Mycobacterial blood culture as the only means of diagnosis of disseminated tuberculosis in advanced HIV infection

Abstract

The diagnosis of disseminated tuberculosis (TB) in advanced HIV infection is often delayed because of difficulty in obtaining suitable specimens for culture. A total of 32 such patients from South India with positive mycobacterial blood cultures were studied over ten years. Almost all (90%) had a febrile illness and the majority (68.7%) had clinical lung involvement, but only 27.3% had positive sputum smears. Liver biopsy yielded a positive diagnosis in only 1/7. Cytopenia was almost universal (96.9%). Bone marrow cultures were, however, positive in 54.8%, of whom one-quarter grew atypical mycobacteria. Mycobacterial blood culture is therefore a useful adjunct test to diagnose TB in advanced HIV.

Keywords

Mycobacteraemia HIV disseminated tuberculosis

Introduction

The number of cases of tuberculosis (TB) has escalated enormously, principally on account of the HIV epidemic. In advanced HIV infection, the disease is usually already disseminated which betrays a high mortality. Even at this stage, the diagnosis is often delayed because of the difficulty in obtaining suitable specimens for analysis. Our study was carried out to evaluate the usefulness of positive mycobacterial blood cultures in TB diagnosis in advanced HIV infection.

Materials and methods

Ours was a descriptive retrospective study from a review of patient records between 2005 and 2015. A total of 32 HIV-infected patients who had positive blood cultures for mycobacteria were collected from a large tertiary care referral centre in South India. All had been admitted with pyrexia of unknown origin. Blood for mycobacterial cultures were processed in Kirchner selective medium. Growth detected was subject to microscopy and subcultures were performed onto a modified Lowenstein Jensen medium to aid identification.

Details about patient demographics, clinical symptoms and signs, duration of symptoms, haematological and biochemical parameters, radiological findings, biopsies and cultures of sampled tissues and body fluids, CD4 count, HIV viral load, use of anti-retroviral therapy (ART) and species of mycobacteria identified on blood cultures were examined. Data are reported as frequencies and percentages of categorical variables and compared for significance using Pearson’s Chi-square test.

The study was approved by the Institutional review board of the Christian Medical College, Vellore on 19 April 2017.

Results

Patients’ baseline characteristics are summarised in Table 1. An outline of their clinical characteristics showed that 90% underwent investigations for a febrile illness. A total of 19 (59.3%) had other opportunistic infections, and in 17 (53.1%) HIV infection was diagnosed for the very first time along with a mycobacteraemic illness. In nine patients (28.1%), this occurred between one and five years after the diagnosis of HIV infection, and in a further five patients (15%) five years later. The CD4 count of these patients was in the range of 1–80.

Table 1.

Baseline characteristics.

Variable	n (%)
Age (years)
15–45	24 (75.0)
>45	8 (25.0)
Sex
Male	24 (75.0)
Female	8 (25.0)
Number on ART	9 (28.1)
Adherent on ART	6 (66.7)
Symptom duration (days)
<15	7 (22.6)
15–45	8 (25.8)
>45	16 (51.6)

ART, anti-retroviral therapy.

ART had already been commenced in nine patients (28.12%) at the time of diagnosis of mycobacteraemia. A past history of TB was mentioned in seven patients (21.8%), and of these five (15.6%) died.

Organ systems found to be involved are shown in Table 2; despite 22 patients (68.7%) having lung involvement (cough in 43.7%, an abnormal chest radiograph in 46.8%), sputum smears for acid-fast bacilli were positive in only six patients (27.3%). Reticulo-endothelial system involvement was frequent (90.6%), with hepatomegaly in 62.5%, splenomegaly in 40% and lymphadenopathy in 50%. Seven space-occupying lesions were detected on imaging (one splenic and six hepatic). Liver biopsy was performed in seven patients but was useful in establishing a diagnosis in only one patient in whom histopathology showed a granuloma. Mycobacterial cultures from this specimen were however unhelpful. A total of 11 lymph node biopsies were performed in 16 patients with lymphadenopathy, but only three contributed positively to a diagnosis of TB on both histopathology and culture. There was evidence of cytopenia in almost all patients (96.9%); anaemia 93.7%, leukopenia in 56.25% and thrombocytopaenia in 40.62%. Some 34.3% had a pancytopenia at presentation, but bone marrow examination revealed granulomas in only 35.5% and positive bone marrow cultures for mycobacteria in only 54.8%. Only four had evidence of neurological TB, but cerebrospinal fluid (CSF) cultures did not grow the tuberculous bacillus in any. However, the diagnosis was confirmed by sampling other involved organs in two patients (50%). One patient had genitourinary TB and 11 (34.3%) confessed to vomiting, loose stools or dysphagia, but the diagnosis could not be obtained by sampling the excreta.

Table 2.

Organs involved clinically and diagnosis obtained by appropriate sampling of organ system.

Organ system	Number involved (clinical/radiological) (%)	Diagnosis obtained from sampling organ involved (%)
Lung involvement	22 (68.8)	Sputum smear/culture 6/22 (27.3)
Reticulo-endothelial	29 (90.6)	Liver biopsy done in 7 Granuloma on HPE – 1/7 (14.2) Tissue AFB C/S – 0(0) Lymph node biopsy/FNAC done in 11 Histopathology and C/S-positive – 3/11 (27.3)
Bone marrow	31 (96.9)	Biopsy Granuloma –11/31 (35.5) AFB C/S – 17/31 (54.8)
Brain	4 (12.9) Probable TB meningitis 2/4 Tuberculoma without meningitis 2/4 Tuberculoma with meningitis1/4	CSF culture – 0 (0) (TB diagnosis by cultures from other sites – 2/4 (50))
Genito-urinary	Sterile pyuria, hydro-ureteronephrosis – 1 (3.2)	Urine 0 (0) (TB diagnosis by cultures from other sites – 1/1 (100))
Gastrointestinal	11 (34.3) (esophageal ulcer – 2, duodenal ulcer –1)	GI mucosal biopsies 0 (0)

AFB, acid-fast bacilli; C/S, culture and sensitivity; CSF, cerebrospinal fluid; FNAC, fine needle aspiration cytology; GI, gastrointestinal; HPE, histopathological examination

Mycobacterial blood cultures grew Myco. tuberculosis in 13 patients (40.6%), and 19 (61.2%) had atypical mycobacteria (probably Myco. avium intercellulare in 14, Myco. simiae in one and Myco. fortuitum in four). Only one of the isolates was multi-drug resistant.

In 12 (37.5%), mycobacterial blood cultures were the only test that yielded a positive diagnosis although other sites had been sampled. Mycobacterial blood cultures as the sole diagnostic test contributed proportionately more to the diagnosis of atypical mycobacteria (10 vs. 2 cultures, P = 0.03).

The mean turnaround time for this method of culture was 55.6 days (range = 24–83 days).

Discussion

Mycobacteraemia is found in approximately 10–20%^1,2 of adult patients with HIV infection. Predictors of are a low CD4 count, male gender and anaemia.^1,2. TB mortality rates in the presence of HIV infection are high (17.9 vs. 3.8 per 100 person-years without TB).³ Therefore, it is crucial to identify this pathology and initiate prompt treatment for both TB and HIV. In disseminated TB, invasive sampling is needed if involved tissue is likely to be accessed. This is a problem in a resource constrained setting; thus, empirical anti-tuberculous therapy (ATT) is usually commenced without such intervention.

The yield of mycobacterial blood cultures remains modest (14.5–30%).^4–6 However, cultures are a useful adjunct especially when the yield from other tissues are poor.

The method of culture used here until recently required growth in a liquid-selective medium with subsequent subcultures onto solid media. The turnaround time was therefore high. Recently, however, with automated liquid cultures such as the mycobacterial growth indicator tube (MGIT) this has been shortened to approximately two to three weeks. The only precaution to be followed is to ensure collection of blood before the start of ATT or at least within one week of its initiation.

Given the design of the study we were unable to provide data on sensitivity and the negative predictive value of mycobacterial blood cultures for the diagnosis of TB.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Nakiyingi

Ssengooba

Nakanjako

et al.

Predictors and outcomes of mycobacteremia among HIV-infected smear- negative presumptive tuberculosis patients in Uganda. BMC Infect Dis 2015; 15: 62–62.

Crump

Kendall

et al.

Predictors and outcomes of Mycobacterium tuberculosis bacteremia among patients with HIV and tuberculosis co-infection enrolled in the ACTG A5221 STRIDE study. BMC Infect Dis 2015; 15: 12–12.

Moore

Liechty

Ekwaru

et al.

Prevalence, incidence and mortality associated with tuberculosis in HIV-infected patients initiating antiretroviral therapy in rural Uganda. AIDS 2007; 21: 713–719.

Oplustil

Leite

Oliveira

et al.

Detection of mycobacteria in the bloodstream of patients with acquired immunodeficiency syndrome in a university hospital in Brazil. Braz J Infect Dis 2001; 5: 252–259.

Gopinath

Kumar

Singh

. Prevalence of mycobacteremia in Indian HIV-infected patients detected by the MB/BacT automated culture system. Eur J Clin Microbiol Infect Dis 2008; 27: 423–431.

Bacha

Cimerman

de Souza

et al.

Prevalence of mycobacteremia in patients with AIDS and persistent fever. Braz J Infect Dis 2004; 8: 290–295.