%) MF recipients reported mild (n=50)/moderate nausea/vomiting and 35/52 (67.3%) received domperidone. Laboratory adverse events were mostly grade 1 ALT(n=19)/AST(n=36) increases in the IM-SSG and MF arms. Transient stage 1/2 acute kidney injury affected 11/24 (45.8%) under-5 s vs. 5/37 (13.5%) ≥5 years MF patients (p=0.007).

Conclusions

SSG was effective by both routes, but partially supervised MF was ineffective. More work is needed to determine factors affecting MF effectiveness against L. tropica in Kandahar.

Clinical trial number The clinical trial number is TCTR20180710007, dated 10/07/2018.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s41182-026-01075-6.

Background

Old World cutaneous leishmaniasis (OWCL) is a sand fly-transmitted disease of war and poverty caused predominantly by Leishmania major, L. tropica, and L. aethiopica that extends from West Africa to parts of western China [1]. The countries with the highest OWCL burdens are Afghanistan, Algeria, Iran, Iraq, Pakistan, and Syria [2].

In Afghanistan, OWCL is caused by L. major and L. tropica and is called “Saldana” or “Kaldana”, “that which lasts 1 year.” L. major is a zoonosis transmitted mostly by Phlebotomus papatasi (the great gerbil and Libyan jird are the main reservoirs) and is mostly rural, whilst L. tropica, transmitted between humans by P. sergenti, is found in large cities like Kabul, Kandahar and Herat [3]. Natural healing rates are ≥1 year for L. tropica and 2–8 months (M) for L. major, with 3-M cure rates of ~1 and ~70%, respectively [4].

Treatment aims to accelerate healing and improve the cosmetic result. Only sodium stibogluconate (SSG) and meglumine antimoniate (MA) are available in public sector clinics in Afghanistan, and, anecdotally, they remain effective. In 2003 in L. tropica-endemic Kabul, SSG achieved per protocol (PP), 100-day (D) cure rates of ~75% for intralesional (IL) SSG (total of 5 injections every 5–7 days), ~45% for intramuscular (IM) SSG [odds ratio 3.75 (1.86–7.54)], and ~70% for thermotherapy; corresponding, approximate intention to treat (ITT) rates were 48, 22 and 54% [5]. These rates are broadly similar to the 6 M ITT cure rates of 82.5% (thermotherapy) and 74% (IL-MA), also in 2012 in Kabul [6], but lower than the 92% PP 12 M rate of IL-SSG (7-weekly doses) in L. tropica, acquired in Rajasthan, India [7].

Oral miltefosine (MF) is unavailable in Afghanistan but has been used in non-immune foreign soldiers. In 5 L. tropica-infected Canadian soldiers, acquired in southern Afghanistan, 4 were cured after 28-day MF, and 1 responded well (MF discontinued after 7 days) [8]. MF was also effective in 34 Dutch soldiers, 27 with confirmed L. major, who were billeted in Mazar-e-Sharif, northern Afghanistan [9]; 30 (~ 88%) were cured at 12 M (28 by 6 M), and the four failures responded to IL/IV-SSG.

L. tropica in Quetta, Pakistan, was cured after 28-D MF in 77% (40/52) evaluable (PP) patients (median age 48, IQR 13–59 years) at 3 M [10]; all had either failed SSG or had SSG contraindications. Tolerability was good, with 28 patients reporting 39 transient, mild (31) and moderate (8), mostly gastrointestinal, adverse events: nausea (16/39, 41%), abdominal pain (5/39, 13%), vomiting (5/39, 13%), and diarrhoea (2/39, 5%). A small series of 8 L. tropica-infected Israeli children used 3 mg/kg/day (maximum 150 mg/day) of MF for 28–90 (mean ~45) days; 6/8 were cured, including 1 with a retreated relapse, and 2 did not respond [11].

Given the paucity of controlled, randomised MF efficacy data in CL in Afghanistan and the uncertainty regarding the optimal dose in children, we conducted a pilot, clinical and pharmacokinetic (PK) study of allometrically scaled MF, comparing its effectiveness to SSG.

Methods

Study design, site & participants

This open-label, randomised, single-centre, parallel-group trial of CL took place from December 14, 2019, to October 20, 2021 (COVID-19 interruption April 6–May 2021), at the National Malaria and Leishmaniasis Control Programme clinic in Kandahar, a city endemic for L. tropica [3]. All children <5 years were enrolled in an observational, MF-only arm.

If routinely seen patients were Giemsa-slide positive for amastigotes and potentially eligible for the study, they were referred to the research team to undergo consent and evaluation for study entry. They were enrolled if they were aged ≥1–60 years (males) or ≥1–<14 years (females, to avoid the risk of pregnancy) who weighed ≥10 kg, gave written informed consent, and were willing to be followed up for 6 M. Exclusion criteria were any one of: (i) severe lesions requiring hospital treatment (physician’s decision), (ii) solitary lesion <2 cm in a child <5 years or <3 cm in a patient aged ≥5 years, (iii) pregnant/breast feeding/intending to have a child in the next 12 M, (iv) known allergy to SSG/MF, (v) known serious underlying disease e.g. liver, renal, cardiac, HIV, tuberculosis, on systemic steroids, (vi) creatinine clearance <30 mL/min, assessed by Cockcroft Gault (adults) and Schwartz (children) formulae, (vii) ALT >3× upper limit of normal (ULN)/total bilirubin >1.5× ULN, (viii) abnormal cardiac rhythm/prolonged Bazett-corrected QT interval (QTcB) for age and sex.

Randomisation and masking

For patients aged ≥5 years, SSG or MF allocation was determined using a computer-generated randomisation list and placed in a sealed opaque envelope that was only opened after patients had been enrolled and given a study number (001, 002, etc.). Within the SSG group, IM-SSG was given for lesions that were prospectively determined to be inaccessible to IL-SSG, namely, on the fingers, toes, nose, ears, eyelids, lips, or genitalia.

Interventions

SSG was obtained from Albert David Limited, Kolkata, India. Each vial contains 30 mL SSG: 1 mL of SSG ≡ 100 mg of pentavalent antimony. SSG dosing followed local/WHO guidelines: (i) IL-SSG was given ×1/week ×4w (5 doses); the administered dose varied between 1 and 5 mL, depending on the size of the lesion, and (ii) IM-SSG was dosed at 20 mg SSG/kg/day ×20 days (Supplementary Table 1).

28-D MF (Impavido® as 10/50 mg capsules, Knight Therapeutics Inc, St. Michael, Barbados) was dosed by total body weight, using a research team-designed, allometrically scaled regimen (Table 1) intended to obtain similar exposures in children as the standard adult dose of 150 mg/day in patients ≥45 kg. We used allometric scaling of exposure with total body weight (BW), in accordance with Anderson and Holford [12], assuming a daily target dose of 150 mg in an adult patient at 60 kg as the target value (Eq. 1).

$$MFdose\left(\mathit{mg}\right)=150mg\times {\left(\frac{\mathit{BW}}{60}\right)}^{0.75}$$ 1

Table 1.

aaa

Weight (kg)	Daily dose (mg)	Given as	mg/kg/day dose
10‒<15	40	20 mg bid	2.7‒4
15‒<25	60	50 mg a.m. + 10 mg p.m	2.4‒4
25‒<30	80	50 mg a.m. + 30 mg p.m	2.7‒3.2
30‒<45	100	50 mg bid	2.2‒3.3
≥45	150	50 mg tid	1.7‒3.3

Bid, twice a day; tid, three times a day

IL- and IM-SSG injections and morning MF doses were given in the clinic. The other MF capsules were given to the patients/guardians to be taken at home, aided by daily phone call reminders as well as pill counts. To reduce gastrointestinal side effects, MF was administered with cake (clinic), and patients were advised to take MF with food. Vomiting within one hour resulted in a repeat MF dose. Domperidone was available for troublesome nausea or vomiting. Patients who failed or poorly tolerated their treatment were referred to the dermatologist for further management.

Conduct of clinical trial

All enrolled participants underwent a history of the current illness, symptoms (symptom checklist), past medical and family histories, a physical examination, and laboratory testing: full blood count, routine biochemistry, serum amylase (to detect SSG-induced pancreatitis), hepatitis B surface antigen (HBsAg), and the glucose-6-phosphate-dehydrogenase deficient Mediterranean variant (G6PDd^Med) by PCR–RFLP (exon 6) [13]. Skin lesions were measured with a ruler at the longest/widest points at the inflamed-normal skin junctions, determined with the naked eye. If there was >1 lesion, the larger/est was selected for serial measurements, photographs, and swabs.

Follow-up was 6 M: D3, D7, D14, D21, D28, D42, then monthly M2–6 for clinical and laboratory evaluations, in addition to the medication days (Supplementary Table S2). At each visit, participants/parents/legal guardians were asked “How are you/is your child today?” and the symptom checklist was administered to detect adverse or serious adverse events (SAEs). All AEs were treated as clinically indicated.

All patients consented to lesion photographs, using a mobile phone and the Antera® camera, a 3D camera that assesses inflammation and lesion height and depth. Lesional swabs were taken pre-treatment for Leishmania species typing and sequencing of the heat shock protein (hsp)70 gene [14]. Briefly, total DNA was extracted from swab samples using the DNeasy extraction kit (Qiagen), and 2.5 μl of DNA was used in PCR reactions of 25 μl, following the published typing protocol and sequencing algorithm of Van der Auwera et al. [14]. PCR parameters were adapted for the High Fidelity Platinum® Taq DNA Polymerase (Life Technologies, Thermo Fisher Scientific, USA), as described by Hoyos et al. [15]. Bidirectional sequencing was conducted at Macrogen Genomics Laboratories (Macrogen, South Korea) using 4F and 6R primers. Sequences were edited and aligned, and a dendrogram was constructed using MEGA version 12 [16]. The nucleotide substitution model that best fitted the sequencing data was the Tamura 3-parameter model. The tree was constructed using three statistical methods: Neighbour Joining, Maximum Likelihood, and Minimum Evolution. The phylogenetic analysis used the 186 available sequences of the Leishmania spp hsp70 gene [14], and bootstrapping with 1000 replicates was used as a test of phylogeny.

Dried blood spots (DBS) were taken from all MF-treated patients, using Whatman 903 protein saver cards (GE Healthcare Europe GmbH, Diegem, Belgium) to measure whole blood MF concentrations at baseline (D0H0), D1, 2, 3, 7, 14, 21, 28, and M2–6, using a liquid chromatography–tandem mass spectrometry method, validated according to United States Food and Drug Administration and European Medicines Agency guidelines and adapted from Kip et al. [17]. MF blood concentrations were measured over a quantification range of 10–2000 ng/mL, with a lower limit of quantification of 10 ng/mL. DBS samples expected to exceed 2000 ng/mL were diluted 100-fold post-extraction using the final extract of a processed blank DBS. Quality control samples were included in each batch of clinical samples. All batches met predefined acceptance criteria, with bias within ±15% and precision ≤15% at all quality control levels.

Patients aged ≥5 years underwent 12-lead ECGs (paper speed 25 mm/s, height calibrated at 10 mm/mV) at D0H0, D0+1‒3 h, D7, D14, and D28 (Jiangsu Rooe Medical Technology Co, Ltd, Suzhou, China).

Outcomes

Adapting current recommendations [18], the primary outcome was the 6 M cure rate, defined as the complete and sustained re-epithelialisation of ulcerated lesions or flattening for non-ulcerated lesions plus the resolution of signs of acute inflammation, e.g. erythema, oedema, scaling, and no residual scab. Post-study, all mobile phone skin lesions were assessed blinded to the treatment allocation by a consultant dermatologist (NW) and the study PI (WT), based on consensus. For patients seen at M6 but without a photograph, the field team diagnosis was used.

The main secondary outcomes were: (i) time to initial cure, and (ii) rates of (a) relapse, the reappearance of the original skin lesion after initial ‘cure’, and (b) failure at M6, i.e. not meeting all of the criteria for cure. In all failures and patients with early study discontinuation, skin lesions were assessed clinically (inflammation signs & size) and assessed as better, worse, or more-or-less the same vs. presentation.

Using the same criteria, we looked at the clinical course over time from baseline to the last time point patients were seen and classed these broadly as: (i) improvement from the start, (ii) worse from the start, (iii) initial improvement followed by worsening, (iv) initial worsening followed by improvement, (v) a fluctuating course, and (vi) no overall change.

Tolerability and safety were assessed using the Common Terminology Criteria for Adverse Events v5.0, which grades clinical and laboratory AEs from 1 to 5 (5 is death). AST, ALT, and total bilirubin grades use the ULN or baseline value, depending on whether the baseline value is within or outside NLs, respectively. Serum amylase grades 2 and 3 use lower values xULN for symptomatic patients. For creatinine, we decided post hoc to use the KDIGO criteria for acute kidney injury (AKI), given its simplicity, independence of sex, and applicability across all ages [19].

Sample size, statistical, and pharmacokinetic analyses

Based on Afghan and Indian data [5], [6], [7], we assumed 6 M cure rates of 85% with a 10% precision, resulting in a sample size of 49/arm. We aimed to recruit 50 SSG and 50 MF patients, including at least 20 under-5 children.

Data were entered onto standardised case record forms, checked against source documents, and entered into Macro® before analysis in Stata v17 (Stata Corporation, College Station, TX, USA). Patients who received ≥1 dose of study drug were included in the safety and efficacy analyses, following a prospectively written statistical analysis plan.

The 6 M cure rates were determined by Kaplan–Meier survival analysis and compared between the arms using the log-rank test, grouping the MF patients. Patients who were lost to follow-up or withdrew were censored at the last time point seen. Inter-arm differences in proportions were compared by chi² or Fisher’s exact test, as appropriate; continuous data were analysed by the paired/unpaired ‘t’ or sign rank/Mann–Whitney U tests, as appropriate. Independent factors for a binary outcome were assessed by logistic regression.

MF PK parameters were determined by non-compartmental analysis, using a validated in-house developed script in R version 4.4.1. Exposures (area under the drug concentration–time curve, µg*day/mL) were determined during two treatment periods: 0–28 days (treatment) and 0–180 days (end of follow-up). We explored the relationships between MF dose and nausea and vomiting, and MF exposures with cured vs. failed, grade 1/2 increases in AST and ALT vs. no increases, and AKI vs. no AKI. Detailed PK and PK-pharmacodynamic (PD) analyses will be presented elsewhere. All statistical analyses were two-sided, and a p value of <0.05 was considered statistically significant.

Role of the funding source

The funders had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to publish.

Results

Patient disposition and baseline characteristics

Of the 1047 screened individuals, 103 were enrolled. A total of 994 patients had 1076 reasons for non-inclusion; the major reason (721/994, 72.5%) was the 6 M follow-up (Fig. 1). 24 under-5 children were enrolled in the observational study, and 79 patients were randomised to IL-, IM-SSG, or MF; 66 patients completed the 6 M follow-up.

Fig. 1.

Consort trial profile by arm

Baseline characteristics (Table 2) were broadly similar across the randomised groups. Overall, the median time to presentation was 112 days, and no patients had previously sought treatment. Few had underlying illnesses, none were HBsAg positive, 12 had G6PD^Med (8 heterozygous females, 4 hemizygous males), and routine haematology and biochemistry were normal in the majority. L. tropica was detected from lesional swabs in 4 of 15 patients; their positions on the phylogenetic tree are shown in Supplementary Fig. 1. The negative samples were of low DNA quality for sequencing.

Table 2.

Baseline characteristics

Characteristics	Intralesional SSG (n=15)	Intramuscular SSG (n=27)	Oral miltefosine ≥5y (n=37)	Oral miltefosine <5y (n=24)	Total (n=103)
Age (years)†	11.9 (7.2‒13.8)	9.9 (8.2‒15.1)	10.0 (7.8‒12.5)	4.3 (2.9‒4.9)	8.5 (5.2‒12.3)
Sex
Male: Female	9: 6	17: 10	25: 12	14: 10	65: 38
Weight (kg)	26 (19‒38)	25 (19‒36)	25 (18‒38)	14 (12‒18)	21 (16‒30)
Height (cm)	135 (115‒144)	126 (116‒156)	126 (115‒147)	96 (89‒106)	119 (106‒139)
BMI (kg/m2)‡	18.1 (6.2)	15.9 (3.2)	16.6 (4.2)	17.2 (4.0)	16.8 (4.3)
Temperature (ºC)†	37.1 (36.5‒37.3)	37.1 (37.0‒37.4)	37.0 (36.8‒37.3)	37.0 (36.3‒37.2)	37.0 (36.5‒37.3)
Respiratory rate (beats/min)†	25 (22‒26)	22 (20‒25)	25 (22‒26)	24 (22‒26)	24 (22‒26)
Heart rate (breaths/min)†	80 (76‒90)	84 (75‒97)	88 (78‒100)	102 (90‒110)	90 (80‒105)
Systolic blood pressure (mmHg)†	100 (90‒110)	90 (80‒105)	100 (80‒110)	80 (80‒88)	90 (80‒105)
Diastolic blood pressure (mmHg)†	70 (65‒75)	60 (60‒70)	65 (60‒70)	60 (55‒60)	62 (60‒70)
Past medical history§	2 (13)	2 (7)	3 (8)	2 (8)	9 (9)
Lung disease	0 (0)	1 (4)	2 (5)	1 (4)	4 (4)
Kidney disease	1 (7)	0 (0)	0 (0)	0 (0)	1 (1)
Visceral Leishmaniasis	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
HIV	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Other	1 (7)	1 (4)	1 (3)	1 (4)	4 (4)
G6PD status§
Normal	14 (93)	22 (81)	34 (92)	21 (88)	91 (88)
Hemizygous males	0 (0)	1 (4)	1 (3)	2 (8)	4 (4)
Heterozygous females	1 (7)	4 (15)	2 (5)	1 (4)	8 (8)
Time lesion/s present before presentation (Days)¶	100 (63‒126; 54‒381)	104 (64‒190; 37‒465)	127 (64‒196; 25‒435)	117 (71‒181; 31‒299)	112 (66‒187; 25‒465)
Previous CL treatment (No)	15 (100)	17 (100)	37 (100)	24 (100)	103 (100)
Haematology†
Haemoglobin (g/dl)	12.3 (2.3)	11.7 (1.4)	11.7 (1.3)	10.8 (1.1)	11.6 (1.5)
WBC (103/µL)	6.5 (5.2‒7.7)	7.2 (6.1‒8.7)	7.7 (6.2‒8.6)	7.9 (6.6‒10.5)	7.5 (6.1‒8.8)
Neutrophil (103/µL)	3.7 (2.7–4.6)	4.5 (3.7–5.6)	4.5 (3.2–5.5)	4.3 (3.3–5.4)	4.3 (3.4–5.5)
Lymphocyte (103/µL)	2.3 (2.1–3.0)	2.3 (1.8–2.9)	2.5 (1.7–3.3)	3.1 (2.2–4.2)	2.4 (2.0–3.2)
Eosinophil (103/µL)	0.04 (0.03–0.05)	0.06 (0.04–0.10)	0.06 (0.04–0.08)	0.10 (0.06–0.17)	0.06 (0.04–0.09)
Platelets (103/µL)	252 (204‒306)	306 (244‒330)	279 (230‒320)	300 (250‒348)	290 (236‒324)
Biochemistry†
Sodium (mmol/L)	137 (135‒139)	135 (135‒139)	137 (136‒138)	136 (134‒138)	136 (135‒138)
Potassium (mmol/L)	3.9 (3.7‒4.0)	4.0 (3.8‒4.3)	4.0 (3.8‒4.1)	3.9 (3.6‒4.1)	4.0 (3.7‒4.1)
Creatinine (mg/dL)	0.5 (0.5‒0.7)	0.6 (0.4‒0.7)	0.6 (0.5‒0.7)	0.5 (0.4‒0.6)	0.5 (0.5‒0.6)
Calcium (mmol/L)	8.6 (8.2‒9.0)	8.1 (8.0‒8.5)	8.3 (8.0‒8.6)	8.0 (7.6‒8.4)	8.2 (8.0‒8.6)
Magnesium (mmol/L)	1.8 (1.6‒2.0)	1.7 (1.5‒2.0)	1.7 (1.6‒1.9)	1.6 (1.5‒2.0)	1.7 (1.5‒1.9)
ALT/SGPT (IU/L)	26 (15‒30)	21 (18‒26)	24 (19‒28)	21 (19‒27)	23 (18‒28)
AST/SGOT(IU/L)	31 (20‒39)	23 (19‒32)	30 (21‒34)	31 (26‒36)	28 (21‒34)
Total bilirubin (mg/dL)	0.7 (0.6‒0.8)	0.6 (0.5‒0.8)	0.6 (0.6‒0.7)	0.6 (0.6‒0.7)	0.6 (0.6‒0.7)
Direct bilirubin (mg/dL)	0.3 (0.3‒0.4)	0.3 (0.3‒0.4)	0.4 (0.3‒0.4)	0.4 (0.3‒0.4)	0.4 (0.3‒0.4)
Amylase (U/L)	89 (71‒105)	90 (65‒122)	80 (69‒123)	90 (64‒120)	88 (66‒120)
ECG
QTcB interval (ms)†	410 (394‒418)	414 (398‒428)	414 (390‒426)	-	412 (394‒424)

^* all patients aged 5 years and above were randomised

^† median (IQR)

^‡ mean (SD)

^§ number (%)

^¶ median (IQR) and range

Skin lesion distribution

The total number of skin lesions seen was 191; 122 patients had one lesion, 17 had two lesions, six had three lesions, and three and one had four and five lesions, respectively. The face (n=113), ears (n=7), and neck (n=4) accounted for 124/191 (64.9%) of lesions, followed by the legs (n=32), arms (n=21), hands (n=13), and back (n=1). At presentation, ulcerated plaques or ulcerated nodules were observed in 65 and 30 patients, respectively (92.2%); the rest were plaques or nodules.

Cure rates and clinical course

The overall 6 M Kaplan–Meier MF cure rate was 25.3% (38.2% <5 years vs. 13.0% ≥5 years), which was significantly lower than IL-SSG 65.0% and IM SSG 77.3% (Table 3); the raw numbers are shown in Supplementary Table S3. The times to initial cure are shown in Fig. 2 and the course of a cured patient is shown in Fig. 3. Nine patients (7 on MF) made good progress and at 6 M had either fully (n=6) or almost fully epithelialised but, with residual signs of inflammation, they were classed as failures (Supplementary Table S4 & Supplementary Figure S2).

Table 3.

Cumulative Cure rates by 6 months

Treatment	Univariate Cox proportional hazard model
	Cumulative Cure rates (95% CI)%	Hazard ratio (95% CI)	p-value
Oral miltefosine (n=61)*	25.3 (14.7–40.9)	Reference
<5 years (n=24)	38.2 (21.2–62.3)
≥5 years (n=37)	13.0 (4.4–35.2)
Intralesional SSG (n=15)	65.0 (36.0–91.5)	3.33 (1.22–9.09)	0.019
Intramuscular SSG (n=27)	77.3 (53.7–94.3)	4.21 (1.85–9.56)	0.001

^* all patients combined

Fig. 2.

The skin changes in a cured patient. This patient had an initial increase in the size of the ulcerated lesion. With the loss of the scab, re-epithelialisation was observed from Month 2, leaving a scar and post-inflammatory hyperpigmentation in months 5 and 6

Fig. 3.

Kaplan–Meier plot of time to cure across the arms over 6 months

During the course of their disease, just over half experienced at least one crescendo of the inflammatory response, diagnosed clinically as a worsening or fluctuation of lesions (Table 4). Improvement from the start was associated mostly with cure, whilst initial improvement followed by worsening was associated with treatment failure and relapse. One relapse was an IM-SSG treated 7-year-old who presented with an ulcerated plaque on the proximal ulnar aspect of his right arm covered by a dark adherent scab and surrounded by a ring of erythema, measuring 4.3 cm in the longest dimension. At 1 M, the surrounding erythema was much less and associated with scaling and, by D42, almost all the scab had come off to reveal a pale patch. Epithelialisation was complete by M3, but one month later, two small dark scabs appeared that persisted to M6 (Supplementary Figure S3). The second relapsed child, age 4 years, also presented with a nasal ulcerated plaque with a small central dark adherent scab and surrounding erythema (1.6 cm). Following MF, the lesion improved, and the patient was cured at M3, but at M4, a light-coloured scab reappeared that increased in size by M6 (Supplementary Figure S4). In the 37 failed/relapse patients, there was an approximately even split between lesions that were better or worse vs. baseline.

Table 4.

Clinical course of patients

Clinical course	Cured (n=29)	Failed (n=35)	Relapse (n=2)	Lost to follow up (n=37)	Total (n=103)
Improvement from the start	20 (69.0)	5 (14.3)	0 (0)	23 (62.2)	48 (46.6)
Worse from the start	0 (0)	2 (5.7)	0 (0)	0 (0)	2 (1.9)
Initial improvement followed by worsening	0 (0)	14 (40.0)	2 (100.0)	7 (18.9)	23 (22.3)
Initial worsening followed by improvement	4 (13.8)	0 (0)	0 (0)	1 (2.7)	5 (4.9)
Fluctuating course	5 (17.2)	14 (40.0)	0 (0)	6 (16.2)	25 (24.3)
No overall change	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Overall clinical impression
Better vs. baseline	27 (93.1)	18 (51.4)	1 (50.0)	32 (86.5)	78 (75.7)
Worse vs. baseline	0 (0)	16 (45.7)	1 (50.0)	5 (13.5)	22 (21.4)
About the same	0 (0)	1 (2.9)	0 (0)	0 (0)	1 (1.0)
Missing	2 (6.9)	0 (0)	0 (0)	0 (0)	2 (1.9)

Harms and adverse events

A total of 60 patients reported 88 clinical AEs; 84/88 were grade 1 (Table 5). 52/61 (85.2%) MF-treated patients reported nausea and/or vomiting, 47 (90.4%) within the first 4 days of treatment; 35/52 (67.3%) were prescribed domperidone. Two domperidone recipients requested study withdrawal: one for mild nausea and vomiting and the other for moderate nausea with mild vomiting. Musculoskeletal symptoms were noteworthy in 4/6 IM-SSG-treated patients.

Table 5.

The frequencies of reported clinical adverse events with a positive relationship to the study drugs*

Adverse events	Possibly, probably, definitely related, x/n (%)
	Intralesional SSG (n=15)	Intramuscular SSG (n=27)	Oral miltefosine ≥5y (n=37)	Oral miltefosine <5y (n=24)	Total (n=103)
Number of participants reporting AEs	1	5	32	22	60
Number of events	2	6	50	30	88
Event†
Vomiting	0 (0)	0 (0)	30 (60)	22 (73)	52 (59)
Nausea	0 (0)	0 (0))	15 (30)	4 (13)	19 (22)
Feeling weak or sick	0 (0)	1 (17)	10 (2)	2 (7)	4 (5)
Myalgia	0 (0)	3 (50)	0 (0)	0 (0)	3 (3)
Pale skin	0 (0)	0 (0)	2 (4)	1 (3)	3 (3)
Weakness	0 (0)	0 (0)	2 (4)	0 (0)	2 (2)
Diarrhoea	0 (0)	0 (0)	0 (0)	1 (3)	1 (1)
Dizziness	1 (50)	0 (0)	0 (0)	0 (0)	1 (1)
Headache	1 (50)	0 (0)	0 (0)	0 (0)	1 (1)
Joint stiffness	0 (0)	1 (17)	0 (0)	0 (0)	1 (1)
Itching	0 (0)	1 (17)	0 (0)	0 (0)	1 (1)

^* a total of 60 individuals reported 88 AEs: 84 were mild AEs in 58 patients & 4 were moderate in 4 individuals

^† reported as N (%) of events

Of the 61 MF recipients, 50 (81.9%) reported mild nausea or vomiting and 2 (3.3%) moderate nausea or vomiting: 19/52 patients had nausea and vomiting, 16/19 were treated with domperidone, and 33/52 patients had vomiting alone with 19/33 receiving domperidone. Overall, 35/52 (67.3%) patients took domperidone for their symptoms

During follow-up, the majority of patients had laboratory values within the normal range (Table 6); none had increases in serum amylase. The IM-SSG group saw most of the increases in ALT and AST. The 5 patients with grade 2 ALT increases also had increases in AST (graded 1–3), and all 4 patients with AST grades 2/3 (n=1) had increases in ALT but normal total bilirubin concentrations, not meeting Hy’s law for potential drug-induced liver failure. Two patients (MF, IL-SSG) had moderate increases in their total bilirubin with normal ALT and AST concentrations. Evidence of AKI (Fig. 4) was detected in 26% of the IM-SSG arm and just under 28% of the MF recipients, with a significantly greater proportion (p=0.007) of children <5 years vs. ≥5 years: 45.8% (11/24) vs. 13.5% (5/37). For all arms combined, there was a downward trend in the median Hb concentration but no consistent Hb change as a function of G6PD status or treatment arm (Supplementary Figures S4 & S5).

Table 6.

Changes in laboratory parameters over time and detected adverse events

Parameter	Intralesional SSG (N=15)	Intramuscular SSG (N=27)	Oral miltefosine ≥ 5 yrs (N=37)	Oral miltefosine <5 yrs (N=24)	Total (N=103)
ALT*
Normal	12 (80.0)	9 (33.3)	34 (91.9)	21 (87.5)	76 (73.8)
Grade 1	3 (20.0)	13 (48.2)	3 (8.1)	3 (12.5)	22 (21.4)
Grade 2	0 (0)	5 (18.5)	0 (0)	0 (0)	5 (4.9)
AST†
Normal	9 (60.0)	5 (18.5)	29 (78.4)	14 (58.3)	57 (55.3)
Grade 1	6 (40.0)	18 (66.7)	8 (21.6)	10 (41.7)	42 (40.8)
Grade 2	0 (0)	3 (11.1)	0 (0)	0 (0)	3 (2.9)
Grade 3	0 (0)	1 (3.7)	0 (0)	0 (0)	1 (1.0)
Total bilirubin
Normal	14 (93.3)	22 (81.5)	36 (97.3)	24 (100.0)	96 (93.2)
Grade 1	0 (0)	5 (18.5)	0 (0)	0 (0)	5 (4.9)
Grade 2	1 (6.7)	0 (0)	1 (2.7)	0 (0)	2 (1.9)
Creatinine#
No AKI	12 (80.0)	20 (74.1)	32 (86.5)	13 (54.2)	77 (74.8)
Stage 1	0 (0)	3 (11.1)	2 (5.4)	9 (37.5)	17 (16.5)
Stage 2	3 (20.0)	4 (14.8)	3 (8.1)	2 (8.3)	9 (8.7)
Amylase
Normal	15 (100.0)	27 (100.0)	37 (100.0)	24 (100.0)	103 (100.0)

^* Alanine transaminase (ALT)

^† Aspartate aminotransferase (AST)

^# KDIGO classification of acute kidney injury (AKI):

Stage 1 – serum creatinine increase of 1.5 to 1.9 × baseline or increase of ≥ 0.3 mg/dL (≥ 27 mol/L)

Stage 2 – serum creatinine increase 2 to 2.9 × baseline

Stage 3 – serum creatinine increase 3 × baseline or increase of ≥ 4 mg/dL (≥ 354 mol/L)

Fig. 4.

Time trends in liver enzymes, total bilirubin, and creatinine in all patients. Upper limits of normal used by the local laboratory: ALT—ULN 42 U/L (male), 32 U/L (female), AST—37 U/L (male), 31 U/L (female), total bilirubin—1 mg/dL, creatinine—1.1 mg/dL (adult male), 0.9 mg/dL (adult female)

Pharmacokinetic–pharmacodynamic relationships

A total of 9 (15%) patients did not complete the full 28-day treatment (all lost to follow-up) and were only included in the determination of AUC_D0-7. Four patients had PK evidence of missing MF doses. The median AUC_0-180 values were similar between failed vs. cured patients: 629 vs. 555 µg*day/mL (p=0.412) and between children <5 years vs. ≥5 years: 602 vs 589 µg*day/mL (p=0.616) as were the median AUC_0-28 values between those with and without laboratory AEs: (i) ALT 455 vs. 396 µg*day/mL (p=0.16), (ii) AST 416 vs. 391 µg*day/mL (p=0.593), and (iii) grade1/2 AKI vs. no AKI 384 vs. 400 µg*day/mL (p=0.71), respectively.

The median mg/kg/day dose and its distributions were similar between patients who reported: (i) nausea 3.07 mg/kg/day vs. no nausea 3.3 mg/kg/day (p=0.087) and (ii) vomiting 3.2 mg/kg/day vs. no vomiting 3.3 mg/kg/day (p=0.868). Age did not influence vomiting (6.9 vs 6.2 years, p=0.878) but was higher in those who reported nausea: 11 vs. 5 years (p=0.0047) for an odds ratio of 1.27 (1.07–1.50) for every 1 year increase in age (p=0.007).

ECG changes

Over time, all QTcB values were <500 ms. Combining post-baseline QTcB values, mean values (Supplementary Figure S6) were not significantly different vs. baseline within the 3 arms: IL-SSG 406.5 vs. 401.1 ms (p=0.613), IM-SSG 413.3 vs. 421.0 ms (p=0.153), and MF 411.3 vs. 404.1 ms (p=0.098). A detailed analysis will be presented elsewhere.

Discussion

This pilot study has shown that the effectiveness of partially supervised MF was low, but the efficacies of clinic-administered IL- and IM-SSG were high. Most patients tolerated their treatments; 2 MF recipients left the study because of upper gastrointestinal upset, but injection pain/distress in the SSG arms was not recorded.

Our Kaplan–Meier MF cure rate was substantially lower than for L. tropica in Pakistan, ~77 (PP) [10] and 92% (ITT) [20], and Iran (68%, ITT 12 M) [21], L. major in Afghanistan (ITT 88% [9]) and Iran (~ 93%) [22], and the species found in Central/South America: L. panamensis (~ 90%) in Colombia [23], L. braziliensis (~ 78%) in Brazil [24], L. guyanensis (~ 72%) in Brazil [25], and ~55% for Guatemalan L. Mexicana [23] and Guatemalan L. braziliensis (~ 33% [23]).

We assessed a novel, 5-dosing band regimen that was allometrically scaled to compensate for the increased dose-normalised MF clearance in young children [26]. It delivered 2.4–4 mg/kg/day in children weighing 10‒<25 kg (~ median ages of 14 months–9 years [27]), the same as Dorlo et al.’s multiband height, weight and sex regimen [28] (used by Kamink et al. in Quetta [10]) and higher than Mohebali’s 4-band regimen, 2–3.3 mg/kg/day (Supplementary Tables S5 & 6, respectively).

Overall, the median exposures over 180 days in our failed and cured patients and under-5 s vs. ≥5 years were similar, suggesting reduced MF adherence does not fully explain the failures and PK differences cannot account for the higher cure rate in the under-5 s. These contrast somewhat with data from Castro et al. who treated Colombian children and adults with the same target dose, 2.5 mg/kg/day, for mostly L. panamensis. In that study, MF exposure (AUC_0-∞) was significantly lower (p<0.0001) in children aged 2 to 8 years (weight >10 kg) compared to adults, 652 µg*day/mL vs. 880 µg*day/mL, for corresponding 7 M efficacies of 80% vs. ~93% (p=0.21); although not significantly different, these data suggest a correlation between exposure and efficacy [26].

More PK/PD data are needed in fully supervised MF patients with L. tropica to determine whether MF is suitable for Afghanistan, including the search for molecular markers of MF resistance and in vitro drug sensitivity testing. The advent of a filter paper method has greatly facilitated field-based PK studies [17].

IL- and IM-SSG were efficacious but the study was underpowered to compare their cure rates. Reithinger et al. reported a higher PP cure rate with IL- vs. IM-SSG (75 vs. 45%) in Kabul [5] whilst an Iranian observational study reports L. tropica cure rates of 100% (90/90) and 90% (46/51) for IL- and IM-MA, respectively [29]; these data are consistent with a retrospective study of L. tropica imported into Sweden: 18/21 (85% IM/IV antimonials) 10/15 (IL antimonials) [30]. SSG remains effective in Kandahar, but we cannot rule out antimonial resistance by L. tropica, which is well-documented in Iran [31].

We observed high rates of mild nausea (~ 30%), vomiting (85%), resulting in a high rate of antiemetic use, and AKI, 25% overall and 45% in the under-5 s. Mild and transient increases in creatinine were absent in one CL study that included 22 children aged 2–12 years [24], very low (2/268, 0.75%) in one study of visceral leishmaniasis, but 33% (vs. 9% for placebo) of patients aged ≥12 years [23]. Rises in ALT and AST were grade 1 and all < 2-fold vs. baseline/ULN. None of these AEs were related to MF exposure and only nausea was related to increasing age, which may be due to better reporting in older children. IM-SSG-related AEs were expected, but there were no increases in serum amylase, and no deleterious haematological changes were observed. G6PD^Med had no impact on the evolution of the Hb concentration in all arms; in vitro, MF causes concentration-dependent haemolysis by disrupting lipid-protein dynamics within the red cell membrane [32].

We used simple efficacy endpoints: cured vs. not cured at 6 M, adopting a strict definition of cure that included the resolution of all signs of acute inflammation that has also been used by Van Thiel, Mohebali and Tayyebi [9] [21] [22]. However, the complete resolution of inflammatory signs is not a prerequisite suggested by CL experts because induration is difficult to standardise and erythema is felt to be unreliable [18]. Our stringent definition saw nine patients who were almost cured with either full or almost full skin closure; other investigators might consider some of these patients as cured. This highlights the need to continue to strive for universally accepted efficacy endpoints. Kaplan–Meier analysis was used to reduce the distortions inherent in PP and ITT analyses and is recommended as the analysis of choice [18].

We elected not to assess changes in lesion size over time as a PD marker; using the clinical photos, lesion measurements were crude, based on the boundary between normal and inflamed skin, including post-inflammatory hyperpigmentation. Moreover, attempts to measure ulceration separate to the penumbra of intact inflamed skin were far too time-consuming and were confounded by scab formation.

There were several limitations in this study. Our study was small, with limited statistical power that was eroded further by the high rate of patient loss to follow-up. Moreover, most patients declining study participation cited the 6 M follow-up as too long. Future studies should consider a more user-friendly follow-up regimen with less time spent in the clinic but ensuring adequate follow-up at home for at least 6 M. Indeed, the optimal follow-up time may exceed 6 M for the slow-healing L. tropica. For convenience, we combined the MF groups for the Kaplan–Meier analysis, cognisant that no under-5 s received IL- or IM-SSG. Nevertheless, it remains likely that MF would still have lower efficacy than SSG. The molecular diagnosis of L. tropica was confirmed only in a handful of patients (more molecular analyses are planned), but, given the urban setting, L. tropica remains the most likely species. Although swab sampling followed an SOP and, therefore, we assume sampling was good, there was poor DNA yield at analysis that may have been partly due to long storage times in hot, humid conditions. It may behove researchers to plan for these analyses to be done in a timely manner to avoid what may be adverse storage conditions.

IM-SSG was reserved for anatomically inaccessible lesions, like those on the eyelids, nose, lips, or fingers, and introduced an unavoidable treatment indication bias so that the IL- and IM-SSG groups were clinically different. Moreover, this bias may have been compounded because limited data report that facial lesions heal more slowly and are more likely to fail treatment [33], [34]. In the absence of these biases, IL-SSG was superior to IM-SSG (~ 75 vs. 45%) in a randomised trial in Kabul of single lesion CL, caused by L. tropica [5].

Conclusions

This study has provided useful data on MF effectiveness and tolerability and their relationships to MF exposure. MF effectiveness was very low, appeared independent of MF exposure, and for unclear reasons was higher in the under-5 s. This result should be reconfirmed, and MF resistance should be sought before MF is ruled out as useful for Afghanistan. MF comes in two capsule strengths, and there are no child-friendly formulations, which would be useful for adherence, given its 28-day course. SSG was effective and remains an essential drug in a country with no accessible alternatives in the public sector. CL is crying out for more effective, tolerated, cheap, child-friendly treatments.

Abbreviations

AE

Adverse event

AKI

Acute kidney injury

D

Day

DBS

Dried blood spots

G6PD

Glucose-6-phosphate dehydrogenase

HSP

Heat shock protein

IL

Intralesional

IM

Intramuscular

ITT

Intention to treat

IV

Intravenous

M

Month

MA

Meglumine antimonial

MF

Miltefosine

OWCL

Old World cutaneous leishmaniasis

PD

Pharmacodynamic

PK

Pharmacokinetic

PP

Per protocol

SSG

Sodium stibogluconate

y

Year

Author contributions

BAR, WRJT, JT, and MM worked on the original conception, funding acquisition & protocol development. BAR (site PI), KB, GMM, AWW, MHW, MSS, and WML did the study execution. WD, TD, FC, EB, TK, MI, EBatty MAG, MoAG, and NJW had substantial intellectual input. MAG, MR-C, EBatty, JThai, and MI did the molecular analyses. WS and TD did the PK analysis. BM, CT, NC, JD, KN, CP, and MM did the data management & analysis. WRJT, MM, and BAR wrote the first draft. MAG, JT, TD, and WS did a critical review of the paper. All authors reviewed the manuscript.

Funding

WRT, MM and JT are supported by the Wellcome Trust of Great Britain through its core grant (220211) to the Mahidol-Oxford Tropical Medicine Research Unit research programme. NJW was a Wellcome Trust Principal Fellow. NJW is a Wellcome Trust Principal Fellow (Ref: 223099/Z/21/Z). TPCD is supported by the Swedish Research Council (VR 2022–01251). TD was supported by the Swedish Research Council, reference VR 2022‐01251. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.

Data availability

Deidentified individual participant data and relevant supplementary data and documents from this publication are available to applicants who provide a sound proposal to the Mahidol Oxford Tropical Medicine Research Unit Data Access Committee (datasharing@tropmedres.ac). A data access agreement will be put in place before sharing.

Declarations

Ethics approval and consent to participate

This study was approved by the Oxford University Tropical Research Ethics Committee (reference 26-18, 26/7/2018) and the Afghan Ministry of Public Health institutional review board (IRB) (number 444804, 04/11/2018). All adults and parents/guardians of patients gave informed consent before study entry and children aged 12–17 years also signed an assent form. The patient information sheet (PIS) was in Pashto, the local language, and a member of the research team gave an initial verbal explanation. The PIS detailed the nature of the study, follow-up schedule, the side effects of SSG and MF, the risks of taking blood, and the total blood volume (47 mL). We emphasised the voluntary nature of the study. Potential participants/guardians were allowed as much time as they needed before making their decision. A copy of the PIS and signed Informed Consent were given to the participant.

Consent for publication

Granted.

Conflict of interest

The authors declare that there are no competing interests.