Introduction

Human immunodeficiency virus (HIV) has continued itsrelentless spread throughout the Republic of South Africa(RSA). An estimated 770 new HIV infections occur daily,and approximately 1.8 million people in RSA are infectedwith HIV.1

Nocardia spp. are soil saprophytes which mainly causeopportunistic disease in humans with compromisedimmunity. Transplant recipients and those infected withHIV account for most cases. Other predisposing disordersof immunity include haematological malignancies andchronic therapy with corticosteroids.2 Pulmonary infec-tion is the most common presentation, with dissemina-tion to multiple sites in the body, especially the brain.In addition, Nocardia spp. can cause a local chronicsuppurative disease of the skin and soft tissues, often of

the foot, in immunocompetent individuals, known as“mycetoma.”

Based on studies from the U.S.A.3–5 and Europe,6

nocardial infection has been regarded as an unusual com-plication of HIV infection. There is little information avail-able from Africa, although one study indicated thatnocardial infection may be an important cause of mortal-ity in African HIV-infected individuals.7

In this study we have attempted, over a 2- year period,to estimate the incidence of nocardial infection compli-cating HIV infection in Soweto, RSA. We have recordedclinical details and where possible the outcome of suchinfections. Microbiologically, we have aimed to identify tospecies level the Nocardia isolates using biochemical meth-ods and to perform in vitro susceptibility testing using abroth microdilution method.

Patients and Methods

The setting for the study was Chris Hani BaragwanathHospital, Soweto. This is a large tertiary referral hospital

Journal of Infection (2000) 41, 232–239doi:10.1053/jinf.2000.0729, available online at http://www.idealibrary.com on

Nocardial Infection as a Complication of HIV in South Africa

N. Jones*1, M. Khoosal1, M. Louw2 and A. Karstaedt3

1Department of Microbiology, Chris Hani Baragwanath Hospital, South African Institute for Medical Research,PO Box 1038, Hospital Street, Johannesburg 2000, RSA, 2Department of Mycology, South African Institute

for Medical Research, PO Box 1038, Hospital Street, Johannesburg 2000, RSA, 3Department of Medicineand the University of the Witwatersrand, Chris Hani Baragawanath Hospital, Johannesburg, RSA

Objectives: To assess the occurrence, clinical and microbiological features of nocardial infections complicating HIV inSoweto, South Africa.Methods: A prospective study was carried out over a 2-year period. Patients were identified after isolation of Nocardiaspp. from a clinical specimen. Clinical details were recorded. The nocardial isolates were identified to species level andsusceptibility tests performed.Results: Ten patients were identified as having nocardial disease complicating HIV. Clinical presentations were pul-monary (five patients), pulmonary and cerebral (one patient), cerebral (one patient) and skin and soft tissue infectionof the lower limb (three patients). Three infections were fatal. The isolates were Nocardia asteroides (seven patients),N. farcinica (two patients) and Nocardia spp. (one). Isolates of N. farcinica demonstrated opacification of Middlebrookagar. All isolates were sensitive to amikacin and minocycline. Most nocardial isolates were susceptible to cefotaxime,imipenem and coamoxiclav. In vitro resistance to cotrimoxazole was present in five.Conclusions: Nocardial infection occurs as a complication of HIV infection in the Republic of South Africa. Pulmonarycases may be difficult to distinguish from tuberculosis. Nocardia asteroides is the most common species isolated.Nocardia farcinica has resistance to multiple antibacterial agents and demonstrates opacification of Middlebrook agar,a useful screening test for this species. Agents with good in vitro antinocardial activity were amikacin, minocycline,cefotaxime, imipenem and coamoxiclav. There was a high level of resistance in vitro to cotrimoxazole.

© 2000 The British Infection Society

0163-4453/00/060232�08 $35.00/0

* Please address all correspondence to: N. Jones, Department of Microbiol-ogy, Level 7 John Radcliffe Hospital, Headley Way, Oxford OX3 9DU, U.K.Accepted for publication 10 August 2000.

© 2000 The British Infection Society

Nocardial Infection and HIV 233

in an urban area of RSA, near Johannesburg. Social con-ditions in the local communities are generally poor: thereare a number of large squatter camps. Estimated sero-prevalence of HIV in this area (from antenatal clinicattenders) is 15.7%.1

A prospective study was carried out from the depart-ment of Microbiology at Chris Hani BaragwanathHospital over a 2-year period, 1995–1997. Adult patientswere identified after isolation of Nocardia spp. from a clin-ical specimen. Clinical records for each patient werereviewed to determine the diagnosis, underlying diseases,complications, outcome and therapy administered. In allcases the clinical presentation or the finding of nocardialinfection was considered sufficient reason for HIV testing.Microbiology laboratory records were reviewed to deter-mine the number of new cases of pulmonary tuberculosisdiagnosed and for the approximate seroprevalence of HIVin medical admissions. Hospital administration recordswere consulted for medical admission numbers.

Microbiology methods

MicroscopySamples of sputum or pus were smeared onto glass slidesand allowed to air dry. Gram stains were examined for thepresence of Gram-positive branching filamentous struc-tures (Fig. 1), which were acid fast with the modifiedKinyoun stain. In the presence of such structures, sam-ples were called “suspicious of Nocardia.”

CultureSputum and pus were plated onto MacConkey agar and5% horse blood agar. Plates were incubated in air at 37 °Cfor 48 h (incubation was extended for 2 weeks if

microscopy was “suspicious of Nocardia”). Colonies resem-bling Nocardia spp. were subcultured onto LowensteinJensen slopes and onto Middlebrook (7H10) agar and fur-ther identified using biochemical reactions. Biochemicalcriteria were resistance to lysozyme, differential hydrolysisof caesin, urea, tyrosine, xanthine and hypoxanthine andtheir fermentation of a range of sugars.8

In vitro susceptibility testingOrganism inoculum was prepared using a method basedon that described by Wallace et al.9 The organisms weregrown in trypticase soy broth with 0.05% Tween 80 for48 h at 37 °C. The suspension of organisms was matchedto a 0.5 MacFarland standard and diluted 100 times, to afinal inoculum of 106 colony forming units per ml.

Broth microdilution method

Serial, two-fold dilutions of antimicrobial agents in cationsupplemented Mueller Hinton broth were dispensed into 96well microtitre plates. Organisms were inoculated from thestandard suspension. The plates were covered, placed inplastic bags and incubated in air at 37 °C for 48 h. The min-imum inhibitory concentration (MIC) was defined as thelowest concentration of the agent (mg/l) that prevented vis-ible growth of the organism (or 80% inhibition in the caseof cotrimoxazole). The antibiotics tested were ampicillin,erythromycin, cotrimoxazole, coamoxiclav, imipenem,cefotaxime, amikacin, minocycline, ciprofloxacin, chlo-ramphenicol and kanamycin. Breakpoint values deter-mined by the National Committee for Clinical LaboratoryStandards (NCCLS) were used to determine susceptibility.10

Results

Over the 2-year period of study, Nocardia spp. was isolatedfrom 10 patients; in all cases it was felt to be an infectingorganism. The 10 patients were ethnically black Africanand were seropositive for HIV antibodies. Six of the 10patients were female. The average age of patients was35.7 years (range 25–53 years), 31.7 years for femalepatients and 41.8 years for males. Clinical details aresummarized in Table I.

No patient was receiving antimicrobials, including cotri-moxazole prophylaxis or antiretroviral medication, at thetime of diagnosis of nocardial infection.

CD4 T-lymphocyte count is not routinely performed atour hospital, and such data was only available for fourpatients. Only one other case of nocardial infection wasseen in this period: an 8-year-old HIV seronegative malepatient with scalp and periorbital infection due to Nocardiaasteroides.

Figure 1. Slide of sputum sample (500� magnification), Gramstained, showing Gram-positive branching filamentous structures, sus-picious of Nocardia spp.

N. Jones et al.234

Clinical syndromes of nocardial infection seen werepulmonary infection, cerebral infection and soft tissueinfection of the lower limb with cellulitis and sinus tractformation. Three patients died.

Pulmonary infection

Six of 10 patients presented with pulmonary infection.The major presenting symptoms were cough, weight lossand night sweats. Chest radiograph appearances werevariable: left upper lobe shadowing (Fig. 2), large cavita-tory lesions with air fluid levels (Fig. 3), hilar lymph-adenopathy and consolidation. Two of six patients wereco-infected with a second HIV-associated pathogen(Mycobacterium avium complex (MAC) (one case), Pneumo-cystis carinii (one case)). Four patients had been com-menced on empirical antituberculous therapy. In twopatients the infection proved fatal; one also had cerebralinfection.

Cerebral infection

Two patients had cerebral abscess diagnosed by culture ofpus aspirated from the brain; in one, this was a complica-tion of pulmonary disease (Fig. 4). Both cases were fatal, inone prior to the administration of antinocardial therapy.The second patient was treated with oral cotrimoxazoleuntil susceptibility tests showed resistance to this agent.Clinical response had been poor so therapy was changed tooral minocycline, but the patient died after a few weeks oftherapy. MAC was subsequently cultured from blood; thisinfection was not recognized prior to death.

Soft tissue infection

Three patients had cellulitis with subcutaneous puscollection. In two, this was associated with sinus tractformation. All three involved the lower limb, either thighor ankle. All three patients were clinically believed tohave pyogenic soft tissue infection, none was thoughtto resemble mycetoma clinically. On the basis of pus cul-ture, Nocardia was felt to be the primary pathogen intwo patients and to be a copathogen with Staphylococcusaureus in the third. One patient was treated with oralcotrimoxazole, a second with oral minocycline, accordingto the results of susceptibility tests. A third patient refusedmedical treatment.

Therapy

Agranulocytosis, thought to be secondary to cotrimoxazole,developed in one patient. Therapy was changed to intra-venous amikacin and cefotaxime followed by oral minocy-cline. Two patients treated empirically with cotrimoxazolesubsequently had susceptibility test results demonstratingin vitro resistance to that treatment. Both patients had afatal outcome; one also had undiagnosed MAC infection.

Incidence of pulmonary nocardia infection comparedto pulmonary tuberculosis

During the 2-year study period there were approximately55 000 medical admissions to Baragwanath Hospital,with an estimated HIV seroprevalence of 26% (per-sonal communication, Dr AS Karstaedt). Over the sameperiod, pulmonary tuberculosis (PTB) was diagnosedmicrobiologically in 1200 patients, 540 (45%) were

Table I. Clinical and microbiological details for 10 patients with Nocardial disease complicating HIV infection.

Patient Age Sex Diagnosis CD4 Therapy Outcome Species Copathology

1 29 M Cerebral – Surgery Died (3 days) N. farcinica –2 30 F 1. Cerebral 102 1. Sulf/trim Died (6 weeks) N. farcinica MAC

2. Pulmonary 2. Minocycline*3 25 F Pulmonary – Sulf/trim Lost to follow-up N. asteroides –4 53 M Pulmonary 71 Sulf/trim Well (15 months) N. asteroides PCP5 36 M Pulmonary – None Lost to follow-up N. asteroides –6 28 F Pulmonary – Sulf/trim Died (2 weeks) N. asteroides –7 29 F Pulmonary 13 1. Sulf/trim Well (2 months) N. asteroides –

2. Amikacin,cefotaxime†3. Minocycline‡

8 38 F Subcutaneous – Sulf/trim Lost to follow-up N. asteroides S. aureus9 49 M Subcutaneous – None Lost to follow-up N. spp. –

10 40 F Subcutaneous 30 Minocycline Well (2 weeks) N. asteroides –

Abbreviations: CD4, CD4 T-lymphocyte count; Sulf/trim, cotrimoxazole; MAC, Mycobacterium avium complex; PCP, Pneumocystis carinii pneu-monia; S. aureus, Staphylococcus aureus.Key to table: *therapy changed when results of in vitro susceptibility tests showed resistance to cotrimoxazole; †therapy changed after agran-ulocytosis, thought to be secondary to cotrimoxazole, developed; ‡therapy changed from intravenous to oral regime.

Nocardial Infection and HIV 235

HIV-infected.11 We can therefore estimate that 3.78% ofHIV-infected medical admissions had microbiologicallyproven PTB, 0.07% had nocardial infection (0.04% pul-monary nocardial infection). This gives a ratio of one caseof pulmonary nocardial infection to 90 cases of PTB inHIV-infected medical in-patients.

Microbiology results

Identification of Nocardia speciesIn all patients Nocardia spp. was isolated from either pus orsputum samples. In seven patients Nocardia asteroides wasisolated, in two N. farcinica, and in one the species ofNocardia could not be determined by biochemical means.Both the N. farcinica isolates demonstrated opacification ofMiddlebrook agar (7H10).12

In vitro susceptibility testing (Table II)Complete results were available for nine isolates: one iso-late was not viable on subculture. Nine isolates weresusceptible to minocycline and amikacin, eight were

Figure 2. Chest radiograph (patient 2) showing consolidation predom-inantly of the left upper lobe.

Figure 4. CT scan of the head with intravenous contrast (patient 1).There is a mass lesion with ring enhancement in the right frontoparietalarea of the brain. The lesion is surrounded by oedema and there is mid-line shift with obliteration of the ventricles.

Figure 3. Chest radiograph (patient 3) showing cavitatory lesions withair fluid levels in the right upper lobe and left lower lobe.

N. Jones et al.236

Tabl

e II

.In

vit

rosu

scep

tibi

lity

resu

lts,

usi

ng

a br

oth

mic

rodi

luti

on m

eth

od, f

or N

ocar

dial

isol

ates

.

Isol

ate

from

pat

ien

t

Age

nt

12

45

67

89

10

Susc

epti

ble,

Bre

akpo

int*

num

ber

(%)

Am

p�

64

.0�

64

.06

4.0

16

.01

6.0

32

.03

2.0

�6

4.0

64

.00

(0)

�0

.25

Ery

�6

4.0

64

.04

.02

.02

.02

.01

6.0

16

.02

.00

(0)

�0

.5A

ug

16

.0/8

.08

.0/4

.00

.25

/8

.0/4

.08

.0/4

.01

.0/0

.52

.0/1

.08

.0/4

.00

.12

5/

8(8

9)

�8

.0/4

.00

.12

50

.06

Cot

8.0

/15

2�

8.0

/15

20

.5/9

.50

.5/9

.5�

8.0

/15

22

.0/3

8.0

0.5

/9.5

�8

.0/1

52

�8

.0/1

52

4(4

4)

�2

.0/3

8.0

Imi

1.0

0.5

1.0

1.0

4.0

0.2

5�

64

.01

.00

.58

(89

)�

4.0

Cef

�6

4.0

16

.00

.25

0.2

5�

0.1

25

0.2

54

.02

.04

.07

(78

)�

8.0

Am

i2

.01

.0�

0.1

25

�0

.12

5�

0.1

25

0.1

25

1.0

1.0

4.0

9(1

00

)�

16

.0M

in2

.02

.01

.01

.01

.0�

0.1

25

1.0

2.0

2.0

9(1

00

)�

4.0

Cip

0.2

50

.25

16

.02

.02

.03

2.0

8.0

2.0

32

2(2

2)

�1

.0C

hl

64

.01

6.0

64

.01

6.0

�6

4.0

32

.06

4.0

64

.06

4.0

0(0

)�

8.0

Kan

�6

4.0

�6

4.0

8.0

64

.08

.06

4.0

�6

4.0

�6

4.0

16

.03

(33

)�

16

.0

Dat

a ar

e m

inim

um

inh

ibit

ory

con

cen

trat

ion

in m

g/l.

*Bre

akpo

ints

for

min

imu

m in

hib

itor

y co

nce

ntr

atio

ns

(mg/

l), f

rom

NC

CLS

gu

idel

ines

.10

Key

to ta

ble:

am

p, a

mpi

cilli

n; e

ry, e

ryth

rom

ycin

; au

g, c

oam

oxic

lav;

cot

, cot

rim

oxaz

ole;

imi,

imip

enem

; cef

, cef

otax

ime;

am

i, am

ikac

in; m

in, m

inoc

yclin

e; c

ip, c

ipro

floxa

cin

, ch

l, ch

lora

m-

phen

icol

; kan

, kan

amyc

in.

Nocardial Infection and HIV 237

susceptible to coamoxiclav and imipenem, and seven weresusceptible to cefotaxime. All isolates were resistant toampicillin, chloramphenicol and erythromycin, sevenwere resistant to ciprofloxacin, six to kanamycin and fiveisolates were resistant to cotrimoxazole. Both isolates ofN. farcinica showed resistance to a number of antibiotics,including third-generation cephalosporins.

Discussion

Nocardia infections have been reported from RSA for manyyears. In 1927 Pijper and Pullinger described “a few casesof chronic lung affection in white persons, where tuberclebacilli were absent, and fungal threads of the Nocardia-typewere observed in sputum”.13 They also comment that “theease with which new and undescribed Nocardias are foundin RSA affords additional proof that such infections are offrequent occurrence.” The authors describe two newspecies, N. transvaalensis and N. pretoriana. More recently apaper from Zimbabwe was entitled “Nocardiosis: a neglectedchronic lung disease in Africa.”14 This paper reportedseveral cases and suggested that nocardial lung disease isunderdiagnosed in Southern Africa.

Investigators in West Africa (Abidjan, Côte d’ Ivoire)found that 10 of 247 (4%) of the post mortems carriedout on HIV seropositive persons had evidence of nocardialinfection. They describe a ratio of 9 PTB cases : oneNocardia case.15 A prospective study of 120 chest clinicadmissions from Côte d’ Ivoire found a similar ratio ofnocardial infections (five cases, 4.2%), with 58 cases ofpulmonary tuberculosis (48%).16

In Nigeria, nocardial isolation rates of 4.1% from unse-lected sputa17 and 5% for sputa from patients with bron-chopulmonary diseases, have been reported.18

It is difficult to draw comparisons given differentmethodology, but nocardial disease would appear to berelatively less common in RSA compared with Nigeria orCôte d’ Ivoire. We calculated that there were 90 cases ofPTB for each case of pulmonary nocardial infection inHIV-infected patients. This finding may in part be due tostudy design. The cases described in Côte d’ Ivoire were atpost mortem, thus were the most severe end of the spec-trum. Gugnani et al.17 cultured sputa from Nigerianpatients with bronchopulmonary disease using a paraffinbaiting technique to improve detection of nocardia. Inour study conventional methods of culture for aerobicrespiratory pathogens were used, which are known to berelatively insensitive in the isolation of nocardia. Thedetection rate can be as low as 15%.19,20 Alternatively, thereason for the apparent low ratio of pulmonary nocardialinfection to PTB may be due to the extremely high inci-dence of tuberculosis in RSA.11

Nocardial infection is fairly unusual in the U.S.A.,occurring in 0.19–0.3% of AIDS patients.5 The use of pro-phylactic cotrimoxazole in HIV cases in the U.S.A. isthought to inhibit Nocardia as well as Pneumocystiscarinii.21 It is unusual for HIV seropositive persons in RSAto be taking long-term antibiotic prophylaxis.

Nocardial pulmonary disease can have a very similarpresentation to pulmonary tuberculosis. In RSA, duringthe study period, incidence rates of tuberculosis wereapproximately 225 per 100 000.11 Four patients with pul-monary nocardial infection were empirically treated forPTB. It seems likely that patients with nocardial infectionwill be mistaken for “smear negative” PTB. The distinctionis important, as sputa sent for TB culture alone areunlikely to grow nocardia. The harsh decontaminationstep of this process usually results in inhibition ofNocardia spp. The difficulty in distinguishing these two dis-orders has been commented on previously.22,23 The radio-logical features24 of pulmonary nocardial infection arealso not diagnostic and may be mistaken for a number ofother chronic respiratory infections, including TB.

Pulmonary presentation was most common in thiscase series, with cerebral disease being the most frequentsecondary site from the lung. Diverse presentations ofnocardial disease in AIDS have been described, includingjoint25 and bone26 infection, paraspinal abscess,5 adrenalinsufficiency27 and renal abscess.28

The patients described with skin and soft tissue infec-tion had an unusual presentation. Pyogenic infection orcellulitis rather than mycetoma was suspected by theadmitting physician in each case; nocardial infection wasnot suspected. Cutaneous disease, both localized and dis-seminated, has been described in HIV.29 In Spain, amongstintravenous drug users with HIV, cutaneous nocardial dis-ease is common.30 Intravenous drug use was not thoughtto be present in any of the patients we have described.

The outcome for the patients in our study was fairly dis-mal. Three patients died soon after diagnosis. Nocardialinfection was felt to be the cause of death in two patients.In the third patient cerebral nocardial infection and undi-agnosed MAC infection were thought responsible fordeath. Follow-up was not achieved in several patients.AIDS was diagnosed clinically in most, but CD4 T-cellcount was only available in four.

Nocardia asteroides was the most frequent isolate in ourstudy and normally accounts for the majority of humannocardial infections.21 Nocardia farcinica has been called anemerging pathogen in Germany, where this species makesup 43.7% of all nocardial isolates.2 In a French study, arecent increase in the isolation of Nocardia farcinica hasbeen noted, and was thought to be related to the increasingprevalence of HIV. This group also found a higher mortality

N. Jones et al.238

in patients with N. farcinica.6 Both of the N. farcinica infec-tions described in our study were cerebral and bothresulted in a fatal outcome. The increasing occurrence ofN. farcinica which is resistant to many antimicrobialagents, including third-generation cephalosporins,31 mayresult in difficulties in treating patients or delay in thecommencement of appropriate therapy. It has been sug-gested that opacification of Middlebrook agar byN. farcinica can be used as an early indicator of possiblemultiresistance in a nocardial isolate.12

In vitro susceptibility tests performed on the study iso-lates showed that intravenous agents with good in vitroactivity against Nocardia spp. were imipenem, amikacinand cefotaxime. Oral agents with good antinocardial activ-ity were minocycline and coamoxiclav. Cotrimoxazoleresistance was found in five of nine isolates studied here.Other groups have detected in vitro resistance to sulph-onamides and cotrimoxazole in 0–100%9,32 of isolatestested. The reason for this wide variation in in vitro suscep-tibility is not clear. Geographical differences betweennocardial isolates and different methodology are possiblecauses. A standardized method for in vitro susceptibilitytesting of Nocardia spp. is still not available. Recently diskdiffusion,33 the E test method34 and radiometric methodshave been used with some success.34 In conclusion, theincreasing incidence of HIV in countries like RSA meansthat clinicians and laboratory staff must be vigilant to thediagnosis and isolation of this unusual organism.Pulmonary nocardial infection may be difficult to distin-guish from tuberculosis, which is prevalent in areas ofhigh HIV seroprevalence. Once diagnosed, nocardial infec-tion is potentially treatable with a number of inexpensiveoral antibiotics. Areas of interest, which require furtherinvestigation, include the methodology to increase detec-tion of Nocardia spp. from sputum samples, the develop-ment of rapid standardized methods for susceptibilitytesting and the choice and duration of antibiotic therapy.

Acknowledgements

The authors would like to thank Professor H Koornhof and ProfessorH Crewe-Brown for advice in the laboratory study and in the preparationof this manuscript.

References1 Department of Health, Republic of South Africa. Sixth National HIV

Survey of women attending antenatal clinics of the public HealthServices in the Republic of RSA. Epidemiological Comments 1997; 23:4–16.

2 Schaal KP, Lee HJ. Actinomycete infections in humans – a review. Gene1992; 115: 201–211.

3 Uttamchandani RB, Daikos GL, Reyes RR et al. Nocardiosis in 30patients with advanced Human Immunodeficiency Virus infection:Clinical Features and outcome. Clin Infect Dis 1994; 18: 348–353.

14 Javaly K, Horowitz HW, Wormser GP. Nocardiosis in patients withHuman Immunodeficiency virus infection. Medicine 1992; 71:128–138.

15 Kim J, Minamoto GY, Grieco MH. Nocardial Infection as a complica-tion of AIDS: report of six cases and review. Rev Infect Dis 1991; 13:624–629.

16 Boiron P, Provost F, Chevrier G, Dupont B. Review of nocardial infec-tions in France 1987 to 1990. Eur J Clin Microbiol Infect Dis 1992;11: 709–714.

17 Lucas SB, Hounnou A, Peacock C et al. The mortality and pathologyof HIV infection in a West African city. AIDS 1993; 7: 1569–1579.

18 Mishra SK, Gordon RE, Barnett DA. Identification of nocardiae andstreptomycetes of medical importance. J Clin Microbiol 1980; 11:728–736.

19 Wallace RJ, Steele LC, Sumter G, Smith JM. Antimicrobial suscepti-bility patterns of Nocardia asteroides. Antimicrob Agents Chemother1988; 32: 1776–1779.

10 National Committee for Clinical Laboratory Standards 1997.Methods for dilution antimicrobial susceptibility testing for bacteriawhich grow aerobically. Approved standard M7-A4. NationalCommittee for Clinical Laboratory Standards, Villanova, Pa.

11 Karstaedt AS, Jones N, Khoosal M, Crewe-Brown HH. The bacteriol-ogy of pulmonary tuberculosis in a population with a high HIV sero-prevalence. Int J Tuberc Lung Dis 1998; 2: 312–316.

12 Flores M, Desmond E. Opacification of Middlebrook agar as an aid inidentification of Nocardia farcinica. J Clin Microbiol 1993; 31:3040–3041.

13 Pijper A, Pullinger BD. South African nocardiases. J Trop Med Hyg1927; 12: 153–156.

14 Baily GG, Neill P. Robertson VJ. Nocardiosis: a neglected chronic lungdisease in Africa? Thorax 1988; 43: 905–910.

15 Lucas SB, Hounnou A, Peacock C, Beaumel A, Kadio A, De Cock KM.Nocardiosis in HIV positive patients: an autopsy study in WestAfrica. Tuber Lung Dis 1994; 75: 301–307.

16 Koffi N, Aka-Danguy E, Ngom A, Kouassi B, Yaya BA, Dosso M.Prevalence of nocardiosis in an area of endemic tuberculosis. RevMal Respir 1998; 15: 643–647.

17 Gugnani HC, Unaogu IC, Emeruwa CN. Incidence of pulmonaryinfection due to nocardial species in Nigeria. Mycoses 1991; 34:359–361.

18 Osoagbaka OU, Njoku-Obi AN. Nocardiosis in pulmonary diseases inparts of Nigeria. J Trop Med Hyg 1985; 88: 367–372.

19 Simpson GL, Stinson EB, Egger MJ, Remington JS. Nocardial infec-tions in the immunocompromised host: a detailed study in a definedpopulation. Rev Infect Dis 1981: 3: 492–507.

20 Coker RJ, Bignardi G, Horner P, Savage M, Cook T, Tomlinson D,Weber J. Nocardia infection in AIDS: a clinical and microbiologicalchallenge. J Clin Pathol 1992; 45: 821–822.

21 McNeil MM, Brown JM. The medically important aerobic actino-mycetes: epidemiology and microbiology. Clin Microbiol Rev 1994; 7:357–417.

22 Hovette P, Camara P, Raphenon G, Burgel PR, Donzel C, N’Dir MC.The case of drug-resistant pulmonary tuberculosis was nocardiosis!Int J Tuberc Lung Dis 1999; 3: 742–743.

23 Aisu TO, Eriki PP. Morrissey AB, Ellner JJ, Daniel TM. Nocardiosismimicking pulmonary tuberculosis in Ugandan AIDS patients. Chest1991; 100: 888.

24 Kramer MR, Uttamanchandani RB. The radiographic appearance ofpulmonary nocardiosis associated with AIDS. Chest 1990; 98:382–385.

25 Ray TD, Nimitongskul P, Ramsey KM. Disseminated Nocardia aster-oides infection presenting as Septic arthritis in a patient with AIDS.Clin Infect Dis 1994; 18: 256–257.

26 Schiff TA, Klirsfeld D, McNeil MM, Brown J. Cutaneous Nocardiosiscaused by Nocardia nova occurring in an HIV infected individual: acase report and review of the literature. J Acquir Immun Defic Syndr1993; 6: 849–854.

27 Arabi Y, Fairfax MR, Szuba MJ, Crane L, Schumann P. AdrenalInsufficiency, recurrent bacteraemia and disseminated abscesses

Nocardial Infection and HIV 239

caused by Nocardia asteroides in a patient with AIDS. Diagn MicrobiolInfect Dis 1996; 24: 47–51.

28 Diego Miralles G. Disseminated Nocardia farcinica infection in anAIDS patient. Eur J Clin Microbiol Infect Dis 1994; 13: 497–500.

29 Boixeda P, Espana A, Suarez J, Buzon L, Ledo A. CutaneousNocardiosis and human immunodeficiency virus. Int J Dermatol1991; 30: 804–805.

30 Sanchez Munoz Torrero JF, Yniguez TR, Garcia Onieva E, Pascua FJ,Crespo L, Bacaicoa A, Martin C. Nocardiosis in patients with humanimmunodeficiency virus infection in Spain. Rev Clin Esp 1995; 195:468–472.

31 Wallace RJ, Tsukamura M, Brown B, Brown J, Steingrube VA, Zhang Y,Nash DA. Cefotaxime-Resistant Nocardia asteroides strains are iso-

lates of the controversial species Nocardia farcinica. J Clin Microbiol1990; 28: 2726–2732.

32 Boiron P, Provost F. In-vitro susceptibility testing of Nocardia spp. andits taxonomic implications. J Antimicrob Chemother 1988; 22:623–629.

33 Wallace RJ, Septimus EJ, Musher DM, Martin RR. Disk diffusion sus-ceptibility testing of Nocardia species. J Infect Dis 1977; 135:568–676.

34 Biehle JR, Cavallieri JR, Saboulle MA, Getsinger LJ. Comparative eval-uation of the E test for susceptibility testing of Nocardia species. DiagnMicrobiol Infect Dis 1994; 19: 101–110.