ANTIMICROBIAL AGENTS AND CHEMOTHERAPY,
Copyright © 2000, American Society for Microbiology. All Rights Reserved.
Dec. 2000, p. 3285–3287Vol. 44, No. 12
Early Bactericidal Activity of Paromomycin (Aminosidine) in
Patients with Smear-Positive Pulmonary Tuberculosis
P. R. DONALD,1* F. A. SIRGEL,2T. P. KANYOK,3L. H. DANZIGER,3A. VENTER,2F. J. BOTHA,4
D. P. PARKIN,4H. I. SEIFART,4B. W. VAN DE WAL,5J. S. MARITZ,6AND D. A. MITCHISON7
Departments of Pediatrics and Child Health,1Pharmacology,4and Internal Medicine,5Tygerberg Hospital and The
University of Stellenbosch, Tygerberg, and National Tuberculosis Research Programme2and MRC Centre
for Epidemiological Research,6Cape Town, South Africa; Department of Pharmacy Practice Colleges of
Pharmacy and Medicine, University of Illinois at Chicago, Chicago, Illinois3; and Department
of Microbiology, St. George’s Hospital Medical School, London, United Kingdom7
Received 4 April 2000/Returned for modification 27 June 2000/Accepted 25 August 2000
The early bactericidal activity of the aminoglycoside paromomycin (aminosidine) in doses of 7.5 and 15 mg/
kg of body weight was measured in 22 patients with previously untreated smear-positive pulmonary tubercu-
losis. The fall in log10CFU per milliliter of sputum per day during the first 2 days of treatment for 7 patients
receiving a paromomycin dosage of 7.5 mg/kg/day was 0.066, with a standard deviation (SD) of 0.216 and con-
fidence limits from ?0.134 to 0.266, and that for 15 patients receiving 15 mg/kg/day was 0.0924, with an SD of
0.140 and confidence limits from 0.015 to 0.170. The difference between the mean and zero was not significant
for the 7.5-mg/kg dose group but was significant for the 15-mg/kg dose group (t ? 2.55, P ? 0.023). Since paro-
momycin has no cross-resistance with streptomycin and has no greater toxicity than other aminoglycosides,
these results suggest that it has the potential to substitute for streptomycin in antituberculosis regimens and
may be a particularly valuable addition to the drug armamentarium for the management of multidrug-re-
The early bactericidal activity (EBA) of an antituberculosis
drug reflects its ability to kill the rapidly multiplying organisms
present in the cavities of pulmonary tuberculosis patients. It is
determined by measuring the rate of decrease in viable CFU of
Mycobacterium tuberculosis per milliliter of sputum during the
first 2 days of treatment with the drug under investigation (10).
It has been used to evaluate and compare both new (13, 14)
and established (3, 15) antituberculosis drugs and to determine
the lowest effective dose of a drug (3).
Paromomycin, also known as aminosidine, is a broad-spec-
trum aminoglycoside closely related structurally to neomycin
and kanamycin and less closely related to streptomycin (5).
Although its main use at present is for the management of
visceral leishmaniasis (kala-azar) (6, 16), it has been shown,
both in vitro (8) and in animal experiments (9), to have con-
siderable activity against M. tuberculosis, including multidrug-
resistant strains, and there are anecdotal reports of its use for
the management of pulmonary tuberculosis (17). It lacks cross-
resistance with streptomycin and other antimycobacterial
agents. The MIC of paromomycin for M. tuberculosis ranges
from ?0.09 to ?1.5 ?g/ml (8). In humans, peak concentrations
of paromomycin of 11.6 to 25.6 ?g/ml 1 h after a 500-mg
intramuscular dose have been reported (4). In this pilot study,
the EBA of paromomycin at dosages of 7.5 and 15 mg/kg of
body weight was evaluated.
MATERIALS AND METHODS
This study was undertaken in the Department of Internal Medicine of Tyger-
berg Hospital, the teaching hospital of the Faculty of Medicine of the University
of Stellenbosch. Tygerberg Hospital serves a number of socioeconomically de-
prived communities in the Western Cape province of South Africa, a region with
a notified tuberculosis incidence of ?500 cases/100,000 people in 1998. The study
was undertaken between November 1997 and June 1998.
Patients. Patients had newly diagnosed, previously untreated, pulmonary tu-
berculosis, were between 18 and 40 years of age, and weighed more than 40 kg.
Patients in poor general condition or suffering from other serious medical com-
plications, women who were pregnant or lactating, and patients unable to pro-
duce at least 10 ml of sputum overnight were excluded from the study, as were
those in whose initial sputum or urine specimens traces of isoniazid or its
metabolites could be detected. Also excluded were those with any known hyper-
sensitivity to any aminoglyocoside and those whose initial M. tuberculosis isolates
were found to be resistant to isoniazid, rifampin, streptomycin, or paromomycin.
Resistance to isoniazid, rifampin, or streptomycin, if found, would indicate that
the patient might have received previous antituberculosis treatment. Thirty-two
patients were randomized to receive either 15 mg of paromomycin per kg (18
patients) or 7.5 mg of paromomycin per kg (14 patients). The results from 10
patients, 7 from the 7.5-mg/kg group and 3 from the 15-mg/kg group, were
excluded from the analysis. The reasons for exclusion were resistance to isoniazid
in the isolates for five patients, resistance to isoniazid and inadequate homoge-
nization of the sputum specimen for one patient, and no growth or only very poor
growth for four patients.
Sputum collection and drug administration. Patients were actively encour-
aged to cough, and a 16-h collection of sputum was done from 1600 on the day
of admission to 0800 the next day (S1 sputum sample). Soon after 0800, paro-
momycin was given by intramuscular injection, and the sputum collection pro-
cedure was repeated to obtain sputum specimens following the first and second
doses of paromomycin (S2 and S3, respectively).
On completion of the study protocol after the S3 sputum collection, the
patients were commenced on isoniazid, rifampin, pyrazinamide, and ethambutol
as recommended by the South African National Tuberculosis Control Pro-
Microbiologic methods. Sputum in the S1, S2, and S3 collections was examined
conventionally by direct smear, culture, and sensitivity testing. CFU counts on
the sputum collections were carried out as described previously (13). Without
preliminary centrifugation, 20 ?l of the dilutions were set up on thirds of trip-
licate plates of selective 7H10 medium. Drug resistance did not develop during
the 3 days of the study. Analysis of sputum and urine specimens for isoniazid and
its metabolites was done by a previously described high-performance liquid
chromatography method (12).
Statistical methods. The EBA for each patient was calculated by first obtain-
ing the mean CFU count (X) at the most appropriate dilution, which was that
permitting counting of between 20 and 200 colonies. Then, we used the equation
Y ? log10(f X), where f is the dilution factor and Y is log10CFU per millititer of
sputum. This calculation was performed for the S1, S2, and S3 sputum collections
to give Y1, Y2, and Y3, respectively, and then EBA was calculated as (Y1? Y3)/2.
The means and standard deviations of Y1, Y2, Y3, and EBA were calculated,
* Corresponding author. Mailing address: Department of Paediat-
rics and Child Health, Faculty of Medicine, P.O. Box 19063, Tyger-
berg, 7505, South Africa. Phone: (021) 9389506. Fax: (021) 9389138.
together with 95% confidence limits (95% CL). These results were compared to
those for three previously described no-drug control groups, all derived from a
similar patient population at our institution and evaluated in 1992 (group A) (13)
and 1993 (group B) (14) and between July 1996 and April 1998 (group C) (15),
and one no-drug control group evaluated between November 1996 and April
1999 (not previously described; group D).
Differences between group means were examined first by one-way analysis of
variance, in which between-group variation was regarded as a random effect (the
no-drug control groups were from different studies). Since there were consider-
able differences in the intragroup variances, the group means were further
compared using the nonparametric trend test for trends across nonparametric
groups using Stata, version 6 (Stata, College Station, Tex). Finally, regression
analysis was performed on the group mean values, regarding the no-drug control
groups as receiving a dose of 0 mg/kg of body weight, to study the relationship
between dose and EBA.
The study protocol was approved by the ethics committee of the Medical
Faculty of the University of Stellenbosch. All patients entered in the study gave
written informed consent for their participation.
The 22 patients included in the final analysis had a mean age
of 36 years and a mean weight of 52 kg. All but one patient
were found to have multicavitary disease by chest radiography.
The means and standard deviations of the CFU per milliliter
sputum for the S1, S2, and S3 sputum samples and the mean
EBA, with standard deviations and 95% Cl, are shown in Table
1 and compared to the results obtained for the no-drug control
groups. A one-way analysis of variance comparing the mean
EBA of the six groups gives an F5.57of 1.75 and a P value of
0.137, indicating no clear evidence of real differences between
group means. Since there was considerable heterogeneity in
the intragroup variances (with variances increasing substan-
tially with the timing of the study), the P value reported above
may not be reliable. Hence, a nonparametric test for trend with
increasing dose was carried out; this test gave an F7of 2.24 and
a P value of 0.03, suggesting that the mean EBA increases with
an increasing dose. In addition, the CL for the mean EBA at a
dose of 15 mg/kg do not include zero, providing definite evi-
dence of an effect (P ? 0.023). There is, however, no evidence
that the mean EBA for the 7.5-mg/kg dose differed from zero
(P ? 0.45). This evidence of a dose effect was examined further
using regression analysis. Figure 1 shows a plot of mean EBA
against dose for the six groups in Table 1. Fitting a line by
weighted least squares, with weights equal to the number of
observations per group, gives an intercept of 0.011 with a
standard error of 0.013 and a slope of 0.00562 with a standard
error of 0.00169. The slope differs significantly from zero, with
a t4of 3.329 and a P value of 0.029. Fitting a line through the
six mean points and basing the test of significance on just four
degrees of freedom is a very conservative approach, yet the
result supports the hypothesis that there is a dose effect. The
fitted straight line and 95% confidence band are also shown in
Of the patients excluded from analysis because of resistance
to isoniazid, three received a dose of 7.5 mg/kg, and their EBA
results were 0.053, 0.019, and 0.156 (mean, 0.076). The remain-
ing two isoniazid-resistant patients received a dose of 15 mg/kg
and had EBAs of 0.049 and 0.102 (mean, 0.076).
In this study, paromomycin produced a small but statistically
significant increase in EBA, as shown by both the nonparamet-
ric trend test (P ? 0.03) and the regression analysis on the
group means (P ? 0.029). Streptomycin, given in a dosage of
approximately 20 mg/kg of body weight to a group of four
patients, had a similar EBA, 0.094 (7). The CL of both esti-
mates are so wide that comparison is inconclusive. The finding
that paromomycin at a dose of 7.5 mg/kg did not have a de-
tectable effect compared to zero suggests that the antitubercu-
losis activity may be limited.
Our findings and the results of earlier in vitro (8) and in vivo
(9) studies suggest, but do not prove, that paromomycin has an
antituberculosis activity similar to that of streptomycin. Impor-
tantly, paromomycin has no cross-resistance with streptomycin
and does not appear to have any higher incidence of ototoxicity
or nephrotoxicity than that found with other aminoglycosides
(1, 2, 6, 11). Paromomycin may therefore be a valuable addi-
tion to the antituberculosis drug armamentarium, particularly
for the management of multidrug-resistant tuberculosis.
This study, FD-R-001167-01, was funded by an orphan products
grant from the U.S. food and Drug Administration, Office of Orphan
FIG. 1. Plot showing linear regression of EBA of the 7.5- and 15-mg/kg
paromomycin groups and the no-drug groups.
Table 1. Viable counts of CFU of tubercle bacilli in sputum collections S1, S2, and S3 and in groups receiving no drug
Dosage (mg/kg) or no-drug
group (reference no.)
Log10CFU/ml of sputum (mean [SD]) for: EBA
S1 S2S3 Mean (SD)95% Confidence
aThe difference between the means and zero was nonsignificant for the 7.5-mg/kg group but significant (t ? 2.55, P ? 0.023) for the 15-mg/kg group.
bND, not determined.
3286DONALD ET AL.ANTIMICROB. AGENTS CHEMOTHER.
1. Arcamone, F., C. Bertazzoli, A. Buogo, G. Cassinelli, and A. Vigevani. 1974.
New studies in the field of aminoglycosidic aminocyclitol antibiotics. Pan-
minerva Med. 16:9–24.
2. Chunge, C. N., J. Owate, H. O. Pamba, and L. Donno. 1990. Treatment of
visceral leishmaniasis in Kenya by aminosidine alone or combined with
sodium stibogluconate. Trans. R. Soc. Trop. Med. Hyg. 84:221–225.
3. Donald, P. R., F. A. Sirgel, F. J. Botha, H. I. Selfart, D. P. Parkin, M. Van
den Plas, B. W. Van de Wal, J. S. Maritz, and D. A. Mitchison. 1997. The
early bactericidal activity of isoniazid related to its dose size in pulmonary
tuberculosis. Am. J. Respir. Crit. Care Med. 156:895–900.
4. Eastwood, J. 1962. Paromomycin: serum levels following parenteral admin-
istration. Antibiot. Chemother. 12:77–84.
5. Fisher, M. W., M. C. Manning, L. A. Gagliardi, M. R. Gaetz, and A. L.
Erlandson. 1960. Paromomycin: experimental antibacterial activity. Anti-
biot. Annu. 7:293–303.
6. Jha, T. K., P. Olliaro, C. P. Thakur, T. P. Kanyok, B. L. Singhania, I. J.
Singh, N. K. Singh, S. Akhoury, and S. Jha. 1998. Randomised controlled
trial of aminosidine (paromomycin) v. sodium stibogluconate for treating
visceral leishmaniasis in North Bihar, India. Br. Med. J. 316:1200–1205.
7. Jindani, A., V. R. Aber, E. A. Edwards, and D. A. Mitchison. 1980. The early
bactericidal activity of drugs in patients with pulmonary tuberculosis. Am.
Rev. Respir. Dis. 121:939–949.
8. Kanyok, T. P., M. V. Reddy, J. Chinnaswamy, L. H. Danziger, and P. R. J.
Gangadharam. 1994. Activity of aminosidine (paromomycin) for Mycobac-
terium tuberculosis and Mycobacterium avium. J. Antimicrob. Chemother.
9. Kanyok, T. P., M. V. Reddy, J. Chinnaswamy, L. H. Danziger, and P. R. J.
Gangadharam. 1994. In vivo activity of paromomycin against susceptible
and multidrug-resistant Mycobacterium tuberculosis and M. avium complex
strains. Antimicrob. Agents Chemother. 38:170–173.
10. Mitchison, D. A. 1998. Basic concepts in chemotherapy of tuberculosis,
p. 15–50. In P. R. J. Gangadharam and P. A. Jenkins (ed.), Mycobacteria:
chemotherapy, vol. 2. Chapman Hall, New York, N.Y.
11. Scott, J. A. G., R. N. Davidson, A. L. T. Moody, H. R. Grant, D. Felmingham,
and G. M. S. Scott. 1992. Aminosidine (paromomycin) in the treatment of
leishmaniasis imported into the United Kingdom. Trans. R. Soc. Trop. Med.
12. Seifart, H. I., W. L. Gent, D. P. Parkin, P. P. Van Jaarsveld, and P. R.
Donald. 1995. High performance liquid chromatographic determination of
isoniazid, acetylisoniazid and hydrazine in biological fluids. J. Chromatogr. B
13. Sirgel, F. A., F. J. H. Botha, D. P. Parkin, B. W. Van de Wal, P. R. Donald,
P. K. Clark, and D. A. Mitchison. 1993. The early bactericidal activity of
rifabutin in patients with pulmonary tuberculosis measured by sputum viable
counts: a new method of drug assessment. J. Antimicrob. Chemother. 32:
14. Sirgel, F. A., F. J. H. Botha, D. P. Parkin, B. W. Van de Wal, R. Schall, P. R.
Donald, and D. A. Mitchison. 1997. The early bactericidal activity of cipro-
floxacin in patients with pulmonary tuberculosis. Am. J. Respir. Crit. Care
15. Sirgel, F. A., P. R. Donald, J. Odhiambo, W. Githui, K. C. Umqpathy, C. N.
Paramasivan, C. M. Tam, K. M. Kam, C. W. Lam, K. M. Sole, D. A.
Mitchison and the EBA Collaborative Study Group. 2000. A multicentre
study of the early bactericidal activity of antituberculosis drugs. J. Antimi-
crob. Chemother. 45:859–870.
16. Thakur, C. P., P. Olliaro, Gothoskars, S. Bhowmick, B. K. Choudhury, S.
Prasad, M. Kumar, and B. B. Verma. 1992. Treatment of visceral leishman-
iasis (kala-azar) with aminosidine (paromomycin)-antimonial combinations,
a pilot study in Bihar, India. Trans. R. Soc. Trop. Med. Hyg. 86:615–616.
17. Tokuoka, S., S. Oka, T. Narusawa, and J. Kondo. 1970. A study of the
effectiveness of aminosidine against pulmonary tuberculosis. Jpn. J. Chest
VOL. 44, 2000 EARLY BACTERICIDAL ACTIVITY OF PAROMOMYCIN3287